Measuring and modelling the energy demand reduction potential of

Measuring and modelling the energy demand reduction potential of
Loughborough University
Institutional Repository
Measuring and modelling the
energy demand reduction
potential of using zonal
space heating control in a
UK home
This item was submitted to Loughborough University's Institutional Repository
by the/an author.
Additional Information:
• A Doctoral Thesis. Submitted in partial fullment of the requirements for
the award of Doctor of Philosophy of Loughborough University.
Metadata Record:
Publisher:
https://dspace.lboro.ac.uk/2134/20326
c Arash Beizaee
This work is made available according to the conditions of the Creative
Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BYNC-ND 4.0) licence. Full details of this licence are available at: https://creativecommons.org/licenses/bync-nd/4.0/
Rights:
Please cite the published version.
Measuring and Modelling the Energy
Demand Reduction Potential of Using
Zonal Space Heating Control in a UK Home
by
Arash Beizaee
Doctoral Thesis Submitted in Partial
Fulfilment of the Requirements for the Award of
Doctor of Philosophy of Loughborough University
15 February 2016
© Arash Beizaee 2016
This is to certify that I am responsible for the work submitted in this thesis, that the
original work is my own except as specified in acknowledgments or in footnotes, and
that neither the thesis nor the original work contained therein has been submitted to
this or any other institution for a degree.
Arash Beizaee
15 February 2016
I
"One important idea is that science is a means whereby learning is achieved, not by
mere theoretical speculation on the one hand, nor by the undirected accumulation of
practical facts on the other, but rather by a motivated iteration between theory and
practice…"
George E.P. Box (1976)
II
Abstract
Most existing houses in the UK have a single thermostat, a timer and conventional
thermostatic radiator valves to control the low pressure, hot water space heating
system. A number of companies are now offering a solution for room-by-room
temperature and time control in such older houses. These systems comprise of
motorised radiator valves with inbuilt thermostats and time control. There is currently
no evidence of any rigorous scientific study to support the energy saving claims of
these ‘zonal control’ systems.
This thesis quantifies the potential savings of zonal control for a typical UK home.
There were three components to the research. Firstly, full-scale experiments were
undertaken in a matched pair of instrumented, three bedroom, un-furbished, 1930s,
test houses that included equipment to replicate the impacts of an occupant family.
Secondly, a dynamic thermal model of the same houses, with the same occupancy
pattern, that was calibrated against the measured results. Thirdly, the experimental
and model results were assessed to explore how the energy savings might vary in
different UK climates or in houses with different levels of insulation.
The results of the experiments indicated that over an 8-week winter period, the
house with zonal control used 12% less gas for space heating compared with a
conventionally controlled system. This was despite the zonal control system resulting
in a 2 percentage point lower boiler efficiency. A calibrated dynamic thermal model
was able to predict the energy use, indoor air temperatures and energy savings to a
reasonable level of accuracy. Wider scale evaluation showed that the annual gas
savings for similar houses in different regions of the UK would be between 10 and 14%
but the energy savings in better insulated homes would be lower.
III
Acknowledgements
I would like to acknowledge the UK Engineering and Physical Science Research
Council (EPSRC) for their financial support for the London-Loughborough (LoLo)
Centre for Doctoral Research in Energy Demand (grant EP/H009612/1) and also
Digital Energy Feedback and Control Technology Optimisation (DEFACTO) research
project (grant EP/K00249X/1) which made this study possible.
I am very thankful to my principal supervisor Dr David Allinson who has always been
available to provide guidance, support and encouragement throughout all stages of
my PhD.
My second supervisor and the director of the LoLo centre, Prof Kevin Lomas
deserves my deepest thanks for believing in me and giving me the opportunity to
study PhD. His insightful guidance, experience and critical appraisal of my research
have been extremely valuable beyond the course of this PhD.
I would also like to express my sincere gratitude to my third supervisor Prof Dennis
Loveday for his help especially during the earlier stages of the work to employ the
test houses which formed the basis of this study as well as his guidance in finalising
this thesis.
During the course of this study I have been grateful to be a part of Loughborough
University’s DEFACTO research project team and I would especially like to thank Dr
Ehab Foda for his valuable comments and suggestions on both measurement and
modelling phases of this work. I would also like to thank Dr Stephen Porritt for his
contribution towards the preparation of the test houses.
Special thanks to my external examiner Prof Staf Roels from KU Leuven and my
internal examiner Prof Malcolm Cook for their constructive feedback.
I’m thankful to all my friends, colleagues, administrators and technicians within the
School of Civil and Building Engineering and the LoLo CDT who have been very
supportive and made this journey an enjoyable one for me.
Finally, I would like to dedicate this thesis to my parents, Dr Nadezda Beizaee and
Mr Kavous Beizaee and my brother Dr Amir Beizaee for their endless love,
unconditional support, encouragement and sacrifice.
IV
Contents
Abstract ..................................................................................................................... III
Acknowledgements ................................................................................................... IV
List of Figures............................................................................................................ IX
List of Tables .......................................................................................................... XIV
Abbreviations ........................................................................................................ XVII
Energy use conversion factors ............................................................................... XXI
1
2
Introduction ......................................................................................................... 1
1.1
Background ................................................................................................... 1
1.2
Justification of the research .......................................................................... 6
1.3
Aim and objectives ........................................................................................ 6
1.4
Outline of the thesis ...................................................................................... 7
Literature review.................................................................................................. 9
2.1
Introduction ................................................................................................... 9
2.2
Space heating methods in the UK homes ..................................................... 9
2.3
Wet central heating system components and configuration ........................ 11
2.3.1
Boiler .................................................................................................... 12
2.3.2
Time switch / programmer .................................................................... 14
2.3.3
Room thermostat / Programmable room thermostat ............................. 14
2.3.4
Thermostatic Radiator Valves (TRVs) .................................................. 15
2.3.5
Motorised valve .................................................................................... 17
2.3.6
Cylinder thermostat .............................................................................. 18
2.3.7
Automatic bypass valve ........................................................................ 18
2.3.8
Boiler interlock ...................................................................................... 19
2.3.9
Pump .................................................................................................... 19
2.3.10
2.4
Heat emitters ..................................................................................... 20
Central heating controls in the UK homes ................................................... 23
2.4.1
Regulations for central heating controls ............................................... 23
2.4.2
Space heating controls in existing homes............................................. 28
2.5
Impacts of space heating controls on energy demand ................................ 31
2.5.1
Conventional space heating controls .................................................... 32
2.5.2
Occupancy based space heating control .............................................. 39
V
2.6
Zonal space heating control ........................................................................ 42
2.7
Modelling domestic energy use ................................................................... 46
2.7.1
Steady state models ............................................................................. 47
2.7.2
Dynamic models ................................................................................... 49
2.7.3
Model calibration and validation ........................................................... 51
2.8
3
Overview of the methodology and test houses ................................................. 56
3.1
Introduction ................................................................................................. 56
3.2
Overview of the methodology ...................................................................... 56
3.2.1
Overview of the space heating trials ..................................................... 56
3.2.2
Overview of the dynamic thermal modelling ......................................... 58
3.2.3
Overview of the wider scale evaluation................................................. 59
3.3
5
Test houses................................................................................................. 60
3.3.1
Description of the test houses .............................................................. 60
3.3.2
Building geometry ................................................................................. 63
3.3.3
Construction materials and properties .................................................. 65
3.3.4
Heating system ..................................................................................... 68
3.3.5
Synthetic occupancy ............................................................................. 68
3.3.6
Experimental characterisation of the test houses ................................. 77
3.4
4
Summary ..................................................................................................... 54
Summary ..................................................................................................... 84
Space heating trials........................................................................................... 85
4.1
Introduction ................................................................................................. 85
4.2
Instrumentation ........................................................................................... 85
4.3
The control strategies.................................................................................. 90
4.4
Comparison of indoor air temperatures ....................................................... 95
4.5
Heating demand, boiler efficiencies and fuel use ...................................... 101
4.6
Summary ................................................................................................... 104
Dynamic thermal modelling ............................................................................. 105
5.1
Introduction ............................................................................................... 105
5.2
Modelling the building envelope of test houses ......................................... 106
5.2.1
Building geometry ............................................................................... 106
5.2.2
Zoning................................................................................................. 108
5.2.3
Ground modelling ............................................................................... 110
VI
5.2.4
5.3
Scheduled Natural Ventilation (SNV) .................................................. 114
5.3.2
Air Flow Network (AFN) ...................................................................... 118
Modelling the heating systems .................................................................. 125
5.4.1
Modelling the heating system for the co-heating test.......................... 125
5.4.2
Modelling the heating systems for the space heating trials................. 126
5.5
Modelling the occupancy ........................................................................... 129
5.6
Weather file Construction .......................................................................... 130
5.7
Summary ................................................................................................... 133
Comparison of the DTM predictions and measurements: DTM calibration ..... 134
6.1
Introduction ............................................................................................... 134
6.2
Comparison for the co-heating test ........................................................... 134
6.2.1
Model with SNV .................................................................................. 134
6.2.2
Model with AFN .................................................................................. 141
6.3
7
Modelling the air flow ................................................................................ 114
5.3.1
5.4
6
Construction materials and properties ................................................ 111
Comparison for the Heating Trial 1 ........................................................... 145
6.3.1
Comparison of the energy demands................................................... 145
6.3.2
Comparison of the indoor air temperatures ........................................ 148
6.4
Model calibration ....................................................................................... 148
6.5
Summary ................................................................................................... 173
Potential savings in other UK locations and better insulated houses .............. 174
7.1
Introduction ............................................................................................... 174
7.2
Evaluation of the empirical results for different UK locations .................... 174
7.2.1
Annual heating fuel and cost savings in different UK locations........... 174
7.2.2
Relationship between measured gas use and weather conditions ..... 175
7.2.3
Effect of different UK locations ........................................................... 177
7.3 Evaluation of the DTM results for different UK locations and comparison with
empirical evaluation ............................................................................................ 181
8
7.4
Implications for better insulated homes ..................................................... 185
7.5
Summary ................................................................................................... 188
Discussion and future work ............................................................................. 190
8.1
Introduction ............................................................................................... 190
8.2
Measuring the energy savings potential of ZC in a UK home ................... 191
8.3
Dynamic thermal modelling and calibration of a UK home with ZC ........... 194
VII
9
8.4
Predicting the energy savings potential of ZC in different UK houses ....... 200
8.5
Summary ................................................................................................... 202
Conclusions .................................................................................................... 204
9.1
Introduction ............................................................................................... 204
9.2
Measuring the energy savings potential of ZC in a UK home ................... 204
9.3
Dynamic thermal modelling and calibration of a UK home with ZC ........... 205
9.4
Predicting the energy savings potential of ZC in different UK houses ....... 206
9.5
Overall conclusions and recommendations for future work ....................... 207
References ............................................................................................................. 209
A.1
Appendix 1: Blower door test reports ........................................................ 232
A.1.1
House 1 .............................................................................................. 232
A.1.2
House 2 .............................................................................................. 233
A.2.
Appendix 2: EMS Code for boiler control ............................................... 234
VIII
List of Figures
Figure 1-1: Contribution of different sectors to the UK’s total carbon dioxide
emissions of 2011(DECC, 2012a) .............................................................................. 2
Figure 1-2: Domestic final energy consumption by end use since 1970 (DECC,
2011a) ........................................................................................................................ 3
Figure 1-3: Space heating energy savings due to better insulation and heating
systems efficiency in UK homes from 1970 to 2006 (DECC, 2011a) ......................... 4
Figure 1-4: Household energy use for space heating and its share of all household
energy use for the UK (Palmer & Cooper, 2011)........................................................ 5
Figure 2-1: A standard domestic wet central heating system configuration (BRECSU,
2001) ........................................................................................................................ 11
Figure 2-2: Efficiency of condensing boilers (Oughton and Hodkinson, 2008) ......... 13
Figure 2-3: Two older thermostat designs with slider bars and analogue display on
the left compared to two state of the art programmable thermostats with LCD or full
touch screen on the right (Peffer, Pritoni, Meier, et al., 2011) .................................. 15
Figure 2-4: Left: Manual on/off radiator valve. Right: Thermostatic Radiator Valve
(TRV) (Munton, Wright, Mallaburn, et al., 2014)....................................................... 16
Figure 2-5: Principle components of a Thermostatic Radiator Valve (TRV) (BSI, 1999)
................................................................................................................................. 17
Figure 2-6: An example of two-port (on the left) and three-port (on the right)
motorized valve (Danfoss, no date) .......................................................................... 18
Figure 2-7: Pressure/flow diagram of a typical domestic three-speed central heating
pump (Mitchell, 2008) ............................................................................................... 20
Figure 2-8: Three common types of heat emitters: panel radiator, fan convector and
underfloor heating coils (Young et al. 2013) ............................................................. 21
Figure 2-9: Example layout for new systems to ensure compliance with the 2010
Building Regulations Part L1A (TACMA, 2010) ........................................................ 26
Figure 2-10: Example layout for replacement boilers to ensure compliance with the
2010 Building Regulations Part L1B (TACMA, 2010) ............................................... 27
Figure 2-11: Percentage of UK households with a boiler with each of the main
heating control types as reported in Munton et al. (2014) ........................................ 29
Figure 2-12: Average air temperature in a building with TRV controlled radiators
(Liao et al., 2005) ..................................................................................................... 36
IX
Figure 2-13: Honeywell’s Evohome system components including PTRV and central
controller .................................................................................................................. 45
Figure 2-14: A number of PTRVs from different manufacturers: from left to right:
Honeywell HR90 (Honeywell, 2014), Salus PH60C (Salus controls, 2013) and
Eurotronic Sparmatic Comet (Eurotronic, 2011)....................................................... 46
Figure 3-1: Bird’s-eye view of the test houses, their surrounding buildings and
vegetation (Google Maps, 2015) .............................................................................. 60
Figure 3-2: Views of the two test houses: front, south-facing (left) and back, northfacing (right) ............................................................................................................. 61
Figure 3-3: The West facing windows in the House 1 which were covered by 50 mm
PIR insulation boards ............................................................................................... 62
Figure 3-4: Blocked original open fire places located in the living room of House 1. 63
Figure 3-5: The floor plans of the test houses with the floor area of each room ....... 64
Figure 3-6: Floor plan of the ventilated subfloors existed below the ground floor of
each house and the location of air bricks ................................................................. 66
Figure 3-7: Examples of test house inspections for understanding details of
construction materials: (a) construction of internal floors; (b) loft (attic) space
construction (before removing debris) ...................................................................... 67
Figure 3-8: Borescope investigation at the test houses: (a) exploring external wall
cavity; (b) exploring subfloor construction through air bricks (Photos by Stephen
Porritt) ...................................................................................................................... 67
Figure 3-9: Z-wave smart home controller used in each house for synthetic
occupancy during the HT1 and HT2 ......................................................................... 71
Figure 3-10: light bulbs with different outputs used to produce the internal heat gains
in the living room of House 2 .................................................................................... 74
Figure 3-11: Total actual heat gains in different rooms of a house during a weekday
................................................................................................................................. 75
Figure 3-12: Internal door operation mechanism used in the test houses ................ 76
Figure 3-13: IP camera which was used in the living room of a test house to check
the operation of synthetic occupancy devices .......................................................... 77
Figure 3-14: The blower door tests set up during the test in House 1 ...................... 78
Figure 3-15: The location of fan heaters and circulation fans during the co-heating
test ........................................................................................................................... 80
Figure 3-16: The co-heating test set up in the living room of House 1 ..................... 81
X
Figure 3-17: Siviour regression analysis for the two test houses ............................. 82
Figure 4-1: Equipment used to measure boiler heat output in the test houses;
consisted of flow meter, temperature sensors and energy integrator ....................... 86
Figure 4-2: Equipment used to measure and record volume of gas use in the test
houses...................................................................................................................... 87
Figure 4-3: The location of temperature sensor used to measure outdoor air
temperature and its shielding ................................................................................... 88
Figure 4-4: The calibration of U type thermistors using water bath calibrator........... 89
Figure 4-5: Test house schematic plans with heating systems and environmental
monitoring equipment as configured during heating trial 1, for heating trial 2 the
PTRVs with their central controller were swapped with TRVs in the opposite house92
Figure 4-6: A PTRV installed on a radiator (on the left) and the interface of the
central controller used to programme the PTRVs (on the right) ............................... 94
Figure 4-7: Air and radiator surface temperature variations in different rooms: heating
trial 1, 21st Feb 2014, ZC in House 1, CC in House 2. ............................................. 97
Figure 4-8: Measured daily heat output from the boilers during the heating trials 1
and 2 and their error bars (based on heat meter’s manufacturer stated accuracy)
together with the average daily outdoor temperature ............................................. 101
Figure 4-9: Daily efficiency of the boilers with zonal control (ZC) and conventional
control (CC) in each heating trial with their error bars together with the daily average
outdoor temperature ............................................................................................... 102
Figure 5-1: Views of the LMP1930 test house model in DesignBuilder: front, southfacing (left) and back, north-facing (right) ............................................................... 106
Figure 5-2 View of the LMP1930 test house model with the effect of shading from the
neighbour blocks (15 March at 16:00) .................................................................... 107
Figure 5-3: Window geometry definition in DesignBuilder and EnergyPlus
(DesignBuilder, 2014)............................................................................................. 108
Figure 5-4: LMP1930 test house model zoning strategy for ground floor and first floor
............................................................................................................................... 109
Figure 5-5: definition of surfaces in determining wind pressure coefficients (CIBSE,
2006) ...................................................................................................................... 122
Figure 6-1: Whole house hourly electricity consumption measured in House 1 and 2
compared with the model prediction along with the hourly outdoor air temperature
(SNV) ..................................................................................................................... 136
XI
Figure 6-2: Whole house hourly electricity consumption measured in House 1 and 2
compared with the model prediction along with the hourly global horizontal solar
radiation (SNV) ....................................................................................................... 137
Figure 6-3: Whole house hourly electricity consumption measured in House 1 and 2
compared with the model prediction along with the hourly wind speed (SNV) ....... 138
Figure 6-4: Measured and predicted whole house daily electricity consumption in
House 1 and 2 during the co-heating test (SNV) .................................................... 140
Figure 6-5: Schematic of the pressure distribution and the air flows in the LMP1930
test houses during the co-heating test.................................................................... 143
Figure 6-6: Whole house hourly electricity consumption measured in House 1 and 2
compared with the model prediction along with the hourly outdoor air temperature
(AFN)...................................................................................................................... 144
Figure 6-7: Measured and predicted daily boiler heat output during Heating Trial 1 in
house with ZC ........................................................................................................ 146
Figure 6-8: Measured and predicted daily boiler heat output during Heating Trial 1 in
house with CC ........................................................................................................ 147
Figure 6-9: predicted and measured indoor air temperatures of the house with ZC
along with measured outdoor air temperatures and global horizontal radiation; 27
Feb to 1 March 2014 .............................................................................................. 151
Figure 6-10: predicted and measured indoor air temperatures of the house with CC
along with measured outdoor air temperatures and global horizontal radiation; 27
Feb to 1 March 2014 .............................................................................................. 153
Figure 6-11: Boiler and its uninsulated pipe work and the position of temperature
sensor on a tripod in the kitchen of house 2 ........................................................... 156
Figure 6-12: predicted and measured mean air temperature over a single day for the
IEA test room with radiator for (a) study by Zhai and Chen (2005) and (b) study by
Beausoleil-Morrison (2000) (Figure was reproduced from Zhai and Chen (2005)). 159
Figure 6-13: predicted daily boiler heat output against measured boiler heat output
for the 28 days of HT: ZC ....................................................................................... 167
Figure 6-14: predicted daily boiler heat output against measured boiler heat output
for the 28 days of HT: CC....................................................................................... 167
Figure 6-15: Indoor air temperatures measured and predicted by the refined model
for the ZC house along with measured outdoor air temperatures; 27 Feb to 1 March
2014 ....................................................................................................................... 170
XII
Figure 6-16: Indoor air temperatures measured and predicted by the refined model
for the CC house along with measured outdoor air temperatures; 27 Feb to 1 March
2014 ....................................................................................................................... 172
Figure 7-1: Weekly gas consumption of the houses with ZC and CC against weekly
average outdoor air temperature for 8 weeks of monitoring, best fit lines and 95%
confidence intervals................................................................................................ 176
Figure 7-2: Measured weekly gas consumption plotted against calculated weekly
HDD for the houses with ZC and CC...................................................................... 177
Figure 7-3: projected residential gas prices between 2014 and 2028 (DECC, 2012b)
............................................................................................................................... 178
Figure A-1: Blower door test report for House 1 ..................................................... 232
Figure A-2: Blower door test report for House 2 ..................................................... 233
XIII
List of Tables
Table 2-1: Census 2011 data for domestic heating systems in England and Wales
(Office for National Statistics, 2011) ......................................................................... 10
Table 2-2: The ranges of heat outputs and heights of different types of radiators
according to a manufacturer (BSMW products Ltd, 2011)........................................ 22
Table 2-3: Proportion of dwelling types reporting primary heating controls
(reproduced from Munton et al. (2014)).................................................................... 30
Table 2-4: Typical average annual fuel and cost savings (£) which could be achieved
from better heating controls (Table reproduced from Good Practice Guide 302
(BRECSU, 2001)) ..................................................................................................... 33
Table 2-5: Studies which compared energy demand and heating practices in homes
with a programmable room thermostat and homes with a room thermostat and their
main findings (Wei et al. 2014) ................................................................................. 38
Table 2-6: A number of systems currently available in the UK market and their prices
(as in 24 February 2015) for a configuration which can apply zonal space heating
control in a typical UK house .................................................................................... 44
Table 3-1: Summary of the DTMs created during the modelling campaign .............. 59
Table 3-2: Summary of construction elements of the test houses, their areas and
calculated U-values according to RdSAP (BRE, 2014) ............................................ 65
Table 3-3: Rated capacities of the radiators in the LMP1930 test houses according to
their manufacturer’s data for 50K temperature difference ........................................ 68
Table 3-4: Weekday and weekend ‘occupied’ hours of each room .......................... 70
Table 3-5: The timing and magnitude of internal heat gains presented in different
rooms of both houses during each trial .................................................................... 72
Table 3-6: Number of heat emitters and their nominal outputs used to deliver internal
heat gains in each room ........................................................................................... 73
Table 3-7: Summary of the house characterisation test results................................ 83
Table 4-1: Accuracy of the equipment and uncertainty in values used .................... 90
Table 4-2: Weekday and weekend ‘occupied’ hours with the number of hours each
room was heated to the set-point or set-back temperatures and, for ZC, the PTRV
set-point and set-back temperatures, and for CC, the TRV position ........................ 95
Table 4-3: Average indoor air temperatures in each room during five different periods,
and the spatially averaged whole house temperature .............................................. 98
XIV
Table 4-4: Summary of daily average boiler efficiencies in each heating trial and
overall efficiency ..................................................................................................... 103
Table 5-1: Construction materials properties used in LMP1930 model .................. 111
Table 5-2: Construction elements of the test houses and their U-values and
thicknesses of the materials used in each construction element ............................ 112
Table 5-3: Characteristics of the glazing in LMP1930 model ................................. 113
Table 5-4: Characteristics of the blinds material chosen for the model .................. 113
Table 5-5: Shielding coefficient (e) reproduced from Table D.8 BS EN 12831 (British
Standards, 2013) .................................................................................................... 116
Table 5-6: Wind pressure coefficients over façade 1 and roof (front) for wind angles
in 45º increments based on the slope of surfaces considering normal exposure of the
site to wind and aspect ratio 1:1 (DesignBuilder, 2014) ......................................... 122
Table 5-7: Crack characteristics according to DesignBuilder’s “poor” crack template
used in the model for walls, floors and the roof ...................................................... 124
Table 5-8: Crack characteristics according to DesignBuilder’s “poor” crack template
used in the model for the doors, windows and vents .............................................. 124
Table 5-9: Measured average air temperature in different zones during the coheating test; theses temperatures were used as the set-point temperature of each
zone in the DTM when modelling the co-heating test ............................................. 126
Table 5-10: Conversion factors of cloud cover from Oktas to tenth ........................ 131
Table 5-11: Summary of hourly weather parameters, their units and sources of data
............................................................................................................................... 132
Table 6-1: MBE (%) and CVRMSE (%) calculated and their acceptable limit (coheating test with SNV) ............................................................................................ 141
Table 6-2: Zone by zone average infiltration rate and exfiltration rate for the
LMP1930 test houses calculated by AFN............................................................... 142
Table 6-3: MBE (%) and CVRMSE (%) calculated and their acceptable limit (Coheating test with AFN) ............................................................................................ 145
Table 6-4: MBE (%) and CVRMSE (%) calculated for each house and their
acceptable limit using each air flow modelling strategy .......................................... 147
Table 6-5: Weather parameters for the selected days and the whole HT1............. 149
Table 6-6: Nominal and new set-point temperatures which were applied for variant 1
............................................................................................................................... 161
XV
Table 6-7: New crack characteristics according to DesignBuilder’s “medium” crack
template used in the variant 9 for walls, floors and the roof ................................... 162
Table 6-8: New crack characteristics according to DesignBuilder’s “medium” crack
template used in the variant 9 for the doors, windows and vents ........................... 163
Table 6-9: MBE (%), CVRMSE (%) and ΔTavg (°C) calculated for each case and
each house using AFN and SNV ............................................................................ 165
Table 6-10: Comparison of MBE (%) and CVRMSE (%) and ΔTavg (°C) between the
base case model and the refined model................................................................. 166
Table 7-1: Estimated gas use for heating the test house, with the same occupancy,
in seven different regions of the UK, using either ZC or CC and, the NPV, IRR or
financial savings, for both a basic and a luxury ZC systems .................................. 180
Table 7-2: Average air temperature, wind speed and global horizontal radiation in
each region during the heating season .................................................................. 181
Table 7-3: Total annual gas use for house with ZC and CC and annual percentages
of savings by ZC in different regions of the UK predicted by DTM and Empirical
Model (EM) and their differences ........................................................................... 184
Table 7-4: NPV, IRR or financial savings for both a basic and a luxury ZC systems
calculated for seven different regions of the UK based on modelling results for the
un-furbished houses ............................................................................................... 185
Table 7-5: Thermal properties of the insulating materials used in the refurbished
model ..................................................................................................................... 186
Table 7-6: Annual gas use and percentages of savings from refurbishment for ZC
and CC houses for different regions of the UK along with percentage of savings from
ZC after refurbishment and its differences compared to the savings in un-furbished
house ..................................................................................................................... 187
Table 7-7: NPV, IRR or financial savings for both a basic and a luxury ZC system
calculated for seven different regions of the UK based on modelling results for
refurbished houses ................................................................................................. 188
XVI
Abbreviations
ACH
Air Changes per Hour
AFN
Air Flow Network
ASHRAE
American Society of Heating Refrigeration and Air-Conditioning
ATTMA
Air Tightness Testing & Measurement Association
BADC
British Atmospheric Data Centre
BI
Boiler Interlock
BRE
Building Research Establishment
BRECSU
Building Research Energy Conservation Support Unit
BREDEM
Building Research Establishment Domestic Energy Model
BSI
British Standards Institution
CC
Conventional Control
CFD
Computational Fluid Dynamics
CHeSS
Central Heating System Specifications
CIBSE
Chartered Institution of Building Services Engineers
CREST
Centre for Renewable Energy Systems Technology
CSV
Comma Separated Value
CT
Cylinder Thermostat
CVRMSE
Coefficient of Variation of Root Mean Square Error
DECC
Department of Energy and Climate Change
DEFACTO
Digital Energy Feedback and Control Technology Optimisation
XVII
DHR
Diffuse Horizontal Radiation
DHW
Domestic Hot Water
DNR
Direct Normal Radiation
DTM
Dynamic Thermal Model
EEBPp
Energy Efficiency Best Practice programme
EFUS
Energy Follow Up Survey
EM
Empirical Model
EMS
Energy Management System (EnergyPlus)
EPSRC
Engineering and Physical Sciences Research Council
EPW
EnergyPlus Weather
ERL
EnergyPlus Runtime Language
GHR
Global Horizontal Radiation
GPS
Global Positioning System
HDD
Heating Degree Days
HT
Heating Trial
HVAC
Heating, Ventilation and Air-Conditioning
IDF
Input Data File (EnergyPlus)
IEA
International Energy Agency
IP
Internet Protocol
IRR
Internal Rate of Return
IWEC
International Weather for Energy Calculations
XVIII
LMP1930
Loughborough University’s Matched Pair of 1930s houses
LPHW
Low Pressure Hot Water
LST
Low Surface Temperature
MBE
Mean Bias Error
MRT
Mean Radiant Temperature
MV
Motorised Valve
NHBC
National House Building Council
NPV
Net Present Value
ONS
Office for National Statistics
PFT
Perfluorocarbon Tracer
PID
Proportional Integral Derivative
PP
Percentage Points
PRT
Programmable Room Thermostat
PTRV
Programmable Thermostatic Radiator Valve
RdSAP
Reduced Standard Assessment Procedure
RECS
Residential Energy Consumption Survey
RT
Room Thermostat
SAP
Standard Assessment Procedure
SBEM
Simplified Building Energy Model
SEDBUK
Seasonal Efficiency of Domestic Boilers in the UK
SNV
Scheduled Natural Ventilation
XIX
TACMA
The Association of Controls Manufacturers
TRV
Thermostatic Radiator Valve
VAT
Value Added Tax
WD
Weekday
WE
Weekend
WGC
Weekly Gas Consumption
ZC
Zonal Control
XX
Energy use conversion factors
1 MJ = 0.27777777778 kWh
1 Mtoe = 11630000000 kWh
XXI
1 Introduction
1.1 Background
As a part of 2008 Climate Change Act, the UK government made a commitment to
reduce the greenhouse gas emissions by at least 80% compared to 1990 levels by
2050 (Office of Public Sector Information, 2008). The Climate Change Act which was
initially targeted to reduce emissions by 26% by 2020 was later tightened to 34%
(Office of Public Sector Information, 2009). Carbon dioxide is the main greenhouse
gas and, in 2013 accounted for 82% of total UK’s man-made greenhouse gas
emissions (DECC, 2014a). Figure 1-1 indicates the contribution of each sector to the
total UK carbon dioxide emissions (DECC, 2012a). Residential fossil fuel use has
been the third largest contributor to the UK’s total carbon dioxide emissions after the
energy supply and transport sectors and accounts for 15% of the total carbon dioxide
emissions (DECC, 2012a). However, the share reported for this sector does not
even include the emissions from the energy supply sector due to generating
electricity for domestic use. Considering the energy supply as well, housing is
responsible for 25% of the UK’s greenhouse gas emissions and therefore it would be
difficult to meet the 2050 target without reducing emissions from residential buildings
(Palmer & Cooper, 2011). Moreover, reduction in residential fossil fuel use is crucial
for the UK’s energy security so that the UK could become less dependent on imports.
1
4%
Energy Supply
15%
40%
Transport
Residential fossil fuel use
15%
Business
Other
26%
Figure 1-1: Contribution of different sectors to the UK’s total carbon dioxide
emissions of 2011(DECC, 2012a)
Achieving the Climate Change Act targets will require substantial reductions in
energy consumption in different sectors; though reductions in the domestic sector
are considered to be “relatively low cost” and “realistically achievable” (Committee on
Climate Change, 2008). Since 1990, emissions from fossil fuel use in the residential
sector have fluctuated but in 2010 they were 8% above the 1990 level (DECC,
2011c). In 2010, the UK residential sector emissions of carbon dioxide increased by
13.4% compared to the previous year (the highest rise for any single sector) due to a
considerable rise in residential gas use for space heating as 2010 was on average
the coldest year since 1986 (DECC, 2011c). In 2013, the emissions from this sector
were estimated to be 3% below the 1990 level (DECC, 2014a).
The UK’s housing stock is one of the oldest and least efficient in Europe (Boardman,
Killip, Darby, et al., 2005). The majority of energy consumption in UK dwellings is
due to space heating which in 2009 accounted for 61% of the total energy
consumption in the domestic sector (DECC, 2011a). Figure 1-2 presents the
domestic final energy consumption in UK by end use since 1970 in which space
heating has been continuously dominant. Therefore as Shipworth et al. (2010)
argues “Any policies and initiatives aimed at significantly reducing residential CO2
emissions must address the largest residential CO2 emitter – central heating”.
2
Figure 1-2: Domestic final energy consumption by end use since 1970 (DECC,
2011a) 1
Improvements in insulation and heating efficiency have saved a considerable
amount of heat energy. Figure 1-3 shows that from 1970 to 2006 improvement in the
efficiency of domestic heating systems and implementing different types of insulation
such as loft (attic), cavity and hot water tank insulation and double glazing kept the
current level of space heating energy consumption to almost half of the amount that
it could have been without these improvements.
1
For conversion of Mtoe to kWh see energy conversion factors, p XXI.
3
Figure 1-3: Space heating energy savings due to better insulation and heating
systems efficiency in UK homes from 1970 to 2006 (DECC, 2011a)
Figure 1-4 shows the energy used for space heating and it’s share in total household
energy use since 1970 for the UK (Palmer & Cooper, 2011). It indicates that despite
the energy efficiency improvements in houses, heating’s share of household energy
use has increased from 58% in 1970 to 66% in 2007. During this period, the
proportion of dwellings with central heating has increased from less than a third to
96%. This increase in heating’s share of domestic energy use is despite the fact that
the amount of electric equipment in homes has significantly increased and also that
gas central heating systems are generally more efficient than individual room
appliances such as open coal fires and are therefore expected to use less energy
(Utley & Shorrock, 2008).
4
Figure 1-4: Household energy use for space heating and its share of all household
energy use for the UK (Palmer & Cooper, 2011)
The rise of central heating has considerably increased the domestic energy use.
According to Andrews et al. (2012), central heating contributed to 30% increase in
energy consumption between 1970 to 2010. This is because it allows people to heat
the whole of their homes rather than just individual rooms and provides expectations
of higher indoor temperatures throughout the house. Hunt & Gidman (1982),
recorded spot measurements of room temperatures in 1000 homes in the UK during
the February and March of 1978 and found that the average temperature in centrally
heated homes was 3°C higher than the homes without central heating.
It is likely that a considerable amount of energy is still being wasted in centrally
heated homes and there is huge potential for further savings via better control
strategies. An example of this waste would be heating all the rooms to maintain the
same temperature even when they are unoccupied. Research has shown that an
average centrally heated home consumes about twice as much energy for space
heating as a similar home with heating only in the living room (Palmer & Cooper,
2011). Utley & Shorrock, (2008) argue that this would be even higher for a house
with poor levels of insulation while in a very well insulated house, it may be only
necessary to have a simple system of one or two room heaters instead of a full
central heating system.
5
Zonal Control of space heating (ZC) is one option when considering more efficient
heating control strategies. ZC could be described simply as restricting the heating of
unoccupied areas of the home in order to reduce wasted energy. For example,
during the day time, when the bedrooms are unoccupied, only the living room could
be heated while the first floor bedrooms would be a separate zone and only heated
during the evening when they are occupied and often to a lower temperature
compared to main living areas. Therefore, ZC could be potentially more energy
efficient as it enables the householders to match their space heating to their lifestyles.
1.2 Justification of the research
Wireless technology and the availability of more powerful batteries have led control
manufacturers to develop retrofit systems for zonal space heating. Although market
deployment is in its infancy, this is a rapidly developing area with many new systems
emerging. The main components of ZC systems are the battery-operated
Programmable Thermostatic Radiator Valves (PTRV) which replace normal TRVs
and have motorised valves to regulate the hot water flow through the radiators
according to a set-point temperature and time schedule. These can be set on the
PTRVs themselves, via a central controller which communicates wirelessly with the
PTRVs, or even remotely via a mobile phone or computer.
There has been little (if any) research to quantify how much energy can be saved
using these devices. These savings are likely to be dependent on house type, size,
location and occupancy pattern. Therefore, this research was conducted to answer
the following questions:
•
How much energy could ZC save in a UK house?
•
Does the effectiveness of ZC depend on the local climate or its level of
insulation?
1.3 Aim and objectives
The aim of this thesis was to quantify the energy demand reduction potential of using
zonal space heating control in a UK home and the implication of this at a wider scale.
This was achieved through the following objectives:
6
1. Design and set up an experimental investigation and measure the energy
savings of zonal space heating control compared to conventional control in
a real house.
2. Predict the energy savings for the same house using a Dynamic Thermal
Model (DTM) calibrated using measurements from the experimental
investigation.
3. Use the experimental results and the calibrated DTM to explore how
savings might vary in UK houses exposed to a different climate or higher
levels of insulation.
1.4 Outline of the thesis
•
Chapter 2 presents a thorough literature review which was conducted for this
study. This covers space heating methods in the UK with a focus on wet
central heating systems and their controls; studies investigated the impacts of
space heating controls on energy use; zonal space heating control systems;
and existing literature on modelling energy use in the domestic sector.
•
Chapter 3 provides an overview of the methodology and describes the test
houses used in this study and their characterisation tests including co-heating
and airtightness tests.
•
Chapter 4 describes the space heating trials conducted in the test houses in
order to measure the energy saving potential of ZC and presents the results of
the trials.
•
Chapter 5 describes the construction of dynamic thermal models (DTMs) of
the test houses for the purpose of modelling the space heating trials and the
co-heating test.
•
Chapter 6 compares the results from the DTMs with the measured results
from the tests. The chapter also describes the processes of calibration and
validation of the DTMs based on these comparisons.
•
Chapter 7 firstly describes the development of an empirical model based on
results from the space heating trials to predict the annual energy and cost
7
savings of ZC in UK homes located in different regions. The differences
between the predictions of the empirical model and the calibrated DTM are
investigated and potential reasons for the differences identified. The calibrated
DTM is then used to predict the likely heating energy and cost savings in
better insulated homes.
•
Chapter 8 discusses the findings from chapters 3 to 7, identifies the key
messages from the research and makes suggestions for future work.
•
Chapter 9 presents the conclusions from the research.
8
2 Literature review
2.1 Introduction
This chapter presents the context for the study of zonal space heating control (ZC) in
UK homes and reviews the relevant academic, governmental and industry based
literature. Firstly, it describes different space heating methods in UK homes
(section 2.2) and the configuration and components of the most common system
which is currently being used: wet central heating (section 2.3). The literature review
then discusses the space heating controls in existing UK homes as well as the
regulations for new homes (section 2.4). In section 2.5, studies which had examined
the impacts of conventional and occupancy based space heating controls are
critically reviewed. Section 2.6 describes ZC and explores different ZC systems
currently available in the UK market. Section 2.7 introduces different techniques and
tools which are being used to model domestic energy use and discusses model
calibration and validation techniques. Finally, section 2.8 summarizes the findings
from literature review which have direct implications on the methods chosen for this
study.
2.2 Space heating methods in the UK homes
Next to food, heating has been among the most important elements in human
existence (Wright, 1964). Since the first fire was lit in a cave, heating the living
spaces to increase thermal comfort has been associated with the life of most people
especially those living in the colder climates. In the UK, homes were commonly
being heated using coal open fires from as early as the 17th Century well into the
1960s (Roberts, 2008 and Wright, 1964). The low pressure gravity hot water heating
was common by 1900 but only limited to larger buildings and the heating in the
middle- and lower-priced homes were “unplanned” and “almost unknown” (Doherty,
1967). Early central heating systems were heated by back boilers situated behind
the grate of open fireplaces which were only able to heat a few radiators (Beattie,
1966). Back-boilers were simple and reliable and a large number of them were
installed in the 1980s but they had low efficiencies (Munton, Wright, Mallaburn, et al.,
2014).
9
In the 1970s, with the introduction of North Sea Gas to the UK, gas fired central
heating evolved (Roberts, 2008). This was a breakthrough into domestic space
heating as, before this, particular rooms were heated when needed but now all the
rooms could be heated, regardless of their occupancy.
Currently, central heating is the main method for space heating in the UK homes.
According to the 2011 Census (Office for National Statistics, 2011), there are more
than 23 million households with at least one usual resident in England and Wales
from which 97.3% of them have one or more types of central heating (Table 2-1).
Domestic central heating systems can be fuelled by mains gas, Liquefied Petroleum
Gas (LPG), oil, electricity or solid fuel. However, the majority of homes in England
and Wales (78.7%) have central heating which is supplied by gas (Table 2-1).
Table 2-1: Census 2011 data for domestic heating systems in England and Wales
(Office for National Statistics, 2011)
Total number of households with at least one resident
23,366,044
Percentage of households with no central heating
2.7%
Percentage of households with Gas central heating
78.7%
Percentage of households with electric central heating
(including storage heaters)
8.1%
Percentage of households with oil central heating
4.1%
Solid fuel central heating (including wood and coal)
0.7%
Other central heating
(including solar, LPG or other bottled gas)
Percentage of households with two or more types of central
heating
1.6%
4.1%
Central heating systems generally fall into 3 main categories: wet (hydronic) systems
with heated water circulating through radiators, convectors or under-floor heating;
warm air systems in which the air is delivered through ducts to rooms using a heat
exchanger with a fan and filter (Doherty, 1967); and electric storage and panel
systems using off peak and on peak electricity.
Wet systems are by far the most common type of heating system in the UK homes
(The Carbon Trust, 2011).
10
2.3 Wet central heating system components and
configuration
A standard domestic wet central heating system typically consists of the following
components (Figure 2-1):
•
Boiler
•
Time switch/programmer
•
Room thermostat / Programmable room thermostat
•
Thermostatic Radiator Valves (TRV)
•
Motorised valve
•
Cylinder thermostat (only in systems with regular boiler)
•
Automatic bypass valve
•
pump
•
Heat emitters
Figure 2-1: A standard domestic wet central heating system configuration (BRECSU,
2001)
11
2.3.1 Boiler
Boilers can be described as ‘regular’ or ‘combination’ (combi). A regular boiler is not
able to provide DHW directly; therefore it does so indirectly via a separate hot water
store (Figure 2-1). Historically, these were the most common boiler type and are
often referred to as conventional or traditional boilers (BRECSU, 2000). A
combination boiler has the capability to provide DHW directly, and some models
contain a small internal hot water store. Combination boilers can be often more
efficient as the standing losses from the hot water tank will be avoided (Munton,
Wright, Mallaburn, et al., 2014). Both regular and combination boilers may either be
condensing or non- condensing. Condensing boilers use the heat from the flue
gasses as secondary heating to heat the water in addition to direct heat transfer via
burning fuel (Hall & Greeno, 2013).They also have a larger heat transfer surface
area compared to non-condensing boilers (Hall & Greeno, 2013).
Condensing boilers are generally more efficient with an overall efficiency of above 90%
compared to the non-condensing boilers with an expected efficiency of 75% (Hall&
Greeno, 2013). The element that defines the efficiency of the condensing boilers in
operation is the temperature at which the water returns to the boiler (Oughton and
Hodkinson, 2008). High efficiency for the condensing boilers would be achieved with
a water returning at a low temperature (Figure 2-2) (Oughton and Hodkinson, 2008).
12
Figure 2-2: Efficiency of condensing boilers (Oughton and Hodkinson, 2008)
Condensing boilers have become mandatory for new and replacement boilers since
2005 according to the UK Building Regulations (The Office of the Deputy Prime
Minister, 2005). The percentage of dwellings with condensing boilers and in
particular condensing combination boilers has increased to about a third of the UK
housing stock in 2010 (Department for Communities and Local Government, 2012).
According to the government’s Energy Efficiency Best Practice programme (EEBPp),
the boiler is one of the main factors influencing energy efficiency of domestic central
heating systems (BRECSU, 2000). The Seasonal Efficiency of a Domestic Boiler in
the UK (SEDBUK) database records the efficiency of boilers which has been
measured in a laboratory.
Internal control of boilers is typically according to the water temperature flowing from
the boiler. Based on the set-point and deadband 2 two temperature threshold for CutIn and Cut-Out can be determined. If the water temperature is higher than Cut-Out,
the boiler is switched off. If the water temperature is lower than the Cut-In, the boiler
is switched on (Liao, Swainson & Dexter, 2005). The water set-point temperature
2
Deadband here means a temperature range that is set around the set-point temperature to avoid
excessive hunting by the controller (Race, 2005)
13
can be fixed, varied based on external air temperature or varied based on heating
load. Liao et al. (2005) discussed that the overall performance of a heating system is
considerably affected by the scheme for determining the value of water temperature
set-point.
2.3.2 Time switch / programmer
A time switch or programmer is the primary way in which the central heating system
can be controlled by the occupants. It allows them to set the times at which the
system will turn on and turn off. A time switch is an electrical switch operated by a
clock to switch only one circuit and therefore control either space heating or hot
water, but not both (Energy Saving Trust, 2008a). A programmer can switch two
circuits (heating and DHW). Depending on the type of programmer (i.e. mini,
standard or full programmer) the heating and DHW time setting can be the same or
fully independent (BRECSU, 2001). A mini programmer allows space heating and
hot water to be on together or hot water alone but not heating alone. A standard
programmer uses the same time setting for space heating and hot water. A full
programmer allows the time setting for space heating and hot water to be fully
independent (BRECSU, 2001).
2.3.3 Room thermostat / Programmable room thermostat
A room thermostat allows the occupants to control the central heating system by
limiting the air temperature when the heating is on. It measures the air temperature,
is often located in a central area of the home such as a living room or hallway and
switches the space heating off when the temperature is above a single target
temperature set by the user (set-point temperature) (Energy Saving Trust, 2008a).
Building services handbook (Hall & Greeno, 2013) suggests that the thermostat
should be installed somewhere away from draughts, direct sunlight and heat emitters
at 1.2 to 1.5 m above the floor level.
A Programmable Room Thermostat (PRT) is a combined time switch and room
thermostat that enables the user to set different periods with different set-point
temperatures for space heating, usually in a daily or weekly cycle (Energy Saving
Trust, 2008a). The use of programmable thermostats was included in the US
Environmental Protection Agency’s EnergyStar Programme in 1995, suggesting that
14
using them the households could save around $180 a year (Meier et al, 2012).
However, programmable thermostats have not been widely used in the UK as it was
believed that they are not necessary considering the milder climate of the UK
(Munton, Wright, Mallaburn, et al., 2014).
During 1990s, the ability of the PRTs to set different temperatures throughout a day
and heating schedules for weekday/ weekends considerably improved (Peffer,
Pritoni, Meier, et al., 2011). Moreover, mobile phones and internet technology have
been developed quickly so that a number of remotely controlled thermostats which
allow occupants to remotely control their heating system are now available from
different manufacturers. Global Positioning System (GPS) in mobile phones allows
the proximity of occupants to home to be identified and transferred to the thermostat
which then can predict arrival times and ensure that the heating is turned on when
the resident is coming home (Consumer focus, 2012). The interface can be remote
via web or smart phone, a large full colour LCD, touch screen or even voice
controlled (Peffer, Pritoni, Meier, et al., 2011).
Figure 2-3: Two older thermostat designs with slider bars and analogue display on
the left compared to two state of the art programmable thermostats with LCD or full
touch screen on the right (Peffer, Pritoni, Meier, et al., 2011)
2.3.4 Thermostatic Radiator Valves (TRVs)
Thermostatic radiator valves (TRVs) are used to provide a degree of temperature
control in individual rooms by adjusting the water flow through an emitter and
controlling its heat output (BRECSU, 2001). TRVs are two-port throttling valves
which can be installed in either the flow or return connection of radiators and are
self-acting and require no external source of power (Figure 2-4) (CIBSE, 2009).
15
Head of a TRV contains a liquid or wax-filled capsule which expands or contract with
the changes in room air temperature (CIBSE, 2009). The expansion of the liquid or
wax filled capsule causes the valve seating to be depressed or elevated and
consequently regulates the flow of hot water in the radiator (Watkins, 2011). TRVs
are manually set at different levels (commonly 1 to 5 including a frost protection level
or 1 to 6) using their temperature selector scale to define a separate target
temperature for each room (Figure 2-4). A temperature setting range is often
available from the manufacturer’s technical data. For example for one of the
products (Drayton RT212) the temperature setting range for levels 1 to 6 was given
between 12 °C to 29 °C (Invensys Controls, no date).
Figure 2-4: Left: Manual on/off radiator valve. Right: Thermostatic Radiator Valve
(TRV) (Munton, Wright, Mallaburn, et al., 2014)
Figure 2-5 shows the components of a TRV with an integral temperature sensor
which means the sensor, transmission unit and temperature selector constitute an
assembly which is incorporated with the valve body assembly (BSI, 2006). This type
of assembly would allow the TRVs to be fitted as direct replacements for manual
on/off radiator valves (CIBSE, 2009).
The accuracy of temperature control achieved by the TRVs is dependent on the
ability of its temperature sensor to sense the real temperature of the room (Watkins,
2011). According to BS7478 (1999), there is a relationship between the temperature
at the thermostatic head assembly and the temperature at the centre of a room
16
which varies between different installations. TRV head which contains its
temperature sensor should be positioned according to manufacturer’s
recommendation to ensure that the sensor is properly exposed to the room
temperature rather than the heat from the radiator (CIBSE, 2009). Since the integral
sensor is very close to the radiator, sometimes the sensor is inevitably affected from
the convective heat flows (Weker & Mineur, 1980). Therefore, in some TRVs the
sensor or both the sensor and temperature selector unit is mounted remotely from
the valve body (BSI, 2006).
Figure 2-5: Principle components of a Thermostatic Radiator Valve (TRV) (BSI, 1999)
2.3.5 Motorised valve
Motorised valves are used to control the water flow from the boiler to heating and hot
water circuits (Energy Saving Trust, 2008b). Motorised valves could be either twoport or three-port (Figure 2-6) and their selection for each system is according to the
system’s pipework layout and preference (Energy Saving Trust, 2008b).
17
Figure 2-6: An example of two-port (on the left) and three-port (on the right)
motorized valve (Danfoss, no date)
A two-port valve controls water flow to one circuit while a three-port valve controls
flow to two circuits (BRECSU, 2001).
When there is only one heating zone, a three-port valve can provide separate
heating and hot water circuits. Most three-port valves are mid-position valves which
means that they have one central inlet port connected to the flow from the boiler and
two outlet ports; one for DHW and one for central heating (BRECSU, 2001). When
there is more than one heating zone as well as hot water zone, a two-port valve for
each heating circuit is required (Energy Saving Trust, 2008b).
2.3.6 Cylinder thermostat
Cylinder thermostats are only used in the systems with a regular boiler and a hot
water tank, as opposed to systems with a combination boiler where hot water is
instantly produced (Consumer focus, 2012). A cylinder thermostat which is often
strapped to the DHW cylinder, measures the temperature of hot water cylinder and
switches the hot water supply on and off using a motorized valve (BRECSU, 2001).
A single target temperature can be set by the user or a combined time switch and
cylinder thermostat can be used to set different period with different target
temperatures for the stored hot water (Energy Saving Trust, 2008b).
2.3.7 Automatic bypass valve
A bypass circuit must be installed if the boiler manufacturer requires one, or specifies
that a minimum flow rate has to be maintained while the boiler is firing (Hall and
18
Greeno, 2013). The bypass circuit must then include an automatic bypass valve
installed between the boiler flow and return considering the direction of the flow
(Energy Saving Trust, 2008b). Automatic bypass valves are necessary when more
than half of the radiators are fitted with TRVs (BRECSU, 2001) as when the TRVs
begin to close the bypass valves opens to maintain a steady flow of water through
the boiler (Hall and Greeno, 2013). Alternatively fixed bypass can be achieved either
by ensuring that one radiator stays open or by adding a short pipe with a fixed
position valve between the flow and return pipe (BRE, 2014). A radiator without a
TRV or hand valve is a common form of fixed bypass (BRE, 2014).
2.3.8 Boiler interlock
Boiler interlock is not a control device but a wiring arrangement of the system
controls (room thermostats, PRTs, cylinder thermostats, programmers and time
switches) in order to prevent the boiler from firing when there is no demand for heat
(Energy Saving Trust, 2008a). For the systems with a regular boiler, the interlock is
usually set so that the room or cylinder thermostat switches the power supply to the
boiler through the motorised valve (BRECSU, 2001). For systems with a combination
boiler, interlock is usually achieved by using a room thermostat. In most cases,
interlock also applies to the pump operation (BRECSU, 2001). TRVs alone are not
sufficient for boiler interlock and needed to be installed together with a room
thermostat (Energy Saving Trust, 2008a).
2.3.9 Pump
The pumps used in domestic central heating systems are simple centrifugal pumps
(Mitchell, 2008). Domestic central heating pumps could be classified into three main
categories; fixed speed, three speed and variable speed (Mitchell, 2008). Fixed
speed pumps are the simplest type and used to be the standard for many years
(Mitchell, 2008). Three speed pumps which are the most common type currently
used have three settings which are related to three different pressure/flow diagrams
as can be seen in Figure 2-7 (Mitchell, 2008). The speed of the pump is selected
manually for the optimal operation of the system and the central heating controls
cannot usually change the pump speed (Mitchell, 2008). Having three settings would
19
enable some flexibility for adjustment to individual installations and allows for
potential changes to the system in future (Hall and Greeno, 2013).
Figure 2-7: Pressure/flow diagram of a typical domestic three-speed central heating
pump (Mitchell, 2008)
Variable speed pumps have a self-regulating output facility which responds
automatically to varying loads throughout a day in modern standard central heating
systems with thermostats, motorized zone valves and TRVs (Hall and Greeno, 2013).
2.3.10
Heat emitters
Heat emitters transfer heat from a heating system into the building spaces by
convection and radiation (Brown, 2011). A wide range of heat emitters are available
for domestic wet central heating systems including panel radiators, column radiators,
Low Surface Temperature (LST) radiators, towel rails, natural and fan convectors
and under-floor heating coils (Figure 2-8). The most common type installed in
modern housing is panel radiators which are available in a wide range of sizes and
outputs to suit different rooms (BRECSU, 2000). Radiators are often installed below
windows to counteract any cold downdraughts (Oughton & Hodkinson, 2008). As
opposed to its name, the majority of the radiator’s heat is transmitted via convection
(about 70%) rather than radiation (about 30%) (Brown, 2011). Fins are often added
to the radiators to increase their surface area in order to increase their output
20
(CIBSE, 2005). LST radiators are used where young children or elderly are at risk in
order to limit the surface temperature to 43ºC and prevent injury (HSE, 2012).
Figure 2-8: Three common types of heat emitters: panel radiator, fan convector and
underfloor heating coils (Young et al. 2013)
The heat output of radiators are dependent on a number of factors including their
size, number of panels (single or double), number of fins and their material
(Table 2-2).
21
Table 2-2: The ranges of heat outputs and heights of different types of radiators
according to a manufacturer (BSMW products Ltd, 2011)
Radiator type
Finned single panel
Finned steel
panel
Double panel, single
fin
Double panel,
double fin
Old steel panel
Height range
(W/m)
(mm)
541-1218
820-1868
483-1042
Double panel
752-1633
1
radiators
300-750
1039-2258
Single panel
Column
1
Heat output range
41-249
300-740
460-910
(Depth: 66-140)
Heat output reported in W/Section
Natural convectors are often 100% convective and consists of a copper or steel pipe
with fins fitted along its length which is installed at the bottom of the casing (Oughton
& Hodkinson, 2008). A convection airflow driven by the warm air above the
convector is moved by the cooler air entering below (Oughton & Hodkinson, 2008).
The fan convectors are similar to natural convectors but includes a fan and thus
have higher outputs compared to natural convectors (Oughton & Hodkinson, 2008).
In under-floor heating (Figure 2-8), circuits of plastic pipes laid in a floor screed or
below a timber floor are fed with low temperature hot water. In under-floor heating,
heat is typically emitted 40% convective and 60% radiative (Oughton & Hodkinson,
2008).
22
2.4 Central heating controls in the UK homes
2.4.1 Regulations for central heating controls
Since the first mandatory UK Building Regulations were introduced in 1966, this has
been revised several times over the last 40 years in order to improve the energy
efficiency of both new and existing dwellings (Boardman, Killip, Darby, et al., 2005).
Many factors such as thermal performance of building envelope, energy efficiency of
boilers and the distribution systems and control systems influence the energy
efficiency of a heating system (BRE, 2014). Central Heating System Specifications
(CHeSS) document (Energy Saving Trust, 2008a) has provided the current “Basic”
and “Best Practice” specifications for the components of domestic wet central
heating systems that are critical to energy efficiency. For example, “Basic” system
must have a regular or combination condensing boiler with minimum SEDBUK
efficiency of 86% (bands A and B) or “Best practice” system must have a regular or
combination condensing boiler with minimum SEDBUK efficiency of 90% (band A
only).
CHeSS (2008) defines “Basic” as “sufficient to comply with Building Regulations Part
L1 that came into effect in April 2002”. The building regulations apply when:
•
A home is built
•
A home has an extension or change of use
•
More than one individual component, such as a boiler is replaced in a heating
system.
CHeSS (2008) also defines “Best Practice” as “the adoption of products and
techniques that are already established in the market, cost effective and able to save
energy without incurring undue risks”. This section focuses on the “basic” and “best
practice” specifications of domestic space heating control systems.
According to CHeSS 2008, a “Basic” central heating system must have following
control specifications:
•
Full programmer and cylinder thermostat (for regular boiler with separate hot
water store)
23
•
Time switch (for combination boilers)
•
Room thermostat
•
Boiler interlock
•
TRVs on all radiators, except in rooms with a room thermostat.
•
Automatic bypass valve
According to CHeSS 2008, a “Best Practice” central heating system must have
following control specifications:
•
Programmable room thermostat (with additional DHW timing capability for
regular boiler)
•
Boiler interlock
•
TRVs on all radiators except in rooms with a room thermostat
•
Automatic by pass valve
•
Cylinder thermostat (only for regular boilers)
The main difference between the control specifications of “Basic” and “Best Practice”
central heating systems is that in “Best Practice”, programmable room thermostat
enables the households to program their heating in order to set different target
temperatures (i.e. set-point temperature) throughout a day. In “Basic” systems,
different set-point temperatures could only be set manually using a room thermostat.
In recent years, more attention has been paid into controlling different zones in
dwellings separately as reflected in Building Regulations Part L1A for new dwellings
which came into force from 1 October 2010 (HM Government, 2013). According to
Domestic Services compliance guide (Department for Communities and Local
Government, 2011), which provides more detailed information on the guidance
contained in approved documents of Part L1A (for new dwellings) and L1B (for
existing dwellings), since 1 October 2010 every new home which is not open plan
must be divided to at least two heating zones such that living and sleeping areas can
be controlled at different temperatures by means of two thermostats. If the house is
smaller (less than 150 𝑚𝑚2 ), then these two zones can be controlled by the same
timer. This means that the flow of heat in each zone is controlled via separate room
thermostats and motorised valves; although the same heating schedule can be
applied for both zones using the same timer. If the house is larger (more than 150
24
𝑚𝑚2 ) then each zone must be controlled by its own timer. This only applies to the new
homes but the minimum requirements for heating controls have not changed for
existing dwellings since 2002 when the Building regulations part L came into force.
Figure 2-9 and Figure 2-10 were adopted from a guide by The Association of
Controls Manufacturers (TACMA) on how to comply with the 2010 Building
Regulations Part L. They show examples of heating system layouts for new
dwellings (layouts 1-6) and existing systems (layouts 7-12) that comply with “Basic”
and “Best Practice” heating controls for different boiler types, dwelling size and valve
types. In Figure 2-9, layouts 1, 2, 5 and 6 comply with “Basic”, and 3, 4 with “Best
Practice”, heating controls for new dwellings. In Figure 2-10, layouts 7, 8 and 11
comply with “Basic” and 9, 10 and 12 with “Best Practice” heating controls for
existing dwellings.
25
Figure 2-9: Example layout for new systems to ensure compliance with the 2010
Building Regulations Part L1A (TACMA, 2010)
26
Figure 2-10: Example layout for replacement boilers to ensure compliance with the
2010 Building Regulations Part L1B (TACMA, 2010)
27
2.4.2 Space heating controls in existing homes
Prior to December 2013, when Energy follow up survey (EFUS) 2011 (BRE, 2013)
was published, there were very few reports or literature on the status of central
heating controls in UK homes (Munton, Wright, Mallaburn, et al., 2014). Most of the
information available was according to control manufacturers which indicated poor
levels of space heating control in UK homes.
One of the largest control manufacturers in the UK, Honeywell, noted that from 26
million homes in the UK, about a third, do not have room thermostats which cause
excessive room temperatures (Enviros Consulting Ltd, 2008). Similarly, work carried
out by TACMA (The Association of Controls Manufacturers) with the Energy Saving
Trust reported that 30% of homes in the UK do not have a room thermostat (Heating
and Hot Water Task Force, 2010). Enviros Consulting Ltd (2008) estimated that
almost a quarter of homes do not have either a programmable thermostat or a room
thermostat. In addition, they estimated that nearly 40% do not have any TRVs
installed (Enviros Consulting Ltd, 2008). Enviros Consulting Ltd (2008) stated that 70%
of the dwellings do not have modern standard heating controls set by building
regulations. More dramatically, according to Heating and Hot Water Taskforce (2010)
there were 4% of homes with a boiler and no controls at all.
These information were mainly in agreement with findings from a literature review by
the statutory consumer champion for England, Wales, Scotland and Northern Ireland
published in July 2012 (Consumer focus, 2012) which shed more light on the
percentages of UK households with each of the main heating control types
(Figure 2-11).
28
Figure 2-11: Percentage of UK households with a boiler with each of the main
heating control types as reported in Munton et al. (2014)
EFUS 2011 which was funded by DECC and carried out by the Building Research
Establishment (BRE) collected new data on the patterns of household and dwelling
energy use including information regarding what heating controls are currently
installed in UK homes. Data was collected from an interview survey of a selfselecting rather than randomly selected sample of 2616 households (BRE, 2013).
The results of EFUS contradict the earlier findings showing that 49% of the
households in their sample have full set of controls compared to only about 30%
found in the previous reports. A report by Munton et al. (2014) who compared the
data from EFUS 2011 with Consumer Focus report from 2008 argued that the
proportion of households with central heating that have a range of controls may have
increased over the recent years. Munton et al. (2014) summarized the most recent
information available regarding the status of space heating controls in UK homes
and its relationship with built type from EFUS 2011 data which are reproduced and
presented in Table 2-3.
Table 2-3 shows that most UK homes in their sample (97%) have a central timer
regardless of the dwelling type. More than two third of the dwellings in each category
have room thermostat with an average of 77% for the whole sample. However, room
thermostats are least common in high rise flats (67%) and most common in
bungalows (83%). It also indicates that above 60% of the central heating systems in
each dwelling type have TRVs installed. The lowest percentage of dwellings which
29
have thermostats was found in high rise purpose built flats which had the highest
percentages of dwellings with TRVs installed.
Table 2-3: Proportion of dwelling types reporting primary heating controls
(reproduced from Munton et al., (2014))
Dwelling/Household type
Room Thermostat
(%)
Central Timer
(%)
TRV
(%)
Full set of
1
controls (%)
Whole sample
77
97
66
49
Purpose built flat, high rise
67
99
78
52
Purpose built flat, low rise
77
98
65
49
End terrace
76
96
69
51
Mid terrace
77
97
69
52
Converted flat
77
97
67
51
Bungalow: all ages
83
97
66
53
Detached house: Pre 1919
76
98
70
52
Detached house: Post 1919
74
96
59
43
Semi-detached & terraced: pre
1919
75
98
66
49
Semi-detached & terraced: 19191944
71
98
63
43
Semi-detached & terraced: 19451964
82
98
61
49
Semi-detached & terraced: 1965
onwards
80
97
66
53
1
Including TRVs, central timer and a room thermostat
30
2.5 Impacts of space heating controls on energy demand
Improving the efficiency of domestic heating systems can be studied by considering
the individual components such as boilers, thermostats, heat emitter controls (TRVs),
pumps etc. or by considering the heating system as a whole (Liao, Swainson &
Dexter, 2005). Liao et al. (2005) argues that although each individual item is
becoming more efficient, the improvement in efficiency of the heating system as a
whole is still unknown to a large extent. They suggested considering all components
when looking for ways to improve energy efficiency rather than concentrating only on
one individual item. Liao et al. (2005, p344) argue that “It is vital therefore that the
interaction of the whole heating system within a building is considered when looking
at controls and that a reliable and repeatable method of testing is developed to allow
claims of performance to be assessed in terms of both energy and thermal comfort
achieved”.
Heating controls have the potential to reduce the heating energy demand in two
main ways; by reducing the length of space heating in a house or altering the heating
demand temperature at different spaces of a house according to its occupancy and
usage patterns (Firth, Lomas & Wright, 2010). Research shows that the length of
heating period and heating demand temperatures are the most influential factors
affecting the amount of heating energy which is consumed in homes and their
relevant CO2 emissions. Firth et al (2010) estimated the length of the heating period
and the heating demand temperature to have normalized sensitivity coefficients of
0.62 and 1.55 on CO2 emissions respectively. This indicates that for every 1%
increase in the heating demand temperature, a 1.55% increase in average dwelling
CO2 emissions will result. Also, a 1% rise in the number of heating hours is
estimated to result in a 0.62% rise in CO2 emissions (Firth et al. 2010).
The studies which investigated the impacts of space heating controls on energy
demand can be divided into two main categories depending on the type of heating
controls tested. A number of studies examined the effects of adding one or more
conventional heating control components such as room thermostat, Programmable
Room Thermostat (PRT) or TRVs to an existing heating system. They will be
discussed in section 2.5.1. Other studies evaluated the energy saving potential of a
31
number of methods to control the delivery of heat in buildings more efficiently
according to the building occupancy. They will be discussed in section 2.5.2.
2.5.1 Conventional space heating controls
Several studies were conducted to investigate the potential space heating energy
savings which can be achieved by employing a number of conventional control
components in homes including room thermostats, programmable room thermostats
and TRVs. Based on their approach, these studies can be divided into three groups.
In the first group, there are studies which used models (either steady state or
dynamic). The second group used test house facilities to conduct a side-by-side
comparison of the effects of different heating controls on energy demand and
thermal comfort of a matched pair of test houses with synthetic occupancy. These
studies were conducted by Building Research Establishment (BRE) in the late 1970s
and 1980s (Rayment et al. 1983 and Rayment & Morgan 1984). In the third group,
there are studies which compared energy demand or factors which influences the
energy demand between real homes with different types of heating controls. The
major difference between groups 1 and 2 and group 3 is that in group 1 and 2,
studies often assume standard occupancy behaviour while the third group takes into
account effects of occupants’ interaction with the heating system controls.
An example of group 1 studies is the Good Practice Guide 302 (BRECSU, 2001)
which used the Standard Assessment Procedure (SAP) for energy rating of
dwellings (BRE, 2014) to estimate the energy savings which could be achieved in
UK homes by applying better controls. According to them, installation of the
minimum standard of controls in a wet system which previously had no controls
reduces fuel consumption and CO2 emissions by 17%. They also argues that turning
down a room thermostat by 1ºC will reduce space heating demand by 6-10% and
reducing the heating on time by two hours a day can reduce demand by 6%
(BRECSU, 2001). Good Practice Guide 302 also provided a Table in which the
average potential savings which could be achieved by adding different features to
improve an existing heating control system is predicted for different house types
depending on their boiler type (Table 2-4). The guide explains that these predictions
were based on assuming normal controls, systems and user behaviours and
therefore actual savings in individual systems may be significantly different. However,
32
the details and assumptions involved in these predictions were not mentioned in the
guide.
Table 2-4 shows that the most energy savings across all the dwelling types can be
achieved when the existing system does not have any type of controls. When the
existing system already has control components such as room thermostat or TRV
the percentage energy savings of adding additional control components reduces. For
example, adding normal TRVs on all of the radiators to an existing heating system
which has a room thermostat and boiler interlock, could on average only save 4% of
fuel consumption regardless of the boiler type.
Table 2-4: Typical average annual fuel and cost savings (£) which could be achieved
from better heating controls (Table reproduced from Good Practice Guide 302
(BRECSU, 2001))
Existing
system has
the following
controls
Improved system add
the following for the
minimum set
Approximate
Typical average Annual fuel cost savings
average
(£)
saving (% of
the existing
fuel
Terraced
Semidetached
Deatched
consumption)
Typical boiler with gravity DHW
-
RT,CT,MV,BI,TRV
17%
51
58
82
RT
CT,MV,BI,TRV
12%
36
41
58
RT,CT,MV,BI
TRV
4%
11
13
18
TRV
RT,CT,MV,BI
9%
27
31
44
Typical boiler-fully pumped
-
RT,CT,MV,BI,TRV
17%
51
58
82
RT, CT, MV
BI,TRV
10%
30
34
48
RT,CT,MV,BI
TRV
4%
11
13
18
TRV
RT,CT,MV,BI
9%
27
31
44
Typical combination boiler
-
RT, BI, TRV
15%
45
52
73
TRV
RT, BI
7%
21
24
34
RT, BI
TRV
4%
11
13
18
RT=Room Thermostat, BI=Boiler Interlock
TRV=Thermostatic Radiator Valve, CT=Cylinder Thermostat, MV=Motorised Valve
33
The estimated typical energy savings of installing TRVs using SAP was considerably
lower than the claimed energy savings by their manufacturers. Tahersima et al.
(2013) noted that TRVs can reduce the heating demand by up to 20%. However,
their reference was only based on a claim on the website of a large manufacturer of
heating controls (Danfoss). Hartmann (no date) in another document written for
Danfoss noted that “according to experience” TRVs save 10-15% of energy and this
could be up to 20% in “individual cases”. It should be noted that although TRVs are
in use for decades, there are only a few published literature which investigates the
energy savings of TRVs (Dentz & Ansanelli, 2015).
Studies which used dynamic thermal modelling estimated higher potential savings by
using TRVs compared to what estimated by Good Practice Guide 302 (BRECSU,
2001). Xu et al. (2008) conducted a modelling analysis based on an existing multifamily building and heating system in China and found that 12.4% of heating energy
can be saved if the TRVs were kept on level 2-3 instead of being fully open (level 5).
This saving was achieved due to the TRVs help in reducing the overheating. The
mean room temperature for the whole building was reduced from 22.8°C when TRVs
were fully open (or in other words when the heating system was operated without
valve control) to 20.5°C with TRVs on level 2-3 (Xu, Fu & Di, 2008). Xu et al. (2008)
also reported a monitoring study by Wang and DI (2002) from the Chinese
government demonstration projects that indicated an average heating demand
reduction of 10% when using TRVs.
A recent study (Monetti, Fabrizio & Filippi, 2015) used EnergyPlus simulation
software to construct and calibrate a dynamic thermal model based on the
monitoring data in order to investigate the effect of TRVs on energy demand of an
old existing multi-family home in Italy. Their case study results showed that the total
heating demand of a heating season can be reduced by up to 10% by using TRVs
and suggested that TRVs can be considered as low cost energy efficiency measure
that can be easily applied to old buildings (Monetti, Fabrizio & Filippi, 2015). Again,
their study was based on theoretical assumptions about occupants’ behaviour. For
example no monitored data regarding the occupant’s interaction with TRVs and
heating temperature set-points was available. They argued that higher quantity and
quality data was needed for better calibration (Monetti, Fabrizio & Filippi, 2015).
34
In contrast with manufacturer claims and model predictions which suggest
considerable potential for energy savings by better heating controls, a number of
studies suggest conflicting results in real world configurations.
Shipworth et al. (2010) in a study of 427 UK households argues that in contrast to
what is currently assumed in policies and regulations, the use of “simple controls”
(thermostats and time clocks) in homes does not reduce energy consumption. They
found that the sample of homes without thermostatic control of the central heating
system had mean estimated thermostat settings of 0.6ºC below those with
thermostatic control. In addition, they found that central heating systems operated by
timer are active 0.4 hours/day longer than those operated manually. They suggested
that alternative forms of heating controls which appeal to householders should be
developed and tested.
In a side-by-side comparison study of BRE (Rayment, Cunliffe & Morgan, 1983),
they found that there is no significant difference between the room air temperatures
and space heating gas demand of a house controlled by a room thermostat and
TRVs compared to a similar house controlled only by TRVs. Rayment et al. (1983)
argues that for the type of occupancy and house tested, room thermostat could be
as effective as TRV control.
Conventional TRVs were found not to perform and operate as designed in real world
set ups after many years in service (Liao et al., 2005 ) & (Dentz & Ansanelli, 2015). A
survey of 35 buildings by Liao et al. (2005) although focusing on non-residential
dwellings found that more than 65% of the TRVs were performing poorly as they
failed to reduce the heating output of radiators when the room temperature was
greater than its desired value and therefore the rooms were overheated.
Figure 2-12 which is adopted from Liao et al. (2005) shows indoor temperatures in
three rooms in a building with TRVs in their study and the corresponding external
temperature. As it can be seen temperatures of up to even 29ºC was observed
showing that the TRVs were not performing well.
35
Figure 2-12: Average air temperature in a building with TRV controlled radiators
(Liao et al., 2005)
In addition, 32% of the TRVs in this study were found to be set at maximum and
more than 65% were found to be set at higher than required. One reason for the
settings higher than required could be due to the fact that occupants do not often
interact with the TRVs. Osz (2014) described a number of factors that could
influence householder’s interaction with the TRVs. These included difficulty in
reading and interpreting the settings (poor design), existence of different styles of
TRV in homes which caused confusion, and householder’s lack of understanding
regarding how TRVs work. Dentz & Ansanelli (2015) argues that even among
experienced professionals there is a range of understandings about TRVs and some
have little knowledge of TRVs.
Another real world example is the PRTs which have been considered as one of the
main components for energy saving in space heating. The basic idea of the PRTs
has been to use two temperature targets and heat the house to a set-point
temperature when the occupants are present and active and let the house to float to
a lower, more energy efficient set-back temperature when the occupants are typically
away or asleep (Lu, Sookoor, Srinivasan, et al., 2010). However, Lu et al (2010)
argues that the households with a simple dial-type thermostat could easily adjust the
temperature settings before going to sleep or leaving the house and save more
36
energy compared to the households with programmable room thermostats in which
the heat is often wasted because it is often not possible to find one or two general
schedules which can be applied for highly dynamic occupancy patterns of most
homes.
A number of studies which compared energy demand and heating practices in
houses with a programmable room thermostats and houses with a simple room
thermostat were reviewed by Wei et al. (2014). The main findings from a number of
these studies were summarized in Table 2-5.
37
Table 2-5: Studies which compared energy demand and heating practices in homes
with a programmable room thermostat and homes with a room thermostat and their
main findings (Wei et al. 2014)
Number of
homes,
Study
Method of
Main findings
collecting data,
country
•
(Nevius &
Pigg, 2000)
not significantly different than homes with a RT
299 homes,
survey &
Houses with PRT had thermostat set-points which were
•
measurement, US
Houses with PRT uses on average 2.5% less energy
than houses with a RT but this difference was not
statistically significant
Jeeninga et
al. (2001) in
Dutch
reported in
(Groot et al.
180 homes,
questionnaire,
Netherlands
•
Preferred set-points are not affected by the type of room
thermostat (programmable or manual)
•
Lower set-point temperatures during long unoccupied
period in homes with RT compared to PRT
2008)
•
Higher temperature settings during the night in houses
with PRT
(Guerra
313 homes,
Santin &
questionnaire,
Itard, 2010)
Netherlands
•
No statistically significant difference between the hours
of use of thermostat and the thermostat setting between
the houses with RT and PRT
•
The type of thermostat affects the number of rooms
heated. In houses with PRT the occupants take less
actions and leave the control to the PRT
(Tachibana,
2356 homes,
2010)
questionnaire, US
•
86% of the homes with PRT applied evening to night
time set-back compared to 66% of the homes with RT
(Lutzenhiser,
Cesafsky,
279 homes,
Chappells, et
survey, US
•
Homes with manual thermostat use less energy
compared to homes with programmable thermostat
al., 2009)
PRT=Programmable room thermostat
RT=Room Thermostat
Munton et al. (2014, p57) discusses that the failure to find consistent evidence that
improved domestic heating control technologies deliver energy savings could be due
38
to poor experimental design. They suggest a number of important factors which need
careful consideration. These include:
•
“having robust and consistent definition of control technologies”,
•
“monitoring actual house temperatures and heating durations”,
•
“an experimental design, at the very least involving a matched comparison
that enables the study to control for intervening variables”
•
“measuring consumer behaviour carefully”.
2.5.2 Occupancy based space heating control
A number of studies have investigated the potential for saving space heating energy
by controlling delivery of the heat more efficiently according to the presence of
occupants in a space. These studies were mainly undertaken in the US where the
majority of buildings are equipped with forced air heating systems. In a monitoring
study of 8 homes in the US it was found that only half of the rooms were occupied for
up to 60% of the time when the home was occupied, and that the occupancy of
these rooms was predictable based on ongoing activities and times of the day (Lu,
Sookoor, Srinivasan, et al., 2010).
Meyers et al. (2010) estimated that 2.7% of all residential primary energy in the US is
spent on heating unoccupied homes assuming that on average, homes are
unoccupied for 4 hours during a weekday. Assuming the percentage of floor space
occupied by bedrooms and living rooms to be 48% and 52% respectively, they also
estimated that 6.2% of total primary energy is wasted for heating or cooling the living
rooms during the night period when unoccupied. Moreover, 9.7% of total primary
energy is wasted for heating or cooling the bedrooms during the 4hours of a day
which was assumed that occupants spent in the living rooms. This shows a total of
15.9% of wasted primary energy for heating or cooling unoccupied spaces of a
typical US home. This was the largest waste amongst different inefficient energy
delivery options and appliances which was investigated in their study including
thermostat oversetting, leakage current and appliance choice.
In addition, Meyers et al. (2010) investigated the energy savings from having
individual control of each zone compared to when there is a single central thermostat
controlling the whole house. Having a central thermostat could result in temperature
39
variations in homes (particularly upstairs cf downstairs) and therefore it may cause
the thermostat to be set at higher temperatures to sufficiently heat the spaces that
are far from the thermostat. Assuming that on average, thermostats in dwellings are
set 1ºC higher in winter and 1ºC lower in summer than the residents desired
temperatures, they estimated that 1.25% of total primary energy can be saved in US
dwellings. However, they did not conduct any measurements or consider the impact
of indoor air temperature reductions on thermal comfort. All the estimations were
based on a framework developed by using the energy data for 4383 US households
collected by Residential Energy Consumption Survey (RECS) (US Department of
energy, 2005) in the US.
Several researchers have investigated the use of occupancy sensors to control
HVAC systems based on either real time occupancy data or occupancy models
integrated into building HVAC systems (Lu et al, 2010; Agarwal et al. 2011;Erickson
et al. 2013).
Scott et al. (2011) developed ‘PreHeat’ system and tested it in five homes, three in
the US and two in the UK. The ‘PreHeat’ system was designed to enable home
heating to be controlled automatically according to occupancy sensors and future
occupancy prediction. All homes tested were family homes with two adults and one
or more children. All US homes had forced air heating. One of the UK homes had a
combination of underfloor heating and radiators while the other had radiators in all
rooms. They compared the ‘PreHeat’ prediction algorithm with a seven day
programmable thermostat with preconfigured heating schedules (i.e. scheduled
algorithm). Individual room heating control according to occupancy sensors were
applied in UK homes while in US homes the whole house air heating system was
controlled according to the occupancy sensors. They alternated the heating control
strategy each day between the ‘PreHeat’’s prediction algorithm and the scheduled
algorithm in order to balance any effects of weather or household schedule changes.
The resulting difference between the average outdoor temperature of PreHeat days
and scheduled days was less than 0.3ºC. However, they did not mention how the
household schedules could have been different from day to day. Over a 61 day
monitoring period, ‘PreHeat’ resulted in little difference in gas use for the homes in
40
the US with a whole house heating control system but resulted in 8% and 18%
reduction in gas use for the individually controlled rooms in the UK homes.
Moreover, Badiei et al (2014) used dynamic thermal modelling to investigate the
effect of changing heating set-points and length of heating using programmable
TRVs on energy demand of a three bedroom UK house. They found that decreasing
heating set-point of every radiator in the house at the same time from 1°C to 5°C
would result in 16% to 64% reduction in annual gas demand for space heating.
Decreasing heating set-point in an individual room showed 3.7% to 14.5% reduction
in annual gas demand. In addition, reducing heating time in all rooms from one hour
to five hour resulted in 5.8% to 28% reduction in annual gas demand. Such reduction
for an individual room showed a potential of 1.1% to 6.2% reduction in whole house
annual gas demand.
Lu et al. (2010) reported average energy savings of 28% from deploying occupancy
sensors in 8 homes. The sensors were designed to automatically turn off the HVAC
system, when the occupants were sleeping or away from home, using a “smart
thermostat” compared to a heating system with “reactive thermostat”. The homes
included both single person and multi person residents as well as houses shared
between students and professionals. They developed an algorithm that analysed
patterns in sensor data in order to recognize people leaving or sleeping so that the
system could be switched off within few minutes of the event. The HVAC system was
heating the whole house when occupied and not sleeping. There was no individual
control of different rooms in their study.
Agarwal et al (2011) used real time occupancy data from a wireless occupancy
sensor network across one floor of a four floor US university building to control and
actuate individual HVAC zones to be conditioned. They reported space heating and
cooling energy savings of 8% to 13%. The authors discuss several applications of
real time occupancy information and combined use of HVAC and IT resources for
commercial buildings.
According to Erickson et al (2013- p1&2) for occupancy based HVAC control,
occupancy detection needs to be accurate, reliable and able to capture occupancy
changes in real time. Moreover, the authors argue that “Unlike lighting, the thermal
41
ramp up or down of a room involves delay. While an optical system of occupancy
monitoring can give occupancy in near real-time, reactively conditioning a room will
likely leave occupants uncomfortable until target temperatures are met. In order to
ensure occupant’s comfort, we must be able to predict when occupants are likely to
enter a room and begin conditioning before-hand” (Erickson et al, 2013).
Holland (2010) described a number of factors which needed to be taken into account
when considering what was named “dynamic zoned control of the heating system”
where a zone is not heated unless it is in use. Three important factors were user
definable set-back temperatures which should be used for the unoccupied periods;
the length of warm up time which is required for the rooms to achieve the comfort
condition from the set-back condition; and the expected time of occupancy for each
room.
2.6 Zonal space heating control
There are currently an increasing number of manufacturers of heating controls that
are providing ZC systems for the new and existing homes with wet central heating
systems (for example: Honeywell, 2015; Heat Genius, 2013; Eurotronic, 2011;
Honeywell, 2014; Salus controls, 2013). These systems can be implemented easily
and quickly and with minimum disruption for households as installing these systems
does not need any pipe change, draining down 3, running wires, plastering to do or
lifting floor boards (Honeywell, 2014). The main component of such systems is
Programmable Thermostatic Radiator Valves (PTRVs) which could replace the
existing conventional TRVs simply by unscrewing the TRV heads and screwing
PTRVs according to their manual (Honeywell, 2015; Heat Genius, 2013; Eurotronic,
2011; Honeywell, 2014; Salus controls, 2013).
PTRVs are battery-operated and have motorised valves and temperature sensor to
control the flow of hot water to the radiators according to a target temperature
schedule assigned for the room where the radiator is located (Figure 2-13)
(Honeywell, 2014).
3
However, if TRVs are not already installed, draining down is required and often a professional
installer is needed.
42
Each room with a PTRV can have a number of different target temperatures
throughout a day and schedules could be different from day to day and weekdays to
weekends (Honeywell, 2015; Heat Genius, 2013; Eurotronic, 2011; Honeywell, 2014;
Salus controls, 2013). Therefore, the rooms can be scheduled to be heated only
when they are occupied and to the level needed.
ZC systems available in the market can be divided into two main categories: modern
luxury systems including PTRVs, a user friendly touch screen wireless central
controller and a boiler relay (type1) (Honeywell, 2015 and Heat Genius, 2013) and
simple “stand alone” PTRVs without any central controller or boiler relay (type2)
(Eurotronic, 2011; Honeywell, 2014 and Salus controls, 2013).
The schedules for the target temperatures can be set via the central controller which
communicates wirelessly with the PTRVs (in type 1 systems) (Honeywell, 2015 and
Heat Genius, 2013) or on the PTRVs themselves (in type 2 systems). The central
controller can be also connected to a tablet or mobile phone wirelessly via internet
and thus, the schedules can be modified remotely in type 1 systems (Honeywell,
2015 and Heat Genius, 2013). In addition, the temperature settings can be manually
overridden by the occupants if needed. In contrast to conventional TRVs which were
adjustable to 5-6 different levels which often left the households without a clear
understanding of what temperature each level is representing (Osz, 2014), exact
temperatures can be adjusted using PTRVs.
The main difference between type 1 and 2 systems is that in type 1 systems the
boiler is switched on when the air temperature in any of the zones with PTRVs drops
below its set-point temperature (Honeywell, 2015 and Heat Genius, 2013) while in
type 2 systems, the boiler operation is conventionally controlled using a room
thermostat and programmer or a programmable room thermostat (Eurotronic, 2011;
Honeywell, 2014 and Salus controls, 2013).
While type 1 systems might be considered as full zonal space heating control, type 2
systems could be more relevant for UK homes where the households often tend to
heat their homes for certain hours during a day and the heating is often switched off
at night with no set-back temperature (Huebner, 2013). Moreover, applying type 1
systems to existing dwellings requires replacing the thermostat and boiler relay
43
already existed in the system with the new wireless central controller and boiler relay
(Honeywell, 2015 and Heat Genius, 2013). This would result in considerably
increasing the capital costs of the system (Table 2-6) as well as installation costs
(Honeywell, 2015 and Heat Genius, 2013).
Table 2-6: A number of systems currently available in the UK market and their prices
(as in 24 February 2015) for a configuration which can apply zonal space heating
control in a typical UK house
System
Type
Honeywell Evohome
1
PTRV price
1
1
Total system price for a
2
boiler relay price (£)
per unit (£)
typical UK house (£)
1
£178.8
£58.69
£531
Heatgenius
1
£249.99
£59.99
£610
Honeywell HR90
2
-
£39.59
£238
Salus PH60C
2
-
£29.38
£177
2
-
£15.95
£96
Eurotronic Sparmatic
Comet
1
Central controller +
Prices are VAT included but do not include the costs of installation and batteries and were sourced
from the main dealers of the products in the UK on 24 February 2015.
2
The house was assumed to have 3 bedrooms, a living room, a dining room, a bathroom and a
hallway as heated spaces which comprises 7 zones, 6 of them controlled using PTRV and one
controlled using a central controller or the existing room thermostat. The house was assumed to have
a combination boiler.
Type 1 systems might be suitable for those homes with no existing heating controls
where upgrading to the cheaper type 2 systems would also need capital costs for a
room thermostat and programmer (Eurotronic, 2011; Honeywell, 2014 and Salus
controls, 2013). While type 1 systems could often be more user friendly as they are
programmed using a touch screen central controller or/and computers, tablets or
phones and also provide advance features such as remote access control
(Honeywell, 2015 and Heat Genius, 2013), type 2 systems could be used as a cheap
energy efficiency measure which can be added to an existing heating system by the
householders themselves, with no need for any electrical work or plumbing to be
done by an external installer (Honeywell, 2014).
Honeywell’s latest ZC product Evohome (Honeywell, 2015) (Figure 2-13) is an
example of type 1 system which features:
44
•
Touch screen central controller with ability to control up to 12 zones.
•
Smart phone application which enables households to monitor and control
their heating whether they are home or not. For example, it allows them to
start heating their homes before they arrive home from work to avoid a cold
home on their return. The connection between the central controller and
smartphone is established using a remote access gateway.
•
Auto window function which realise if a window has been left open and stop
heating that zone in order to save energy.
•
Optimum start and stop: According to Honeywell (2015), Evohome is able to
understand how a home heats up and cools down and thus, works out the
exact time when a room needs to start heating up or cooling down to be at the
desired temperature set for a time in a day.
However, additional features such as auto window function, optimum start or remote
access which could add extra energy savings to the zonal control systems were out
of the scope of this study and were not investigated in this work.
Figure 2-13: Honeywell’s Evohome system components including PTRV and central
controller
Heat Genius (Heat Genius, 2013) is another type 1 system with comparable features
to Evohome which is currently available in the UK market. According to Heat Genius
(2013), one of the unique features of the system compared to its counterparts is that
wireless room sensors could be added to the system which detects when people are
45
using different rooms, thus can automatically schedule the radiators in each room to
only come on at times when people normally use these rooms. However, the
algorithms which lead to such automatic schedules were not described by the
manufacturer.
Apart from a limited number of type 1 ZC systems available in the UK market, there
are good number of “stand alone” programmable thermostatic radiator valves (PTRV)
products (type 2 system) which all have the same function though they have different
designs and prices (Figure 2-14). Honeywell’s HR90 (Honeywell, 2014) have similar
to PTRVs existed in the Evohome system but they can be programmed using the
keys and displays on the PTRV heads and thus do not need a central controller for
assigning the heating schedules and set-point temperatures (Figure 2-14). Similar to
PTRVs in Evohome system they use two 1.5 Volts batteries and also have autowindow function.
Figure 2-14: A number of PTRVs from different manufacturers: from left to right:
Honeywell HR90 (Honeywell, 2014), Salus PH60C (Salus controls, 2013) and
Eurotronic Sparmatic Comet (Eurotronic, 2011)
2.7 Modelling domestic energy use
The modelling techniques for estimating energy use in houses can be divided into
two main approaches: top-down and bottom-up. The top-down approach considers
the residential sector as an energy sink and is not concerned with the individual
dwellings (Swan & Ugursal, 2009). It uses historical statistics of energy use and
46
households on a national level and predicts the influence of changes in top level
factors such as energy price, climate and macroeconomic indicators such as gross
domestic product, unemployment and inflation on energy consumption of the whole
housing stock (Swan & Ugursal, 2009). Therefore it is not suitable for investigating
the effects of energy efficiency measures on energy demand.
On the other hand, the bottom-up approach is based on principles of building physics
and calculates the energy use of a representative group of individual houses,
allowing extrapolating the results to regional or national levels (Swan & Ugursal,
2009). The bottom-up models require a large number of input parameters such as
building geometry, fabric, characteristics of the heating systems, internal
temperatures and heating patterns, ventilation rates, individual appliances, external
temperatures etc.
The bottom-up models can be divided into steady state and dynamic models.
2.7.1 Steady state models
Current approaches in bottom-up domestic stock modelling in the UK typically
employ steady state or quasi steady state calculations to estimate the monthly or
annual energy demand (Taylor, Allinson, Firth, et al., 2013). The majority of bottomup residential stock models developed to date in the UK such as BREHOMES
(Shorrock & Dunster, 1997), The Johnston model (Johnston, 2003), UKDCM
(Environmental Change Institute, 2009), The DECarb model (Natarajan & Levermore,
2007) and CDEM (Firth, Lomas & Wright, 2010) have used the same calculation
engine known as Building Research Establishment Domestic Energy Model
(BREDEM) (Kavgic, Mavrogianni, Mumovic, et al., 2010). BREDEM has different
versions such as: BREDEM-8, which is developed for monthly analysis; BREDEM-12,
for annual analysis; and BREDEM-9 which is a monthly version and the basis of the
UK government’s Standard Assessment Procedure (SAP) (Kavgic et al. 2010). The
Standard Assessment Procedure (SAP) 2012 is the latest version of the UK
government’s approved methodology for rating the energy performance of new
dwellings (BRE, 2014). Reduced Standard Assessment Procedure (RdSAP) is used
for the energy performance assessment of existing dwellings (BRE, 2014). RdSAP is
based primarily on SAP procedures and has additional standard data tables which
47
are added to the SAP model to replace the information which is not available for the
existing dwellings.
Much research has been conducted across the world using the bottom-up approach
to evaluate the potential for energy savings and economic benefits of using different
energy efficiency measures (Swan & Ugursal (2009)). Bottom-up models could
provide good estimates of the effectiveness of different energy efficiency measures
for policy makers (Kane, 2013). In the UK, SAP, RdSAP and BREDEM have been
used in a number of key energy and environmental policy initiatives such as Warm
Front (2014c), Green Deal and Energy Company Obligation (DECC, 2011b), and
code for sustainable homes (Department for Communities and Local Government,
2006).
However, SAP’s procedure to model the energy use in houses with ZC is simplified
and may not be suitable for detailed analysis of specific houses. SAP 2012 (BRE,
2014) defines “time and temperature zone control” as “a system of control that allows
the heating times of at least two zones to be programmed independently as well as
having independent temperature control”. SAP 2012 (BRE, 2014) discusses that this
could be achieved by “separate plumbing circuits, either with their own programmer
or separate channels in the same programmer” or “programmable TRVs or
communicating TRVs”.
SAP 2012 (BRE, 2014) considers fewer hours of heating for the “rest of the house” 4
with a system with “time and temperature zone control” (7 hours per day; from 07:00
to 09:00 and 18:00-23:00) for all days compared to other conventional control
options with 9 hours during the weekdays (from 07:00-09:00 and 16:00 – 23:00) and
16 hours during the weekends (from 07:00-23:00). In addition, it uses a lower mean
temperature for the “rest of the house”.
SAP’s procedure does not take into account a number of factors. For instance, it
does not take into account the number of rooms which are controlled separately
using programmable TRVs. As long as the house has two zones or more, the
4
In SAP 2012 (BRE, 2014), monthly heating requirements of a house are calculated using mean
internal and external temperatures and the heat transfer coefficient allowing for internal and solar
gains. The mean internal temperature is calculated separately for the living area (often the living room)
and the rest of the house. The mean living room and rest of the house temperatures will then be
combined to find the mean internal temperature for the whole house.
48
procedure will remain the same without differentiating between numbers of zones
which are separately controlled. In addition, since SAP estimations are independent
of occupant behaviour, different set-back temperatures used in the zones controlled
by PTRVs which could influence the energy saving potential of a “time and
temperature control” system are not reflected in the SAP’s procedure. The length of
the period when each room is heated to set-back temperature could also be different
from house to house.
As discussed here, weaknesses exist in SAP 2012 regarding the assumptions made
about the occupant behaviour, hours of occupancy and the use of heating system
suggests that steady state building physics models such as SAP should not be used
for detailed analysis of energy savings before and after installing ZC.
2.7.2 Dynamic models
Dynamic thermal modelling could be used for more detailed analysis of the energy
demand reduction potential of applying ZC as they offer the highest flexibility to
model any system and occupancy. A number of dynamic thermal modelling tools
such as DOE-2, EnergyPlus, TRNSYS, ESP-r and IES<VE> have been widely used
in the past decade in early building design as well as analysis of retrofit opportunities
(Crawley, Hand, Kummert, et al., 2008). The main focus has been on modelling
larger commercial buildings rather than modelling domestic energy use (Porritt,
2012). Taylor et al. (2013) were one of the first to try using dynamic thermal
modelling for modelling a whole English region housing stock. They found the level
of details available for the model inputs as one of the factors which affected the
energy predictions with higher level of details resulted in higher energy predictions.
It is important to consider the capabilities of each dynamic thermal modelling tool
and choose the one which suits the most for the specific problem under investigation.
The main feature of ZC is that different rooms are kept at different temperatures
throughout a day. Any model should be dividable into various zones (i.e. each room
with a radiator will be separate zone) where the set-points temperature of each zone
could be altered throughout a day. Most of the current dynamic thermal modelling
tools such as DOE-2, EnergyPlus, eQuest, TRNSYS and Trace700 are based on
multi-zone thermal models and have such capability.
49
The tools mentioned are often focused on representing building characteristics
accurately but to a lesser extent on the heating systems and controls. As a result,
detailed hydraulic behaviours of the heating systems (e.g. The TRVs or PTRVs
dynamic control process) are often represented in a simplified way.
The other important factor to consider when modelling houses with ZC is the interzone heat transfer. Thermal energy is transferred by convection from one zone of a
building to another via air flow through doorways and windows (Allard & Utsumi,
1992). This inter-zone convection could be either natural convection due to
temperature differences between spaces, or forced convection by the pressure
differences which are caused by mechanical ventilation or air distribution systems in
buildings or a combination of both (Barakat, 1987). Keeping the rooms of a naturally
ventilated house at different air temperatures throughout a day such as in ZC will
result in natural convective heat flows through different rooms. Previous research
has shown the significance of natural convection through door openings. For
example flow rates of more than 1200 W have been observed to occur through a
0.9*2.05 m doorway as a result of a 4 K temperature difference between the spaces
on either side of the door (Barakat, 1987). Thus, the selected program should have
been able to model the inter zone heat transfer and its influence on energy use of
the building.
EnergyPlus (US Department of Energy, 2012) is a well-known and powerful multizone building simulation tool that was first released in 2001 by the US department of
energy as a replacement for the two existing simulation tools; BLAST and DOE-2
(Crawley, Hand, Kummert, et al., 2008). One of the main advantages of EnergyPlus
to its predecessors is that in EnergyPlus, heat balance simulation is coupled with
building systems simulation which means that at each time step (down to one minute)
the building loads which is calculated by a heat and mass balance module is passed
to building systems simulation module which has a variable time step (down to
seconds) where the system responses are calculated (Crawley, Lawrie, Winkelmann,
et al., 2001). The information from the building systems simulation module on the
loads not met by the system is fed back to heat and mass balance module and will
be reflected in the next time step of load calculations by adjusting the space
temperature if required and thus result in more accurate space temperature
50
predictions (Crawley, Lawrie, Winkelmann, et al., 2001). This integrated feature
discussed, allows realistic system controls to be modelled (Crawley, Lawrie,
Winkelmann, et al., 2001). In addition, EnergyPlus uses Air Flow Network (AFN)
model which allows simulation of inter- zone air flows and its influence on building
energy use.
2.7.3 Model calibration and validation
Although building simulation has been widely used during the past three decades to
investigate the effect of retrofit measures on energy savings and comfort, without
calibration of the base case model, results produced are not reliable (Westphal &
Lamberts, 2005). A large number of studies have shown discrepancies (which were
often significant) between the model predictions and measured building energy use
(Coakley, Raftery & Keane, 2014). Reddy (2006) defines calibrated simulation as
“the process of using an existing building simulation program and “tuning” or
calibrating the various inputs to the program so that observed energy use matches
closely with that predicted by the simulation program”. The purpose for calibration is
to ensure that the model could reasonably represent the thermal and energy
behaviour of the real building and thus achieve confidence in model predictions
(Westphal & Lamberts, 2005).
Coakley et al. (2014, p. 127) conducted an extensive literature review on current
approaches for building simulation calibration and classified them into four classes:
1. Calibration based on manual, iterative and pragmatic intervention.
2. Calibration based on a suite of informative graphical comparative displays
3. Calibration based on special tests and analytical procedures.
4. Analytical/mathematical methods of calibration
In addition, a number of techniques and tools were suggested by Coakley et al.
(2014) to support the calibration process of building simulation models such as track
and record the changes made to the model during the calibration process, in order to
improve the reliability and reproducibility of the calibration process, and conducting
sensitivity analysis. Sensitivity analysis can be simply described as varying the
model inputs and verifying the consequences of that change on the model outputs
(Calleja Rodríguez et al. 2013). A number of sensitivity analysis techniques such as
51
differential sensitivity analysis, Monte Carlo analysis and stochastic sensitivity
analysis have been used during the past two decades to understand the parameters
which should be carefully considered when modelling a building. Lomas & Eppel
(1992) were among the earliest to use these sensitivity analysis techniques for
dynamic thermal modelling.
ASHRAE (2009) discusses three methods to assess the accuracy of building
simulation models: empirical validation, analytical verification and Inter-model
comparison. In empirical validation results from building simulation are compared
with the data measured in real buildings (ASHRAE, 2009). There are various
published literature on validation mainly for residential buildings rather than large
commercial buildings; where conducting detailed measurements require
considerable efforts and costs (ASHRAE, 2009). A number of empirical validation
studies are summarized by Neymark and Judkoff (2002). One of the main challenges
researchers have been faced with to calibrate building energy models using
empirical validation is the lack of detailed empirical data particularly for residential
buildings which is necessary to understand the operational complexities and develop
better models (Buswell, Marini, Webb, et al., 2013). In majority of the cases, even
when the measured data is available, it has not been measured by end use and for
example the gas use measured include the use for space heating, hot water and
cooking which makes the calibration difficult. In addition, the measured data has also
an uncertainty and the differences observed between the models and measurements
will be due to errors in either set of data (ASHRAE, 2009). In Analytical verification
simulation results are compared with the results of a solved analytical solution
(ASHRAE, 2009). In Inter-model comparison simulated results are compared with
simulated results using other models (ASHRAE, 2009). This method is particularly
useful to test new models against the well established ones (Clarke, 2011).
Recent research suggests comparison of hourly data measured and predicted using
building simulation models rather than monthly or annual comparisons as it allows
better comparison of buildings’ dynamic energy characteristics (Yoon et al. 2003).
ASHRAE Guideline 14 (ASHRAE, 2002) which was initially developed to calculate
the energy saving potential of retrofit measures defines the acceptance criteria for
the calibration of building simulation models (Royapoor & Roskilly, 2015). When a
52
model meets these criteria, there is a reasonable agreement between measured and
simulated data and the model can be considered ‘calibrated’ (ASHRAE, 2002).
The guideline introduces two standardised statistical indices that should be used to
compare measured data and simulation results:
1. Mean Bias Error (MBE) (%): This is the sum of errors between measured and
predicted energy use for each hour. MBE captures the mean difference
between measured and predicted hourly energy use and thus is a good
indicator of the overall bias in the model (Coakley, Raftery & Keane, 2014).
MBE is calculated by equation (2-1):
𝑀𝑀𝑀𝑀𝑀𝑀 (%) =
𝑁𝑁
𝑝𝑝
∑𝑖𝑖=1
(𝑚𝑚𝑖𝑖 − 𝑠𝑠𝑖𝑖 )
𝑁𝑁
𝑝𝑝
∑𝑖𝑖=1
(𝑚𝑚𝑖𝑖 )
(2-1)
Where:
𝑚𝑚𝑖𝑖 = measured data point for each model instance ‘i’
𝑠𝑠𝑖𝑖 = simulated data point for each model instance ‘i’
𝑁𝑁𝑝𝑝 = number of data points at interval ‘p’
A limitation of this method is that positive and negative errors will cancel each other
when summed which means the positive bias compensate for negative bias.
2. Coefficient of Variation of Root Mean Square Error (CVRMSE) (%): This index
does not suffer from the cancellation effect mentioned above and allows one
to determine how well a model fits the energy use data by capturing offsetting
errors between measured and simulated data which were existed in MBE
method (Coakley, Raftery & Keane, 2014). CVRMSE (%) is calculated by
equation (2-2):
𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 (%) =
𝑁𝑁
𝑝𝑝
�(∑𝑖𝑖=1
(𝑚𝑚𝑖𝑖 − 𝑠𝑠𝑖𝑖 )2 /𝑁𝑁𝑝𝑝 )
(2-2)
𝑚𝑚
Where:
53
𝑚𝑚𝑖𝑖 , 𝑠𝑠𝑖𝑖 and 𝑁𝑁𝑝𝑝 are as defined in equation (2-2)
𝑚𝑚 = average of the measured data points.
According to ASHRAE Guideline 14 the acceptance criteria for hourly calibration of
building energy simulation models are: MBE 10% and CVRMSE 30%. These are 5%
and 15% for monthly calibrations.
It should be noted that “the current calibration criteria relate solely to the predicted
energy consumption and do not account for uncertainties or inaccuracies of input
parameters or the accuracy of the simulated environment (e.g. temperature profile)”
(Coakley, Raftery & Keane, 2014).
2.8 Summary
The main findings from the literature review conducted here with direct implications
on this study can be summarized as follows:
•
The majority of UK houses (above 97%) have central heating and a large
number of them are wet (hydronic) systems.
•
In recent years, the importance of having more than one heating zone in
dwellings have been realized in the UK as reflected in the Building
Regulations Part L1A for new dwellings which came into force from 1 October
2010. However, this does not apply to the existing dwellings.
•
A significant number of existing UK homes have poor levels of space heating
controls and there is a great potential for improvement. However, there is
evidence that the proportion of homes that have a range of controls may have
increased over recent years.
•
In theory, energy can be saved in houses by advanced space heating controls
as they could reduce the length of the heating period, the volume of house
which is heated and the heating demand temperature. A number of studies
have used models to prove that. However, there are a number of studies
which shows that the predicted savings could be hardly achieved in real world
settings.
•
There is a lack of a robust and repeatable methodology for measuring the
energy savings which could be achieved by enhanced heating controls.
54
•
There are a number of products currently available in the market which could
be used to establish ZC in the existing houses with wet central heating
systems. However, the energy and thermal comfort implications of using them
instead of conventional space heating controls are unknown.
•
Dynamic thermal modelling could be used as the most detailed tool for
calculating the energy savings of ZC in different houses. However,
reconciliation of the model predictions with the measured data is crucial
before the results could be trusted.
55
3 Overview of the methodology and test
houses
3.1 Introduction
This chapter consists of two main sections. In the first section (section 3.2), an
overview of the methods adopted in this study to achieve its aim and objectives is
provided. The methods adopted were based on the use of a pair of full size test
houses with synthetic occupancy. The second part of this chapter (section 3.3),
describes these houses and discusses the characterisation tests which were carried
out in them to evaluate and compare their thermal performances. Section 3.4
summarizes the work presented in this chapter.
3.2 Overview of the methodology
This study combined space heating trials to measure the energy savings of zonal
space heating controls, dynamic thermal modelling and a wider scale evaluation in
order to achieve the aim and objectives described in section 1.3. This section
provides an overview of these three components and discusses how they were
interconnected. Further details of the methods for the trials, modelling and wider
scale evaluation are given in chapters 4 to 7.
3.2.1 Overview of the space heating trials
The purpose of the space heating trials was to achieve the first objective of this study
which was to measure the energy savings (if any) of applying zonal space heating
control (ZC) in a UK house.
It was decided to measure the energy savings of a house when its space heating is
controlled by ZC compared to when it is controlled with a conventional system in
comply with Building Regulations Part L1B (here referred to as Conventional Control
(CC)). The reason for this choice was that although space heating is not controlled in
the same way in every UK house, all new homes need to comply with the regulations.
56
In addition, nearly half of UK existing homes already have such sets of controls
according to EFUS 2011 (BRE, 2013).
Such comparison was not ideal to be conducted in a single house. Unless the house
was built in a fully controlled environment such as in an environmental chamber, the
changes in the weather for the periods when the house is controlled by ZC and then
by CC could largely influence the energy consumption of the house and any potential
energy savings measured. Therefore, this comparison was conducted using a
matched pair of side-by-side test houses which will be fully described in section 3.3.
In order to ensure that the two houses had a similar thermal performance, a side-byside co-heating test and air tightness tests were carried out in both prior to the space
heating trials (see section 3.3.6).
Two space heating trials (HT1 and HT2) each lasted four weeks were conducted
during the winter of 2014. During the HT1 and HT2, the same synthetic, yet realistic,
occupancy schedule was applied to both houses (see section 3.3.5). The two test
houses were each equipped with the same new wet central heating system. In one
house the space heating was controlled conventionally (CC) in compliance with
requirements in UK Building Regulation Part L1B for existing dwellings, whereas in
the other house ZC was used to heat the rooms only when they were ‘occupied’. In
the HT1 ZC was applied to House 1 and CC to House 2 then, for the HT2, the
heating control strategies were swapped with CC in House 1and ZC in House 2. This
was done to negate any small differences between the thermal performances of the
building fabric of the two test houses (see section 3.3.6). The energy use for space
heating and indoor air temperatures of the two houses were measured and
compared.
The potential energy savings of a house heated by ZC instead of CC were quantified
for one particular house in one location and over one winter period during the space
heating trials. However, conducting further experimental studies in order to measure
the annual energy savings of ZC or energy savings in houses located in different
regions of the UK was not possible considering the time, budget and scope of this
work. Instead, dynamic thermal modelling was used as an alternative method to
assess the potential energy savings of ZC for better insulated houses and those
exposed to different climate.
57
3.2.2 Overview of the dynamic thermal modelling
EnergyPlus was used to create a Dynamic Thermal Model (DTM) of the building
envelope of the test houses according to the existing knowledge of the test houses
and information obtained via detailed house audits (chapter 5). In order to verify the
building envelope model, the co-heating test was simulated and the predicted energy
use during the co-heating test period was compared to the measured energy use
according to ASHRAE guideline 14 acceptance criteria for hourly calibration of
building energy simulation models (chapter 6, Section 6.2). The effects of employing
two different air flow modelling strategies (i.e. scheduled natural ventilation (SNV)
and Air flow network (AFN)) on model predictions were studied (model 1 and 2,
Table 3-1).
The verified building envelope model was then used and the HT1 (the period when
data was successfully measured continuously for the period of 4 weeks) was
simulated (chapter 6, section 6.3). Similar to the co-heating test, predicted energy
use and indoor air temperatures using the two different air flow modelling strategies
were compared to the measured data (model 3 and 4, Table 3-1). Any observed
discrepancies between the predictions and measurements were then explored and
modelling limitations and potential solutions were discussed.
Based on the discrepancies observed between the predictions and measurements, a
sensitivity analysis was conducted to investigate the effects of a number of
parameters on improving model predictions of energy use and indoor air
temperatures (chapter 6, section 6.4). Based on the results from the sensitivity
analysis, a refined model was constructed and assessed against the ASHRAE
guideline 14 calibration criteria (Refined model, Table 3-1).
The second objective of this research was achieved as the refined model which was
calibrated using the measurements was able to closely predict the energy savings of
applying zonal control in the same house under the same conditions.
58
Table 3-1: Summary of the DTMs created during the modelling campaign
Model
Experiment
Heating
Occupancy
Air flow
Heating
Heating
modelled
modelling
control in
control in
(Yes/No)
strategy
House 1
House 2
Constant air
Constant air
temperature
temperature
Constant
Constant
temperature
temperature
1
Co-heating
Electrical
No
SNV
2
Co-heating
Electrical
No
AFN
3
HT1
Wet
Yes
SNV
ZC
CC
4
HT1
Wet
Yes
AFN
ZC
CC
HT1
Wet
Yes
AFN
ZC
CC
Refined
model
3.2.3 Overview of the wider scale evaluation
Two different approaches were undertaken in order to achieve the final objective of
this study which was to explore how the energy savings would vary in better
insulated houses and in different UK locations. In the first approach, an empirical
model was developed using the Heating Degree Day (HDD) method based on the
experimental data collected from the heating trials (chapter 7, section 7.2). The
empirical model was used to extend the measured gas consumptions with CC and
ZC to annual values, and to make an initial estimate of the effect of the weather in
different parts of the UK on the potential savings. The empirical model estimated the
annual gas savings of ZC and the corresponding cost savings. The model was also
used to estimate the pay back periods of upgrading a same size house with
conventional heating controls to zonal heating control in different UK regions.
In the second approach, the calibrated DTM was used to investigate energy savings
of applying ZC instead of CC in the same house for different regions of the UK using
the same weather data as in the empirical model (chapter 7, section 7.3). The cost
benefits were also recalculated based on the DTM results. The predictions of the
DTM were then compared against the predictions of the empirical model. The
59
potential reasons for the differences observed in the predictions of the DTM and
empirical model were then discussed.
Finally, the calibrated DTM was employed and the effects of better insulated building
envelope on the potential energy and cost savings of ZC in different regions were
investigated (chapter 7, Section 7.4).
3.3 Test houses
This section introduces the test houses which were used for this study and describes
their geometries, construction materials and properties, heating systems and their
synthetic occupancy.
3.3.1 Description of the test houses
Loughborough University’s Matched Pair of 1930s houses (LMP1930) are a pair of
adjoining semi-detached homes which were used for this research. The houses,
which are typical family homes of the 1930s period, are located in the East Midland’s
town of Loughborough, UK (Figure 3-1).
LMP 1930 test houses
Figure 3-1: Bird’s-eye view of the test houses, their surrounding buildings and
vegetation (Google Maps, 2015)
Semi-detached houses are the most common house type in England representing 26%
of the housing stock with over 30% of them built between 1919 and 1944
(Department for Communities and Local Government, 2001). However, semi60
detached house layouts and construction methods remained largely unchanged from
the 1930s through to as late as the 1960s (Rock, 2005) and so the house layouts
could be representative of a larger proportion of homes.
The test houses had the same geometry, size and construction and had not been
significantly modified since they were built (Figure 3-2).
Figure 3-2: Views of the two test houses: front, south-facing (left) and back, northfacing (right)
The fronts of the houses faced south and the windows were unshaded except for
those on the West facade of House 1 and the East facade of House 2; these
windows were covered by 50 mm of Polyisocyanurate (PIR) insulation boards from
the inside of the houses to minimize the effect of different morning and afternoon
solar heat transfer to the two houses (Figure 3-3).
61
Figure 3-3: The West facing windows in the House 1 which were covered by 50 mm
PIR insulation boards
Original open fire places were located at the party wall of the two houses in the living
room and dining room of each house (Figure 3-4). These were blocked to avoid
unnecessary air leakage.
62
Air vents were blocked using adhesive tape
Figure 3-4: Blocked original open fire places located in the living room of House 1
3.3.2 Building geometry
Internal dimensions of the test houses were measured at the beginning of this study
and their floor plans were drawn (Figure 3-5). Each house had a total floor area of
91.2 𝑚𝑚2 (including both floors) and a total volume of 240 𝑚𝑚3 . Each house had three
rooms located on the ground floor including living room, dining room and kitchen plus
a hallway and four rooms on the first floor including three bedrooms and a bathroom
plus a WC and a hallway (Figure 3-5).
63
4.5 m^2
4.5 m^2
16.5 m^2
5.3 m^2
5.3 m^2
16.5 m^2
15.2 m^2
15.2 m^2
4.5 m^2
4.5 m^2
5.4 m^2
5.4 m^2
16.0 m^2
10.4 m^2
10.4 m^2
16.0 m^2
14.8 m^2
14.8 m^2
Figure 3-5: The floor plans of the test houses 5 with the floor area of each room
5
The floor plans (here and throughout the thesis) are schematics and not to scale. The blocked fire
places were not shown
64
3.3.3 Construction materials and properties
Test house audits were conducted in the test houses to understand details of the
construction materials used in the test houses. The areas and thicknesses of the
bricks, cavity air gap, plaster, floor boarding, carpets and doors were measured and
the materials used for each construction element were noted. A summary of
construction elements, their areas and their calculated U-values according to RdSAP
(BRE, 2014) are presented in Table 3-2.
Table 3-2: Summary of construction elements of the test houses, their areas and
calculated U-values according to RdSAP (BRE, 2014)
Total
Element
Description
Area
(m2 )
External walls
(W/m2 K)1
Brick Cavity
81.6
1.6
Suspended Timber
40.2
0.8
Floor (kitchen)
Solid floor
5.4
0.7
Roof
Pitched roof covered with clay tiles
45.62
2.3
Windows
Single glazing with wooden frames
20.7
4.8
Entrance doors
Wooden
3.4
3.0
Party walls
Brick Cavity with closed air vents
42.2
0.5
Internal partitions
Solid Brick covered with gypsum plaster
56.1
2.1
Floor (except
kitchen)
1
U-value
Approximate U-values from UK Government’s Standard Assessment Procedure for energy
rating of the existing dwellings (RdSAP) (BRE, 2014).
2
The horizontal, not pitched, area.
As found by the test house audits, both houses had 100% single glazed windows,
un-insulated cavity external walls, and no floor or loft (attic) insulation (Table 3-2). In
contrast, many UK homes have been refurbished, such that in 2011, of the 3.6
million UK homes built between 1919 and 1944, only 4% had no loft insulation, only
6% were still fully single glazed, and only 28% had uninsulated cavity walls
(Department for Communities and Local Government, 2012). Therefore, the test
65
houses would represent the un-furbished interwar houses in the UK which account
for about 180 thousand homes.
The ground floor of LMP1930 test houses was mainly suspended timber ventilated
with outdoor air using six cast iron air bricks located around the perimeter of each
house (Figure 3-6). The underfloor void was 0.2 m deep (from the bottom of the floor
boarding to the ground). In the UK, naturally ventilated floors are used to control the
moisture from the ground. The Kitchen floors were constructed from solid concrete.
Figure 3-6: Floor plan of the ventilated subfloors existed below the ground floor of
each house and the location of air bricks
Figure 3-7 shows examples of test house inspections when a part of carpet and floor
boards were temporarily removed in a bedroom to measure the thicknesses of the
floor materials6 (Figure 3-7 (a)) or the inspection in the loft (attic) space where no
insulation was found (Figure 3-7 (b)).
6
The photo was taken in an adjacent house which was built in the same year by the same builder
66
Figure 3-7: Examples of test house inspections for understanding details of
construction materials: (a) construction of internal floors; (b) loft (attic) space
construction (before removing debris)
Figure 3-8 shows the inspection of cavity walls and underfloor void using a
borescope. No insulation was found in the cavity (Figure 3-8 (a)). Acquiring better
knowledge of the ground surface material (under the suspended floor), by taking
photos and videos using a borescope inside the air bricks (Figure 3-8 (b)), was not
successful due to filth existed under the suspended floors.
Figure 3-8: Borescope investigation at the test houses: (a) exploring external wall
cavity; (b) exploring subfloor construction through air bricks (Photos by Stephen
Porritt)
67
3.3.4 Heating system
Each house was equipped with an identical low pressure hot water (LPHW) wet
central heating system consisting of a 30 kW condensing combination boiler
(Worcester Greenstar 30 CDi combi) located in the kitchen, identical Eco-Compact
radiators sized to suit each room, and a Horstmann wireless C-stat 17-B
programmable room thermostat located in the hallway. Drayton RT212 TRVs were
installed on all radiators apart from the ones located in the hallway of each house.
The boilers were less than seven years old.
Rated capacities of radiators which were selected according to each radiator’s height
and width from their manufacturer’s data for a 50 K temperature difference between
the room’s air and mean water temperature were reported in Table 3-3.
Table 3-3: Rated capacities of the radiators in the LMP1930 test houses according to
their manufacturer’s data for 50K temperature difference
Room
Radiator rated capacity (W)
Living room
1372
Dining room
882
Hallway ground floor
1568
Bedroom 1
1568
Bedroom 2
1764
Unoccupied room
980
Bathroom
588
3.3.5 Synthetic occupancy
Both houses were equipped with synthetic occupancy to represent heat gains from
people, domestic equipment and lighting, internal door opening/closing and window
blind operation in both houses.
Reviewing previous published reports and papers, it was found that a wide range of
occupancy profiles have been used but mostly without any detailed information
about the sources of their assumptions. Capon & Hacker (2009) assumed partial
daytime occupancy and full evening and weekend occupancy for a case study house
68
without presenting any specific schedules. Hacker et al. (2008) provided more
detailed occupancy profile for a family in a case study dwelling assuming the house
is occupied all the time with one adult at work from 08:00 to 18:00. Adult bedrooms
were assumed to be occupied from 23:00 to 07:00 and children bedrooms from
20:00 to 07:00. The sources of these assumptions are not known. A relatively old
report by Building Research Establishment (Allen and Pinney, 1990) provides
occupancy periods for each room to be used in modelling. However, the profiles
were constructed more than 20 years ago when due to absence of personal
computers and TVs or game consoles, it cannot reflect realistic occupancy patterns
of nowadays (Porritt, 2012).
Porrit (2012) was the only recent study that was found to report a detailed occupancy
schedule for each room. Porritt (2012) derived two occupancy profiles using data
from the Time Use Survey 2000 which recorded, in ten minutely slots, the daily
activity of over 6000 households as a representative sample of the population of
households and individuals in the UK (ONS, 2002) (Table 3-4).
Two occupancy profiles were assumed by Porritt: an occupancy profile for a family
consisted of 2 working adults and school age children (number of children depending
on house size), who are out of dwelling during the day time and an occupancy profile
that assumed two elderly residence who occupy the dwelling all the time.
Although Time Use Survey 2000 had detailed information regarding the type of
activity and whether it happened inside or outside the house, it did not contain any
detail regarding which room the activity had taken place. Therefore, Porritt’s
occupancy profiles were based on a number of assumptions:
•
When sleeping is recorded the occupant is in their bedroom.
•
When children recorded that they are using computer or watching TV, it was
assumed that they are in their bedrooms.
•
When adults recorded that they are using a computer it was assumed that
they are in their bedroom and, when watching TV, are in their living room.
•
Cooking activities were happening in the kitchen.
•
Eating activities was happening in the dining rooms in the terraced and semidetached houses and in the living rooms in Flats as the kitchen in these
69
house types are small and does not allow the occupants to eat in the kitchen.
In detached homes, eating was assumed to happen in the Kitchen where
larger space would let the occupants to use it for dining.
Moreover, Porritt (2012) assumed slightly different occupancy patterns for the
weekends compared to the weekdays for a typical family. In weekends, the bedroom
occupied periods were extended to consider morning lie-ins for some occupants.
The chosen occupancy profile for the two houses in this study represented a family
of two working adults, and two school-aged children. The ‘occupied hours’ for each
room was set according to Porritt (2012) (Table 3-4).
Table 3-4: Weekday and weekend ‘occupied’ hours of each room
Room
Weekday ‘occupied’ hours
Weekend ‘occupied’ hours
Living Room
18:00-22:30
18:00-22:30
08:00-08:30
09:30-10:00
17:00-18:00
17:00-18:00
07:30-08:00
09:00-09:30
16:00-17:00
16:00-17:00
Dining Room
Kitchen
19:00-22:30
22:30-08:00
Bedroom 1
08:30-09:00
16:00-17:00
Bedroom 2
Bathroom
Bedroom 3
19:00-22:30
22:30-09:30
10:00-10:30
16:00-17:00
22:30-07:30
22:30-09:00
07:30-08:00
09:00-09:30
08:30-09:00
10:00-10:30
19:00-20:00
19:00-20:00
-
-
Bedroom 1 was assumed to be used only by the two children and Bedroom 2 by the
two adults. It was assumed that bedroom 3 was unoccupied all the time. Although
the occupancy patterns of the rooms were the same for all the weekdays, for the
70
weekends, bedroom occupied periods were extended by 1.5 hours thus shifting
morning gains in other rooms 1.5 hours forward compared to the weekdays. During
the day time (09:00 to 16:00 hrs on weekdays and 10:30 to 16:00 hrs on weekends)
all the occupants were assumed to be out of the house. Evening occupancy patterns
were the same for all days of the week including weekdays and weekends (Table 34).
The occupancy profile was mimicked in each house using a z-wave smart home
controller: Vera 3 (Vera control Ltd, 2014) (Figure 3-9). Each house was equipped
with its own Vera 3. A z-wave network was established in each house by linking Vera
3 to a number of z-wave enabled smart plugs and motor controllers which allowed
Vera 3 to send on/off commands to each plug or motor controller.
Figure 3-9: Z-wave smart home controller used in each house for synthetic
occupancy during the HT1 and HT2
Tables published by the American Society of heating Refrigeration and Air
conditioning Engineers (ASHRAE) (ASHRAE, 2009) were used to estimate the heat
output rates from occupants, equipment and lighting. Similar to Porritt (2012), each
house was assumed to have a refrigerator in the kitchen which was rated at 60 W, a
150 W modern LCD TV in the living room and a computer or game console in the
children’s bedroom with 100 W heat output. Cooking gains were 1.6 kW for period of
one hour during the evening, 160 W for the 30 minutes breakfast period and no
cooker use at lunch time (Table 3-5).
71
Table 3-5: The timing and magnitude of internal heat gains presented in different
rooms of both houses during each trial
Room
Kitchen
Time of day
weekday
Time of day
weekend
07:30-08:00
09:00-09:30
16:00-17:00
16:00-17:00
18:00-19:00
18:00-19:00
19:00-22:30
19:00-22:30
08:00-08:30
09:30-10:00
Living
Room
Dining
Room
Bedroom 1
Bedroom 2
1
Gain source: estimated rate
(W)
Total
estimated
gains
(W)
Total
actual
gains
(W)
Morning cooking: 160
Adult cooking: 189
Fridge: 60
409
400
Evening cooking: 1600
Adult cooking: 189
Lighting: 54
Fridge: 60
1903
1900
60
60
556
580
396
400
448
460
478
480
160
160
Lighting: 30
1
Children seated: 80 *2
190
200
Fridge: 60
TV: 150
Lighting: 30
1
Adult seated: 108*2
1
Children seated: 80*2
TV: 150
Lighting: 30
1
Adult seated: 108*2
1
Hot food: 18 *4
1
Adult seated: 108 *2
1
Children seated: 80 *2
1
Hot food: 18 *4
Lighting: 30
1
Adult seated: 108 *2
1
Children seated: 80 *2
1
Children seated: 80 *2
17:00-18:00
17:00-18:00
08:30-09:00
10:00-10:30
16:00-17:00
16:00-17:00
19:00-20:00
19:00-20:00
Lighting: 30
Child seated: 80
110
120
20:00-22:30
20:00-22:30
Lighting: 30
1
Children seated: 80 *2
Computer: 100
290
300
22:30-08:00
22:30-09:30
Children sleeping: 54*2
108
120
144
140
22:30-07:30
22:30-09:00
1
Adult sleeping: 72 *2
1
Multiplied by the number of people
72
The total amount of heat required at any time and in each room was delivered using
a series of incandescent, halogen and low energy light bulbs, oil-filled radiators or
fan heaters (Table 3-6). The light bulbs were used instead of other potential heat
emitters with small outputs due to the university’s health and safety policies which
did not allow the researcher to use any other type of heaters in the unoccupied test
houses. However, similar to any other heat emitter, all the electricity used by the light
bulbs would end up as heat in the space.
Table 3-6: Number of heat emitters and their nominal outputs used to deliver internal
heat gains in each room
Heat emitters
Number of heat emitters in each room
and their
nominal heat
Living room
Dining room
Kitchen
Bedroom 1
Bedroom 2
400W
1
1
-
-
-
60W
3
1
1
5
2
20W
-
1
-
2
1
-
-
1
-
-
-
-
1
-
-
output (W)
Light bulbs
Heaters
Oil-filled radiator
(400W)
Fan heater
(1500W)
All the light bulbs used were placed on tripods for safety reasons (Figure 3-10). The
location where the heat emitters were located and the height of the light bulbs on the
tripods were matched between each room of the two houses. All the wire runs on the
floors were covered by adhesive tape to avoid trips or falls.
73
Shielded temperature sensor
400 W light bulbs
(Only one of them was in use)
Radiator surface
temperature sensor
60 W light bulbs
Figure 3-10: light bulbs with different outputs used to produce the internal heat gains
in the living room of House 2
The heat emitters were controlled from the home automation controller to produce
the repeating weekday and weekend total heat gain profiles (Figure 3-11). The zwave enabled smart plugs were AN148 by Everspring and used to switch on and off
heat emitters in order to produce internal heat gains in different rooms. The actual
total heat gains produced in each room were identical in each house and were within
±10% difference of the total estimated values calculated from the ASHRAE tables
(Table 3-5) (ASHRAE, 2009). This was due to the sizes of heat emitters that were
available (Table 3-6). Variations in the mains electricity supply voltage also resulted
in small differences in the heat gains achieved; however, this discrepancy was also
the same for both test houses.
74
Figure 3-11: Total actual heat gains in different rooms of a house during a weekday
All windows were fitted with internal roller blinds. The roller blind fabric was cut to the
appropriate sizes to fit each window. The fabric used was thin (1 mm thick), with
closed weave and had a grey colour. Blind rotary motors which were controlled by Zwave motor controllers (DBMZ Hunter Douglas) were used to move the blinds up
and down. The roller blinds in the living room and bedrooms 1 and 2 were opened
every weekday at 08:00 hrs and at 09:30 hrs on Saturday and Sunday. All blinds
were closed at 16:00 hrs every day. The blinds in the dining room, bathroom and
kitchen which all were facing north and the unoccupied spare bedroom were always
remained closed.
The internal doors were operated using electrical actuators controlled by the motor
controllers which were receiving commands from the home automation controller.
The internal doors of the living room, dining room and bedrooms 1 and 2 (Figure 3-5)
were closed when the room was ‘occupied’7 and open otherwise (Table 3-4). The
internal door of the kitchen was open at all times whilst the doors of the unoccupied
7
Throughout, ‘occupied’ means that the room had synthetic occupants present.
75
spare bedroom (bedroom 3), the bathroom and the two doors to the outside, were
closed at all times.
In order to avoid any possibility of entrapment, the internal doors were needed to be
manually operable as well as automatically. Therefore, a mechanism was designed
by attaching the actuators to support rods using cable ties to firmly hold the actuators
in an appropriate position (Figure 3-12). In order to open a door, the actuator chain
pulled the door using a rope which was attached to the actuator chain from one side
and to the door from the other side (Figure 3-12). When a door needed to be closed,
the actuator chain was released and the door closer installed on the top of each door
pushed the door back to its fully closed position.
Support rod
Actuator
Motor controller
Figure 3-12: Internal door operation mechanism used in the test houses
Aspects of occupancy that were not mimicked include outside door openings,
window opening, domestic hot water use, bathroom heat gains and occasional
electrical usage such as dish washers, clothes washing and kettles. Windows and
doors could not be simulated due to security concerns. The potential heat gains from
hot water use and occupants in the bathroom were considered to be negligible as
any heat produced was assumed to be transferred directly to the outside by extract
fans or window openings or drainage. Most importantly however, as both houses
76
were operated in the same manner, their heating energy demands were not
differently affected by the occupancy.
The assumption of blinds in the dining room, kitchen and bathroom being always
closed might not reflect the behaviour of real occupants. The blinds will reduce
radiative and convective heat losses but, as these rooms were all facing north, the
closed blinds have negligible effect on solar gains. The net effect is the same in both
houses.
All the synthetic occupancy equipment had been tested both in the laboratory and in
situ prior to the start of the heating trials. In addition, Internet Protocol (IP) cameras,
which were located in the living room of each house, were used to check the
operation of some synthetic occupancy equipment such as internal door or window
blinds opening/closing and the lighting status (Figure 3-13).
Figure 3-13: IP camera which was used in the living room of a test house to check
the operation of synthetic occupancy devices
3.3.6 Experimental characterisation of the test houses
Characterisation tests were conducted to assess and compare the thermal
performance of the test houses. These tests consisted of a standard blower door test
in accordance with ATTMA Technical Standard L1 (2010) and a standard co-heating
77
test as described by Wingfield et al (2010). No occupancy was simulated during the
characterisation tests.
The blower door tests were carried out on the same day (3 July 2013) for both
houses (Figure 3-14). During the tests, the openings of the passive ventilation,
extractor fan in the kitchen and original open fire places were sealed and all drainage
traps were filled by water, as required by the standard test protocol. Thus the
measured air leakage rate does not measure the in-use ventilation rate of the
dwelling.
Figure 3-14: The blower door tests set up during the test in House 1
The front door of the test houses were arc shaped which did not allow the
rectangular shaped blower door to be fitted to the front doors. Therefore, as it can be
seen in Figure 3-14, a piece of wood were carefully cut and fitted above the
rectangular blower door to cover all the open area of the front door.
The co-heating tests were conducted simultaneously in the two test houses during
the period of 23 November to 1 December 2013. Seven electrical fan heaters which
were set on a level to emit a nominal heat output of 1500 W were used in each
78
house during the co-heating test (Figure 3-15). Four of them were placed in the
ground floor rooms (i.e. living room, dining room, kitchen and hallway) and three of
them were placed in the first floor; one in each bedroom (Figure 3-15).
79
Circulation fans
Electrical fan heaters
Figure 3-15: The location of fan heaters and circulation fans during the co-heating
test
80
These electrical fan heaters were used to maintain a nominal internal air temperature
of 25ºC in each room for a period of 9 days, plus 2 days of pre-conditioning. The two
days pre-conditioning period was considered in order to allow the houses to achieve
the steady state conditions required for the test. Circulation fans were used in each
room to mix the air in the whole house and reduce stratification (Figure 3-16); the
doors to all rooms were left open.
The heat output of each fan heater was controlled using a thermostat located in the
centre of each room, 1.5 m above the floor level. The thermostats were 4-wire
PT100 resistance thermometers. They were shaded from direct sunlight and the hot
air from the fan heaters (Figure 3-16).
The thermostat was connected to a PID temperature control unit (InstCube 3216 L
SSR Temperature Control Unit by TMS Europe Ltd) to switch the fan heater on and
off using their PID control algorithm to maintain the constant air temperature of 25ºC.
The electrical energy supplied to each house was measured at the meter (see
section 4.2).
Thermostat
Circulation fan
PID temperature
control unit
Fan heater
Figure 3-16: The co-heating test set up in the living room of House 1
81
The internal air temperature of every room was measured at 1 minute intervals using
calibrated thermistors (see section 4.2). Minutely values for outdoor air temperature
and hourly values for global horizontal solar irradiance during the test period were
sourced locally (see section 4.2).
The “Siviour” linear regression method described in Butler & Dengel (2013) was
used to calculate the solar-corrected heat loss coefficient of each house by plotting
Q�
S
ΔT against �ΔT for each day (i.e. 23 November to 1 December 2013) of the coheating test (Figure 3-17) where:
Q: Average daily measured power consumption (W)
ΔT: Average daily air temperature difference between indoor and outdoor (ºK)
S: Average daily global horizontal solar irradiance (W/𝑚𝑚2 )
450
400
y = -9.9x + 382
350
y = -11.8x + 361
Q / ΔT (W / K)
300
250
House 2
200
House 1
Linear (House 2)
150
Linear (House 1)
100
50
0
0
0.5
1
1.5
2
2.5
3
3.5
S / ΔT (W / m² K)
Figure 3-17: Siviour regression analysis for the two test houses
The resulting slope of the plot is the solar aperture R in 𝑚𝑚2 and the Y intercept is the
solar corrected total heat loss coefficient in W/K.
82
The results of the characterisation tests are presented in Table 3-7 and show that
the two houses had very similar overall heat loss coefficients that were within 6%.
This is a remarkably similar performance, especially given the uncertainty of coheating tests, which may be greater than 10% (Butler & Dengel, 2013). In National
House Building Council’s (NHBC’s) review of co-heating test methodologies (Butler
& Dengel, 2013) solar corrected whole house heat loss coefficients found by 6
independent co-heating tests conducted by different project partners ranged from -17%
to +11% of the calculated steady state heat loss based on as-built dimensions and
specific fabric element U-values and infiltration rates (BRE, 2014). It should be noted
that although Wingfield et al. (2010) recommends the use of at least one week’s
worth of co-heating test data and 9 days’ worth of data was used in this study, a
longer period could have led to slightly different heat loss coefficients.
The blower door test results also showed air leakages for the two houses were within
3% (Table 3-7). Full reports of the blower door tests were presented in Appendix A.1.
An estimate of the background air infiltration rate, for the houses in the blower door
test state (with the large purpose made openings blocked), can be achieved by
dividing the air change rate at 50 Pa (N50) by 20 (CIBSE, 2000); which gives 1.07
ACH and 1.1 ACH for Houses 1 and 2 respectively. The test houses were less
airtight than the average for UK houses of a similar age as reported by Building
Research Establishment (Stephen, 2000): the mean air leakage rate of 58 dwellings
built between 1930 and 1939 was 15.9 ACH at 50 Pa.
Table 3-7: Summary of the house characterisation test results
Performance measure
House 1
House 2
% difference
382
361
+5.6%
20.761
21.392
-2.9%
Infiltration rate (ACH)
1.07
1.1
-2.9%
Solar aperture (𝑚𝑚2)
9.9
11.8
-16%
Total heat loss coefficient
(W/K)
Air leakage
(m³/ h*m² Surface area at
50Pa)
1
Equals to 21.5 ACH at 50 Pa
2
Equals to 22.1 ACH at 50 Pa
83
3.4 Summary
This chapter provided an overview of the space heating trials, modelling and wider
scale evaluation campaigns which formed the methodology of this research. It also
described the LMP1930 test houses used in this study, their geometries,
construction materials, heating systems and the synthetic occupancy regime. In
addition, the characterisation tests which were carried out in the houses were
described and their results were discussed. The building envelope of the two houses
showed a close thermal performance which was considered suitable for a side-byside comparison of the energy performance of different heating control strategies.
84
4 Space heating trials
4.1 Introduction
This chapter describes the two space heating trials (HT1 and HT2) and present their
results. It starts with describing the test houses’ instrumentation set up (Section 4.2)
and space heating control strategies (Section 4.3) during the HT1 and HT2 trials.
The chapter then discusses the results by comparing the indoor air temperatures
(Section 4.4), heating demand, boiler efficiencies and fuel use (Section 4.5) of the
two houses. The chapter finishes by providing a summary of the findings from the
space heating trials in section 4.6.
4.2 Instrumentation
Identical instrumentation was used in each house. Indoor air temperature was
measured throughout the testing period, in each room, at 1 minute intervals, using U
type thermistors. These were located in the volumetric centre of each room using
tripods and were shielded from any direct sunlight using aluminium sheets
(Figure 3-10).
The surface temperature of each radiator was measured at 10 minute intervals using
I-button temperature loggers (Hindman, 2006). They were attached to the centre of
each radiators surface using adhesive tape (Figure 3-10).
Boiler heat output was measured at 1 minute intervals using a heat flow meter
consisting of Supercal 531 energy integrator (Sontex SA, 2014a) programmed for
10Wh per pulse, Superstatic 440 flow meter (Sontex SA, 2014b) installed at the
return water going to the boiler and Pt500 temperature sensors inserted into ½” BSP
pockets both at supply and return water pipes to the boiler (Figure 4-1). The active
measuring temperature sensor tips were placed in the centre of the pipe cross
section and the water pipes were insulated around the area where the temperature
sensors were inserted according to the manufacturer’s guidance to increase the
accuracy of measurements. Supercal 531 calculated the heat captured into the water
(i.e. boiler heat output) from the mean flow rate, the water temperature difference
and the heat coefficient.
85
Supercal 531
Energy integrator
Superstatic 440
flow meter
Pt 500 temperature
sensors
Figure 4-1: Equipment used to measure boiler heat output in the test houses;
consisted of flow meter, temperature sensors and energy integrator
The volume of gas consumption for the boiler was measured every 10 minutes at the
supply company gas meter of each house using an intrinsically safe pulse counter.
This consisted of a Technolog Zmart Link gas flow transmitter which transmitted the
pulse outputs from the gas meter to a gateway used for data recording and
monitoring 8. The gas pulse data could be then downloaded via web. The gas
consumption (in kWh) was then calculated using the natural gas calorific value of
39.6 𝑀𝑀𝑀𝑀𝑚𝑚−3 (DECC, 2014b).
8
The gas flow transmitter and the gateway were sourced from Loughborough University’s DEFACTO
(Digital Energy Feedback and Control Technology Optimisation) project partners and were not
commercially available in the market
86
Gas pulse
transmitter
Gateway
Figure 4-2: Equipment used to measure and record volume of gas use in the test
houses
Electricity consumption was recorded every 5 minutes using LED pulse loggers
(Enica Ltd, 2014) installed on the supply company electricity meter of each house,
and at the individual appliance level using Plogg energy meters (Constable & Shaw
2011). This provided a measure of the heat delivered to the houses as electricity. All
supplied electricity emerges as heat in the house.
Outdoor air temperature was measured every minute using a thermistor located
adjacent to the houses but far enough away to avoid any thermal effects from the
external walls. The thermistor was shaded from direct solar radiation and the sky and
was shielded to protect it from rain and moisture (Figure 4-3).
Data logging at each house was carried out using a DT 85 Datataker data logger
with in-built web server. The recorded data could be accessed online and
downloaded at any time. Data collected was checked on a daily basis during the
space heating trials. Checking the data on a daily basis was particularly useful on an
occasion during the HT2 when it was found that there is no heat output from the
boiler in one of the test houses, and immediate inspection of the test house revealed
a leak in the pipes; which was quickly fixed with minimum loss of testing time and
data.
87
Figure 4-3: The location of temperature sensor used to measure outdoor air
temperature and its shielding
Hourly global horizontal solar irradiance was sourced from the MIDAS Land Surface
Observation database at the British Atmospheric Data Centre (BADC) operated by
the UK Meteorological Office (UK Meteorological Office, 2012). The nearest weather
station was Sutton Bonington located 8 km away from the test houses.
All the temperature sensors used had been calibrated by the researcher before and
after the experiments using a controlled water bath calibrator (Figure 4-4).
88
Figure 4-4: The calibration of U type thermistors using water bath calibrator
The accuracy of the equipment and uncertainty in values used in this study is
indicated in Table 4-1.
89
Table 4-1: Accuracy of the equipment and uncertainty in values used
Equipment /
values used
U type
thermistors
Parameter
measured /
calculated
Accuracy /
uncertainty
Source
Manufacturer
Air temperature
±0.2ºC
stated
accuracy
Manufacturer
Data logger
Air temperature
0.1%
stated
accuracy
I-buttons
Radiator surface
temperature
Manufacturer
±0.5ºC
stated
accuracy
National
Gas meter
Volume of gas
±2%
Measurement
Office (2014)
Gas calorific
value
Energy of gas
±1.5 MJ𝑚𝑚−3
Buswell
(2013)
Manufacturer
Heat meter
Boiler heat output
±2%
stated
accuracy
4.3 The control strategies
Two space heating trials (HT1 and HT2) were conducted in the test houses. HT1
was conducted continuously from 16 February to 15 March 2014. HT2 started on 18
March 2014, was stopped for 1 week due to equipment failure (9 to 15 April) and
then continued afterwards until 21 April 2014. Thus each heating trial consisted of 4
weeks of reliable data including 20 weekdays and 8 weekend days.
90
The CC system consisted of the programmable room thermostat (PRT) in the
hallway and TRVs on all radiators apart from the one located in the hallway
(Figure 4-5). This enabled the heating system to be operated on a daily schedule
using the PRT. The PRT controlled the boiler, which delivered hot water to all the
radiators, while the individual TRVs provide some room-by-room temperature control.
91
Figure 4-5: Test house schematic plans with heating systems and environmental
monitoring equipment as configured during heating trial 1, for heating trial 2 the
PTRVs with their central controller were swapped with TRVs in the opposite house
92
The PRT was set to switch the heating on for 10.5 hours per day on weekdays
(06:00- 09:00 and 15:00 – 22:30) and 17 hours per day during the weekends (06:00
– 23:00) (i.e. the ‘Heating on’ periods) and the boiler was switched off during the rest
of the day (i.e. the ‘Heating off’ periods)9. There was no set-back temperature during
the heating off period (Table 4-2). This is similar to the heating durations specified in
the UK standard calculation method (SAP) (BRE, 2014) but with each heating period
starting one hour earlier. This was because the poorly-insulated house needed
longer time to achieve suitable temperatures for the assumed periods of occupant
activity.
Suitable TRV positions were found for each radiator by trial and error in order to
achieve the comfort temperature specified by CIBSE Guide A (CIBSE, 2006a) for
winter comfort: i.e. 21oC in the living room and bathroom and 19oC in the bedrooms.
In the unoccupied spare room a setting that yielded approximately 12°C was used to
as this was assumed to be the lowest temperature that would avoid frost and
condensation (BRECSU, 2001) (Table 4-2). The TRV settings were determined for
the radiators in house 2 before starting the HT1 and were not changed when they
were transferred to the radiators in house 1 for the HT2.
For ZC, the whole system ‘heating on’ and ‘heating off’ periods were set by the PRT,
and were the same as for the CC. The difference between ZC and CC was that
programmable thermostatic radiator valves (PTRV) replaced the normal TRVs in 6 of
the rooms (Figure 4-5). Room temperature set-points were the same as for CC but
were set only for the ‘occupied’ hours (Table 4-2). However, the PTRVs’ central
controller adjusted the set-point temperature of the PTRVs 30 minutes before each
‘occupied’ period in order to allow the room to reach the set-point temperature
(Figure 4-6). The set-point temperatures were held whilst the room was ‘occupied’,
but allowed to fall to the set-back temperatures when the heating system was on but
the room scheduled to be unoccupied. The set-back temperature was 16°C in all
rooms except the unoccupied spare room for which 12°C was used (as for CC).
When the heating system was off according to the PRT there was no set-back
9
Throughout, ‘Heating on’ and ‘Heating off’ periods means the times given here.
weeks with ZC in House 1 and 4 weeks in House 2, and likewise for CC.
10
This is thus the average of 4
93
temperature. In other words, PTRVs could not cause the heating system to turn on
as it was controlled by the central thermostat.
Figure 4-6: A PTRV installed on a radiator (on the left) and the interface of the
central controller used to programme the PTRVs (on the right)
Compared to CC, in which all the rooms were heated to their set-point temperatures
for 10.5 hours during weekdays and 17 hours during weekends (i.e. ‘Heating on’
hours), ZC established shorter periods of time when each room was heated to its
set-point temperature (see Table 4-2).
94
Table 4-2: Weekday and weekend ‘occupied’ hours with the number of hours each
room was heated to the set-point or set-back temperatures and, for ZC, the PTRV
set-point and set-back temperatures, and for CC, the TRV position
ZC
Number
of hours
heated
to the
set-point
1
(WD
2
WE )
Number
of hours
heated
to the
set-back
1
(WD ,
2
WE )
CC
Number
of hours
TRV
heated
level
to the
3
(1-6)
set-point
1
(WD ,
2
WE )
Room
Weekday
‘occupied’
hours
Weekend
‘occupied’
hours
Setpoint
(ºC)
Setback
(ºC)
Living
Room
18:00-22:30
18:00-22:30
21
16
5, 5
5.5, 12
10.5, 17
4
Dining
Room
08:00-08:30
17:00-18:00
09:30-10:00
17:00-18:00
19
16
2.5, 2.5
8, 14.5
10.5, 17
3
Kitchen
07:30-08:00
16:00-17:00
09:00-09:30
16:00-17:00
-
-
-
-
-
-
Bedroom
1
19:00-22:30
22:30-08:00
08:30-09:00
16:00-17:00
19:00-22:30
22:30-09:30
10:00-10:30
16:00-17:00
19
16
8.5, 10
2, 7
10.5, 17
3
Bedroom
2
22:30-07:30
22:30-09:00
19
16
2, 3.5
8.5, 13.5
10.5, 17
4
Bathroom
07:30-08:00
08:30-09:00
19:00-20:00
09:00-09:30
10:00-10:30
19:00-20:00
21
16
3.5, 3.5
7, 13.5
10.5, 17
4
Unoccupied
Bedroom
-
-
12
-
10.5, 17
-
10.5, 17
1
1WD – weekdays
2 WE - weekends
3 The TRV settings provided the same set-point temperatures in each room as the set-points with ZC
4.4 Comparison of indoor air temperatures
The air temperature and radiator surface temperatures varied throughout a typical
weekday and weekend according to the heating strategy set on the PRT, but there
were distinct room-by-room temperature differences depending on whether CC or
ZC was used (e.g. Figure 4-7). In the morning, the radiators started to warm up when
the heating came on and with CC continued to emit heat until the set-point
temperature was reached. With ZC however, if the room was not scheduled to be
95
‘occupied’, the PTRV stopped the flow of water to the radiator when the set-back
temperature was reached (see Figure 4-7, dining room and living room, morning
heating period). If the room remained unoccupied, ZC only provided heat when the
air temperature fell below the set-back temperature whereas, with CC, heat was
provided to maintain the higher, set-point temperature (Figure 4-7, living room,
morning on period, bedroom 2 evening heating period). If a room with ZC became
occupied during a ‘heating on’ period, the PTRV would enable flow to the radiator to
bring the room temperature up to the set-point (Figure 4-7, dining room, living room
and bedroom evening heating on periods). It is the difference in the energy needed
to achieve the set-point temperature compared to the set-back temperature when the
heating is on but rooms are unoccupied, that leads to potentially lower heating
energy consumption. The lower the set-back temperature and the shorter the
occupied time relative to the heating on time, the more energy ZC might, in principle,
save. However, in ZC the heated rooms would lose more heat to neighbouring
spaces that are at a lower temperature compared to CC when the neighbouring
spaces are at a higher temperature.
The houses exhibited other characteristics common to UK centrally heated homes,
especially poorly insulated homes. For example, even though bedroom 2 was
‘occupied’ from 06:00 hrs to 07:30 hrs and the heating was on, the room failed to
reach the set-point temperature with either ZC or CC. In fact, the set-point wasn’t
reached even after 3 hours of heating using CC. Bedroom 2 has a particularly large
single-glazed bay window and therefore high rates of heat loss. In the middle of the
day, when the house was unheated, the temperatures in the north-facing rooms fell
to below the set-back temperature in the case of bedroom 1. In contrast, the solar
gain through the large, south-facing window of bedroom 2, and the similarly sized
window in the living room, caused the temperatures in the middle of the day to
exceed the heating set-point; especially in the house with CC (Figure 4-7). In the
evening heating period, with both CC and ZC, the living room, and to a lesser extent
the dining room temperatures exceeded the set-point during the occupied hours.
This was most likely due to the internal heat gains.
96
Figure 4-7: Air and radiator surface temperature variations in different rooms: heating
trial 1, 21st Feb 2014, ZC in House 1, CC in House 2.
97
Table 4-3 shows the average air temperature in each room for the 8 weeks trial
periods10. These are broken down into five different averaging periods: the whole of
each day; when the PRT had switched the ‘Heating on’; when the heating was on
and the space occupied, ‘Heating on and occupied’; when the heating was on but the
space was unoccupied, ‘Heating on and unoccupied’; and, finally, the average
during the ‘Heating off’ hours. The table also gives the floor area-weighted 11 average
temperature for the whole house during each of these five periods.
Table 4-3: Average indoor air temperatures in each room during five different periods,
and the spatially averaged whole house temperature
‘Heating on’
‘Heating on’
‘occupied’
ZC
CC
ZC
CC
ZC
CC
ZC
CC
ZC
CC
(ºC)
(ºC)
(ºC)
(ºC)
(ºC)
(ºC)
(ºC)
(ºC)
(ºC)
(ºC)
Living Room
19.2
20.0
20.3
21.5
22.3
22.5
18.7
20.5
18.0
18.4
Dining Room
18.2
18.7
19.0
19.5
20.4
20.1
18.8
19.4
17.4
17.7
Bedroom 1
18.0
18.3
18.9
19.2
18.9
19.2
18.7
19.4
17.1
17.3
Bedroom 2
17.2
18.2
17.6
19.1
16.3
18.1
17.9
19.3
16.5
17.1
Bathroom
16.5
17.7
17.3
18.9
19.7
19.1
17.2
18.9
15.5
16.4
14.8
15.3
14.9
15.5
-
-
14.9
15.5
14.6
15.0
19.1
19.5
20.3
20.8
-
-
20.3
20.8
17.8
18.1
Kitchen
19.6
20.0
20.7
21.2
23.0
23.6
20.4
20.8
18.4
18.6
Whole house2
18.1
18.7
18.9
19.7
19.7
20.1
18.6
19.6
17.1
17.5
Room
Unoccupied
room
Circulation
areas1
1
Average air temperature in hallways on the ground and first floors.
2
Floor area weighted average air temperature.
‘unoccupied’
‘Heating off’
Whole day
The averages are across four weeks with the control system in one house and four weeks in the
other house.
Considering the whole day, the average air temperature of all the rooms and the
whole house was lower with ZC than with CC. The temperatures were also lower
10
This is thus the average of 4 weeks with ZC in House 1 and 4 weeks in House 2, and likewise for CC.
Calculated as: (T1 ∗ A1 + T2 ∗ A2 + ⋯ + Tn ∗ An )/(A1 + A2 + ⋯ + An ) where: T1 to Tn are the average air
temperature of different rooms during each of the 5 periods and A1 to An are the floor area of those rooms
11
98
with ZC during periods when the heating system was on and when the heating
system was off. This was because ZC kept space temperatures low when rooms
were scheduled to be unoccupied, but provided similar air temperatures to CC (not
less than the set-point temperature) when the rooms were scheduled to be occupied.
During the ‘occupied’ hours when the heating was on, for both control strategies, the
average indoor air temperatures measured in the living room and dining room were
higher than their set-point temperature, which is thought to be due to the effect of
internal heat gains and closing the doors when the rooms were occupied.
The average air temperature in bedroom 2 was lower than its set-point temperature
during the ‘occupied’ hours especially in the house with ZC. This was because this
bedroom was ‘occupied’ mostly during the night when the occupants were assumed
to be sleeping and the heating was switched off (it is usual to sleep in an unheated
bedroom in the UK (Huebner et al. 2013). Therefore, the daily period when the
heating was on and the room was ‘occupied’ and thus heated was too short for the
room to achieve its set-point temperature (Table 4-2).
On a similar basis, the average air temperature in bedroom 1 during the occupied
hours was higher than bedroom 2 and close to the set-point temperature because it
was ‘occupied’ for longer each day, when the heating was on, for purposes other
than sleeping.
The average air temperatures during the sleeping periods are worth noting. In the
house with ZC they were 15.5ºC and 14.3ºC, in bedrooms 1 and 2, respectively,
which was lower than the averages of 16.2ºC and 14.6ºC found for CC. Bedroom air
temperatures in both homes are thus lower than the CIBSE recommendation for
bedrooms of 17ºC. However, Humphreys (1979) reports good sleep quality even for
bedroom temperatures as low as 12ºC while Collins (1986) and Hartley (2006)
indicate the world health organization’s bedroom temperature limit of 16ºC to reduce
the risk of decreasing resistance to respiratory infections which can occur at lower
temperatures (Peeters et al. 2009).
Bathroom average air temperatures were lower than the designed set-point
temperature with both ZC and CC during ‘occupied’ hours. This could be due to an
undersized radiator. Also, there were no internal heat gains as it was assumed that
99
in real houses any heat gain produced in this room would be quickly transferred to
outdoor via extract fans or window opening.
The mean temperatures in the unheated rooms (i.e. unoccupied room and kitchen)
were found to be lower for ZC during all the periods of the day. Again this was
assumed to be due to higher rates of heat loss and lower rates of heat gain to and
from the adjacent rooms in which were cooler in ZC compared to CC. The mean
temperature of the kitchen was much higher than all other rooms during the
‘occupied’ hours (23ºC and 23.6ºC for ZC and CC respectively). This was clearly due
to the considerable heat gains from cooking.
The daily average air temperatures in the circulation areas on the ground floor and
first floor were lower in the house with ZC compared to the house with CC. This
could again be explained by the lower temperatures in adjacent rooms acting as a
heat sink.
It is important to quote the energy savings of ZC when the same level of comfort as
CC is being provided. In this work, it is assumed that indoor air temperature alone is
a good proxy for thermal comfort. However, in this experimental work, it was not
possible and in fact intended to provide identical temperatures at the same time in
the two homes using the different control strategies. The consequence, as can be
seen from Table 4-3, is that the whole house average air temperature during
‘occupied’ hours was slightly lower with ZC (19.7°C), than it was with CC (20.1°C).
However, the main reason for the whole house average air temperature during the
“occupied” hours being slightly lower in ZC compared to CC was that ZC provided
lower air temperatures in bedroom 2 which was mainly occupied for the purpose of
sleeping as it was discussed earlier.
Considering the hours of ‘active occupancy’ (i.e. when the occupants are assumed to
be present and awake) for the entire 8 weeks of the trials the average air
temperatures of the whole house was 21.0ºC for ZC and 20.8ºC for CC. Therefore,
on average, for this experiment ZC provided a slightly higher air temperature
compared to CC during the time period of most interest (i.e. ‘active occupancy’).
Therefore, it was assumed that both control strategies provided the same level of
thermal comfort to the occupants.
100
4.5 Heating demand, boiler efficiencies and fuel use
During the heating trials the daily average outdoor air temperature ranged from a
minimum of 2.5ºC (Day 14) to a maximum of 13.1ºC (Day 48) with an average of
7.1ºC (Figure 4-8). As expected, whole house heating demand, as measured by the
boiler heat output, was greater on colder days than on warmer days. During the
weekends, the heat output was generally higher than for weekdays because the
heating was switched on for longer (Figure 4-8).
The daily heat output with ZC varied from 22.6 to 80.6 kWh/day with an average of
53.6 kWh/day, while with CC it varied from 25.0 to 90.8 kWh/day with an average of
62.4 kWh/day. On every day of the trials the daily boiler heat output in the house with
ZC was lower than the boiler output in the house with CC (Figure 4-8). Overall, daily
heat output with ZC was between 2.6% (Day 7) and 22.1% (Day 25) lower than with
CC, giving a daily average of 14.1% lower heat output.
Figure 4-8: Measured daily heat output from the boilers during the heating trials 1
and 2 and their error bars (based on heat meter’s manufacturer stated accuracy)
together with the average daily outdoor temperature
The efficiency of boilers when operating with ZC was lower than the efficiency of the
boilers when operating with CC (Figure 4-9). However, the difference was quite small,
101
being on average 1.5 percentage points (pp) less efficient during the first trial (HT1)
and 3.3pp less in the second trial (HT2). The larger difference during the HT2 is
perhaps due to the warmer weather which meant the boiler outputs were less and so
they were operating further away from the peak efficiencies for longer. At part load,
small differences in power output lead to larger differences in efficiency than at, or
near, peak load. There may also be some small differences between the boilers
installed in the two houses as they were less than seven years old.
Averaged over both trials, the efficiency of the boilers associated with ZC were 2.4pp
less efficient than the boilers controlled conventionally (CC) (Table 4-4). A standard
chi-square test was conducted to determine if the results were statistically significant.
This difference was found to be statistically significant (p<0.01) and is likely to be
because boilers operated under ZC, experiencing lower heating loads, and so
operate further away from the peak load capacity – at which they are most efficient.
Figure 4-9: Daily efficiency of the boilers with zonal control (ZC) and conventional
control (CC) in each heating trial with their error bars 12 together with the daily
average outdoor temperature
12
Uncertainty in daily boiler efficiencies are calculated as the quadratic sum of the uncertainties in calorific
value of gas, gas meter and heat meter (Table 5)
102
Table 4-4: Summary of daily average boiler efficiencies in each heating trial and
overall efficiency
1
Heating Trial 1,
Heating Trial 2,
Overall Average
Boiler Efficiency (%)
Boiler Efficiency (%)
Boiler Efficiency (%)
Daily Average
Daily Average
Daily Average
(minimum,
(minimum,
(minimum,
maximum)
maximum)
maximum)
Zonal Control
84.2%
82.8%
83.5%
(ZC)
(82.5%, 85.7%)
(80.4%, 85.6%)
(80.4%, 85.7%)
Conventional
85.7%
86.1%
85.9%
Control (CC)
(83.7%, 88.3%)
(82.9%, 89.3%)
(83.7%, 89.3%)
Difference
1.5pp1
3.3pp1
2.4pp1
Percentage points
The total gas consumption across both heating trials was 11.8% less with ZC than
with CC. This resulted from the combination of a reduced heat demand of 14.1% but
a reduction in boiler efficiency of 2.4pp. Average daily gas consumption was
significantly less (p<0.05) with ZC (64.2 kWh) rather than CC (72.8 kWh). During the
40 weekdays of monitoring, average daily gas consumption was significantly less
(p<0.01) in the house operating with ZC (61.8 kWh) rather than the house operating
with CC (71 kWh); a difference in gas consumption of 13%. During the 16 weekend
days the house with ZC used on average 70.3 kWh/day while the house operating
with CC used 77.3 kWh/day; a difference of 9.1% . However, this was not found to
be statistically significant; due to the relatively small number of weekend days (n=16)
for testing any statistical significance. Compared to weekdays, at the weekends
rooms are occupied for a greater proportion of the time that the heating is on
(Table 4-2) and the programmable thermostat (located in the hallway) tends to reach
the set-point more often with CC than with ZC and so the heating system is cycled
off for slightly longer with CC. These results suggest that houses that are more
intermittently occupied and which have rooms that are used infrequently might
benefit more from ZC than homes that are occupied extensively and for longer (see
chapter 8).
103
4.6 Summary
This chapter has described the space heating trials conducted in the LMP1930 test
houses during an 8-weeks winter test period and has presented the trials results.
The main findings from the space heating trials can be summarised as:
•
The average air temperature of all the rooms and the whole house was lower
with ZC than with CC considering the whole day, the period when the heating
system was on and the period when the heating was off.
•
In most rooms and in both houses, the average air temperature measured
during the occupied period when the heating was on was different from their
set-point temperatures.
•
The average air temperatures in bedrooms in both houses during the sleeping
period were below air temperatures recommended by CIBSE.
•
Whole house average air temperature during ‘occupied’ hours was slightly
lower with ZC (19.7°C), than it was with CC (20.1°C). However, these were
very close when excluding the air temperatures during the sleeping period.
•
Daily boiler heat output of the house with ZC was lower than that of the house
with CC on every single day. On average, daily heat output of the boiler in the
house with ZC was 14.1% lower than the boiler in the house with CC.
•
The average efficiency of the boilers associated with ZC were 2.4pp lower
than that of the boilers controlled conventionally (CC)
•
The total gas consumption across both heating trials was 11.8% less with ZC
than with CC.
•
The average gas savings of ZC were found to be higher during the
intermittently heated weekdays rather than weekends when the houses were
heated for longer periods.
104
5 Dynamic thermal modelling
5.1 Introduction
This chapter describes the use of of dynamic thermal models (DTMs) to simulate the
co-heating test (section 3.3.6) and space heating trials (chapter 4) conducted during
the experimental campaign of this study. The modelling approach adopted here was
according to recommendations by Lomas et al. (1997) in which the experimental
work was firstly simulated in a so called “blind phase” where the modeller is unaware
of the actual measured performance of the building. The empirical validation was
then conducted in an “open phase” (chapter 6) in which the measurements were
made available.
EnergyPlus version 8.1.0.009 which was released in October 2013 was used in this
research. EnergyPlus is a freely available dynamic thermal modelling tool which has
undergone a number of revisions and the current version 8.3 was released in March
2015. The input data for EnergyPlus simulations is contained in a text file called the
Input Data File (IDF). This enables the user to change sections of the input file and
control these changes using a text editor or a third party such as IDF editor.
DesignBuilder (2014) is a commercially available software package that offers
detailed dynamic thermal simulations, for which it uses the EnergyPlus simulation
engine and provides a user friendly graphical user interface. In this study,
DesignBuilder version 3.4.0.0.41 which was released in April 2014 was used to input
the building geometries, construction materials and input parameters for modelling
the air flow and heating systems. The model created in DesignBuilder were then
converted to the EnergyPlus IDF files, which were modified further using a text editor
and the EnergyPlus IDF Editor in order to construct the final EnergyPlus model and
run simulations.
The chapter starts with the description of modelling the building envelope of the test
houses (section 5.2). Then in section 5.3, it describes two different air flow modelling
methods which were used to model air flows in the houses. In section 5.4, modelling
of the heating systems which were used during the co-heating test and HT1 are
discussed. In section 5.5, the procedure for modelling the synthetic occupancy of the
105
houses is discussed. In section 5.6, the generation of the weather file for the periods
of co-heating test and the HT1 is described. Finally, section 5.7 presents a summary
of this chapter.
5.2 Modelling the building envelope of test houses
In this section, construction of a DTM of the building envelope of the LMP1930 test
houses is described including details of geometry, zoning, ground modelling and
construction materials.
5.2.1 Building geometry
Internal dimensions of the LMP1930 test houses were entered in to the
DesignBuilder software. The semi-detached test houses were modelled together
(Figure 5-1) as this allows influences of the adjacent house on thermal behaviour of
each house to be considered in the model. In addition, the neighbouring houses
were modelled as component blocks in order to take into account their potential
shading and reflection effects on LMP1930 houses (Figure 5-2).
Figure 5-1: Views of the LMP1930 test house model in DesignBuilder: front, southfacing (left) and back, north-facing (right)
106
Figure 5-2 View of the LMP1930 test house model with the effect of shading from the
neighbour blocks (15 March at 16:00)
A number of simplifications were made in the models. The party wall between the
two houses was modelled as a partition wall (see section 5.2.4 for details of
construction materials). The chimneys and sealed fire places were not considered in
the model.
The width and height of each window was entered separately, including the frame
according to the window corner definition in DesignBuilder (Figure 5-3). The window
frames and dividers were also entered separately for each window (Figure 5-3). The
window area which is provided to EnergyPlus IDF input file after conversion have
slightly smaller area compared to the area defined in DesignBuilder in order to take
into account the frames which is not considered in definition of window area in
EnergyPlus (Figure 5-3). Internal doors and external doors were also entered into
the model.
107
Figure 5-3: Window geometry definition in DesignBuilder and EnergyPlus
(DesignBuilder, 2014)
5.2.2 Zoning
In each house, the lower storey was divided into 4 zones including the living room,
dining room, kitchen and hallway while the upper storey was divided into 6 zones of
hallway, bathroom, bedroom 1, bedroom 2, bedroom 3 (unoccupied room) and a WC
(Figure 5-4). The subfloor and the loft (attic) space of each house was considered as
additional unheated zones.
Each EnergyPlus zone is defined as a common air mass at a specific temperature
(i.e. the air is fully mixed). In space heating with ZC, each room with a radiator and
PTRV is controlled to a temperature which is often different from the temperatures at
which other rooms are being controlled. Although few zones such as bedroom 1 and
2 which have the same set-point temperature could have been merged into one zone
for the house with CC, keeping the same zoning configuration for the houses with
CC and ZC would enable room by room comparison of the two control strategies. In
addition, having separate zones enabled the internal heat gains of each zone to be
modelled more accurately.
108
Figure 5-4: LMP1930 test house model zoning strategy for ground floor and first floor
109
5.2.3 Ground modelling
The suspended timber ground floor of each house was modelled explicitly as a
separate zone (subfloor zone) which was added below the ground floor. The subfloor
zone had six air bricks each having an open area of 0.01 𝑚𝑚2 as measured at the test
houses. The height of the subfloor was 0.2 m according to the measured depth of
the underfloor void existed (from the bottom of the floor boarding to the ground). It
was assumed that the ground under the suspended timber floor is just bare earth
(see section 3.3.3).
The solid floors of the kitchens were represented by a 100 mm concrete slab. The
thickness of concrete slab could not be directly measured and was assumed to be
100 mm, according to a document by University of the West of England (2009).
Average monthly ground surface temperatures under the building are used by
EnergyPlus as the outside surface temperature for all surfaces adjacent to the
ground to calculate the heat transfer between the ground and any adjacent zone.
Average monthly ground surface temperatures could be calculated using 3D ground
heat transfer program of EnergyPlus for slabs (US Department of Energy, 2013b).
The 3D slab program included in EnergyPlus produces outside surface temperature
of the core and perimeter of a slab in contact with the ground. The programme uses
twelve separate average monthly indoor temperatures as inputs for the calculation of
the ground temperature. However, this programme could not be used to calculate
the ground temperature under a ventilated suspended timber floor, and this was not
measured during the experimental work.
According to EnergyPlus documentation, the undisturbed ground temperatures
calculated by EnergyPlus’s weather converter program are often not appropriate for
building loss calculations as these values are too extreme for the soil under a
conditioned building (US Department of Energy, 2013b). EnergyPlus documentation
(US Department of Energy, 2013b) suggests using ground temperatures of 2°C
below mean internal temperatures for large commercial buildings in the US. However,
it does not suggest any method for calculating or estimating ground surface
temperature under a ventilated suspended timber floor or for small residential
buildings such as this case. An article by (Lstiburek, 2008) published in ASHRAE
110
Journal of Building Sciences suggests that a reasonable rule of thumb to estimate
the ground surface temperature of ventilated crawlspaces is to use the average
annual ambient air temperature of that location. In absence of any other reference,
the average annual ambient air temperature measured at Sutton Bonnington
weather station for the year 2014 which was 11.8ºC (UK Meteorological Office, 2012)
was assumed as the monthly ground surface temperatures.
5.2.4 Construction materials and properties
Construction materials properties were selected from DesignBuilder’s library as
shown in Table 5-1.
Table 5-1: Construction materials properties used in LMP1930 model
Material
Conductivity
Density
Specific heat
(W/m. K)
(kg/𝒎𝒎𝟑𝟑 )
capacity (J/kg. K)
800
1700
Brick (outer leaf)
0.84
Brick (inner leaf)
0.62
800
1700
Plaster (dense)
0.50
1000
1300
Clay tile
1.00
800
2000
Roofing felt
0.19
837
960
Glazing
0.9
-
-
Polyisocyanurate
0.022
1470
45
Timber flooring
0.14
1200
650
Cast concrete
1.13
1000
2000
Carpet
0.06
1300
200
Plasterboard
0.25
896
2800
0.19
2390
700
Painted oak (doors
and windows)
DesignBuilder models each building element as one or more layers of construction
materials with a specific thickness. The U-value of each element was automatically
calculated by DesignBuilder (Table 5-2).
111
Table 5-2: Construction elements of the test houses and their U-values and
thicknesses of the materials used in each construction element
Materials
Building element
(outermost to
innermost layer)
External cavity
Brick, air gap,
walls
brick, dense plaster
Internal partition
Plaster, brick,
walls
dense plaster
Plaster, brick, air
Party wall
gap, brick, dense
plaster
Ground floor (semi
Timber flooring,
exposed)
carpet
Kitchen’s solid floor
Cast concrete
Internal floor
Plasterboard, air
(between ground
gap, timber
floor and first floor)
flooring, carpet
First floor ceiling
(semi exposed)
Pitched roof
Clay tile, air gap,
roofing felt
U-value1
innermost layer) (m)
(W/𝒎𝒎𝟐𝟐 𝑲𝑲)
0.105, 0.07, 0.105, 0.013
1.666
0.013, 0.105, 0.013
2.077
0.013, 0.105, 0.07, 0.105,
0.013
1.281
0.02, 0.005
2.015
0.1
3.35
0.013, 0.1, 0.02, 0.005
1.373
0.013
3.1
0.025, 0.02, 0.005
2.93
Glazing
Single glazing
0.003
5.894
Window Frame
Wooden (oak)
0.02
3.633
0.003, 0.01, 0.05
0.377
0.044
2.034
Window covered
with insulation
board
Doors (internal &
external)
1
Plaster board
Thicknesses (outermost to
Glass, air gap,
Polyisocyanurate
Wooden (oak)
U-values were calculated by DesignBuilder for simple calculation methods such as SBEM
112
For the windows, 3 mm single layer clear glass was selected from DesignBuilder
glazing type templates for the whole house model. Characteristics of the glazing
material selected were presented in Table 5-3.
Table 5-3: Characteristics of the glazing in LMP1930 model
Type
Conductivity
(W/m K)
3mm
0.9
clear
Solar
transmittance
(SHGC)
0.837
Outside/
Outside/
inside
Visible
inside
solar
transmittance
visible
reflectance
0.075
reflectance
0.898
Outside/
inside
emissivity
0.081
0.84
Sub-surfaces in DesignBuilder define areas which have a different construction to
that of the main. The windows on the East and West facades, which were covered
from inside with insulation boards during the experiments, were modelled using a
sub-surface with 3 layers: 3 mm glass, 10 mm air gap and 50 mm
Polyisocyanurate insulation boards (thermal conductivity of 0.022 W/m𝐾𝐾 (Celotex,
2015)). The total U-value of the sub-surface was calculated as 0.377 W/𝑚𝑚2 𝐾𝐾
(Table 5-2).
The blinds in the houses were modelled as a closed weave, medium coloured shade
from the DesignBuilder database. The transmittance and reflectance characteristics
matched those in the ASHRAE handbook of fundamentals (ASHRAE, 2009)
(Table 5-4).
Table 5-4: Characteristics of the blinds material chosen for the model
Characteristics
Values
Thickness (m)
0.001
Conductivity (W/m-K)
0.1
Solar / visible transmittance
0.05
Solar / visible reflectance
0.3
Long wave emissivity
0.9
Long wave transmittance
0
113
5.3 Modelling the air flow
Air flows in buildings happen when there is a pressure difference between two points
and a continuous flow path or opening which connects the points (Straube, 2008). In
a naturally ventilated building, the pressure difference can be caused by wind and air
density differences between the points due to their temperature difference (buoyancy
or stack effect) (Straube, 2008).
EnergyPlus has three approaches to modelling the air flow in buildings: scheduled
natural ventilation (SNV), Air Flow Networks (AFN) and Computational Fluid
Dynamics (CFD). However, only two of them (i.e. SNV and AFN) could be used
when the model is used for the purpose of predicting the energy consumption of the
building. Each approach has its own advantages and disadvantages and one
important decision was to select the most appropriate method of modelling air flows
for this research. In order to test the suitability of the two air flow modelling
approaches, both approaches were used to simulate the co-heating test and space
heating trials and the results were compared with each other and the measured data.
5.3.1 Scheduled Natural Ventilation (SNV)
Scheduled natural ventilation is the simplest approach for modelling air flows. A
design air infiltration rate for each zone is input directly in units such as flow per zone
(𝑚𝑚3 /𝑠𝑠) or flow per zone floor area (𝑚𝑚3 /𝑠𝑠 𝑚𝑚2 ) or flow per exterior surface area
(𝑚𝑚3 /𝑠𝑠 𝑚𝑚2 ) or air change rates per hour. EnergyPlus then modifies these design flow
rates using equation (5-1).
Where:
𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 = 𝐼𝐼𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 ∗ 𝐹𝐹𝑠𝑠𝑠𝑠ℎ𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 ∗ (𝐴𝐴 + 𝐵𝐵|𝑇𝑇𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧 − 𝑇𝑇𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑟𝑟 |
+ 𝐶𝐶 𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 + 𝐷𝐷 𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 2 )
(5-1)
𝐼𝐼𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = specified infiltration of the zone as a design level
𝐹𝐹𝑠𝑠𝑠𝑠ℎ𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 = schedule fraction which can modify the infiltration volume flow rate for
each time step according to a defined schedule for each zone.
114
A= constant term coefficient with a default value of 1
B= temperature term coefficient with a default value of 0
𝑇𝑇𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧 − 𝑇𝑇𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 = temperature difference between the zone and outdoor
C= velocity term coefficient with a default value of 0
D= velocity squared term coefficient with a default value of 0
As default EnergyPlus assumes the values of 1 for coefficient A and 0 for coefficients
B, C and D which gives a constant volume of infiltration air flow under all conditions.
According to EnergyPlus input output reference (US Department of Energy, 2013c) a
detailed analysis is needed to determine a custom set of coefficients. Therefore, the
default coefficients were not changed for LMP1930 model.
Measuring infiltration rates of the individual zones of the LMP1930 was not possible.
Instead, the whole house infiltration rate, as measured during the airtightness test,
was used in the model. DesignBuilder uses equation (5-2), sourced from BS EN
12831 (British Standards, 2013), to convert the whole house infiltration rate
measured at 50 Pa to infiltration rate at normal operating conditions for each zone.
Equation (5-2) uses a shielding coefficient (𝑒𝑒) which takes into account the number
of exposed openings in each zone and wind exposure and a height correction factor
(ε).
𝑉𝑉̇𝑖𝑖𝑖𝑖𝑖𝑖,𝑖𝑖 = 2. 𝑉𝑉𝑖𝑖 . 𝑛𝑛50 . 𝑒𝑒𝑖𝑖 . 𝜀𝜀𝑖𝑖
[𝑚𝑚3 ⁄ℎ]
(5-2)
Where:
𝑉𝑉̇𝑖𝑖𝑖𝑖𝑖𝑖,𝑖𝑖 = infiltration air flow rate of heated space (i) induced by wind and stack effect
on the building envelope
𝑉𝑉𝑖𝑖 = volume of heated space (i) in 𝑚𝑚3 calculated on the basis of internal dimensions
𝑛𝑛50 = air exchange rate per hour (ℎ−1 ), resulting from a pressure difference of 50 Pa
between the inside and outside of the building
115
𝑒𝑒𝑖𝑖 = shielding coefficient obtained from Table 5-5. For the case of LMP1930 model,
moderate shielding was set in the model.
𝜀𝜀𝑖𝑖 = height correction factor which takes into account the increase in wind speed with
the height of the space from ground level. 𝜀𝜀𝑖𝑖 =1 when the centre of zone height to
ground level is below 10 m which was the case for all the zones in LMP1930 model.
Table 5-5: Shielding coefficient (e) reproduced from Table D.8 BS EN 12831 (British
Standards, 2013)
Shielding class
No shielding (buildings in windy areas, high
rise buildings in city centres)
𝑒𝑒
Heated
Heated
Heated space
space
space with
with more
without
one
than one
exposed
exposed
exposed
openings
opening
opening
0
0.03
0.05
0
0.02
0.03
0
0.01
0.02
Moderate shielding (buildings in the country
with trees or other buildings around them,
suburbs)
Heavy shielding (average height buildings in
city centres, buildings in forests)
As there was no significant difference between the infiltration rates measured in the
two test houses (see section 3.3.6), the mean result (i.e. 21.75 ACH at 50 pa) was
used in the model for the both houses.
The ventilation rates of the subfloor and the loft (attic) space could not be estimated
by this method as they were not measured in the airtightness test. Measurements of
the ventilation rates of suspended floors (either concrete or timber) are very limited
(Hartless, 2004 & Edwards et al, 1990). In a study by Edwards et al. (1990), subfloor
ventilation rates of a 45 𝑚𝑚2 low energy UK house was measured between about 0.1
to near 2 ACH for different wind speeds and wind directions. However, the total
effective area of the air bricks was only 0.018 𝑚𝑚2 compared to 0.06 𝑚𝑚2 in the
LMP1930 houses with the same floor area. Also the void depth was 1.0 m compared
to 0.2 m for the LMP1930 houses. The smaller total effective area of the air bricks
116
(about 1/3 of the total effective area of the LMP1930) and considerably larger volume
of the void (about 5 times larger) in the house examined by Edwards et al (1990)
suggests that the subfloor ventilation rates of the LMP1930 houses could be
considerably higher in air changes per hour.
The only study which was found to report the measured ventilation rates beneath a
suspended timber floor of a UK house with a similar void depth (i.e. 0.022 m
compared to 0.02 m in LMP1930) and total effective area of air bricks (0.07 vs 0.06
𝑚𝑚2 in LMP 1930) reported that the subfloor ventilation rate was widely fluctuating;
ranging from about 3 air changes per hour (ach) to over 13 ACH (Hartless & White,
1994). Hartless & White (1994) argues that the subfloor ventilation rate of the house
examined was heavily influenced by the subfloor/external temperature difference
rather than the wind speed. Infra-red thermography showed that the air was moving
from the subfloor void to the gap behind the plasterboard in the walls due to a
leakage path at the wall/floor junction. Hartless & White (1994) discussed that this
problem has been also observed in other UK homes and could explain the high
subfloor ventilation rates found in their study. However, there was no plasterboard
used in the walls of LMP1930 test houses.
Considering the lack of comprehensive data regarding the subfloor ventilation rates,
ventilation rate of 8 ACH which was the mean value of 3 and 13 ACH found as lower
and upper limits of subfloor ventilation rate in Hartless’s (1994) study was chosen in
this study as the constant subfloor ventilation rate of both LMP1930 test houses.
Ventilation rates of loft (attic) spaces were measured in a number of studies; mainly
in the US. Dietz et al. (1986) conducted detail multi-zone PFT gas measurements in
a number of homes in the US and reported 3 ACH as “typical” for ventilation rate of
loft spaces. I’anson et al. (1982) measured loft space ventilation rate of 4.3 ACH in a
middle terraced three bedroom house using three tracer gases. The loft space of this
house was ventilated by a continuous gap with 10 mm width behind the fascia board.
Allinson (2007) modelled ventilated pitched roofs during low wind speed conditions in
the UK and chose a ventilation rate of 2 ACH according to assumptions by Burch
(1980). Sanders et al. (2006) developed a number of broad rules for estimating the
loft ventilation based on a series of measurements of the ventilation rates of the
houses (including loft) using tracer gas techniques which were conducted in about
117
eighty properties in England and Scotland during 1970s and 1980s. According to this
document, where the loft is not sealed, but with no eaves or ridge ventilators, the loft
ventilation rate in air changes per hour (ACH) is approximately equivalent to the wind
speed in m/s. This is similar to the case of the LMP1930 houses where there were
no eaves or ridge ventilators. Therefore, the average wind speed during each test
was used as the constant ventilation rate of the loft space of the LMP1930 test
houses. The average wind speed measured during the co-heating test and heating
trial 1 were 2.7 m/s and 4.0 m/s which suggested 2.7 and 4.0 ACH for the loft space
ventilation rate of the LMP1930 houses during the co-heating test and heating trial 1
respectively. These were close to the suggested typical loft space ventilation rates
measured or assumed in the other studies.
In SNV, the air exchange between zones through openings such as internal doors,
windows or holes (i.e. stairs) is modelled using the concept of mixing where equal
amounts of air are transferred from one zone to another and vice versa. It is not
possible to model unidirectional air flow from one zone to another using this method.
The design flow rate is the maximum air exchange between the two zones and is
explicitly defined for each opening as flow rate per zone (𝑚𝑚3 /𝑠𝑠), flow rate per zone
floor area (𝑚𝑚3 /𝑠𝑠 𝑚𝑚2 ), flow rate per person or air changes per hour (ach). This
maximum value is then modified by a schedule fraction which defines the operating
schedule of the opening.
DesignBuilder’s default value of 0.1 𝑚𝑚3 ⁄𝑠𝑠. 𝑚𝑚2 was selected as the air flow rate per
opening area which exchanges between each two adjacent zone through openings.
The same value of 0.1 𝑚𝑚3 ⁄𝑠𝑠. 𝑚𝑚2 was also considered for the air flow rate per square
meter of the opening which connected the lower storey to the upper storey. The
opening has a measured area of 2.25 𝑚𝑚2 . This value of 0.1 𝑚𝑚3 ⁄𝑠𝑠. 𝑚𝑚2 was
automatically multiplied by the area of each opening by DesignBuilder to provide the
air flow rate of each zone in 𝑚𝑚3 ⁄𝑠𝑠 which is used in IDF file.
5.3.2 Air Flow Network (AFN)
A second, more detailed approach to modelling the air flows through a building is to
establish an Air Flow Network (AFN). The AFN consists of a number of nodes
connected by air flow components through surface linkages (Gu, 2007). Each heat
118
transfer surface in a building, with both faces exposed to air, works as a surface
linkage through which air flows (Gu, 2007).The associated air flow component for
each surface can be one crack (or surface effective leakage area) at the average
height of the surface, one opening in an exterior or interior window or door, or a
horizontal opening. In EnergyPlus, each linkage surface specifies two connected
nodes: two zone nodes based on inside and outside face environment for an interior
surface, or a zone node based on inside face environment and an external node (US
Department of Energy, 2013c). Since AFN assumes that air flows from one node to
another, it simplifies airflows through its pathways and cannot predict internal air
circulation within a thermal zone (Gu, 2007).
DesignBuilder was employed in this study to facilitate the process of defining the
nodes and linkage surfaces via its “calculated natural ventilation” simulation option.
The air flow through cracks in the walls, floors and the roof is calculated by AFN
model as a function of the pressure difference across the crack according to power
law in form of equation (5-3) (US Department of Energy, 2013c).
𝑄𝑄 = (𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑘𝑘 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹) ∗ 𝐶𝐶𝑇𝑇 ∗ 𝐶𝐶𝑄𝑄 (∆𝑃𝑃)𝑛𝑛
(5-3)
Where:
𝑄𝑄 = air mass flow rate (kg/s)
Crack factor = multiplier for a crack
𝐶𝐶𝑇𝑇 = reference condition temperature correction factor (dimensionless)
𝐶𝐶𝑄𝑄 = air mass flow coefficient (kg/sat1 Pa)
∆𝑃𝑃 = pressure difference across crack (Pa)
n = Air flow exponent (dimensionless): The valid range is 0.5 for fully turbulent flow to
1.0, for fully laminar flow (US Department of Energy, 2013c).
Air flows through doors, windows and vents when they are open or closed are
calculated by a similar method. When these openings are closed, AFN model
automatically generates a crack around the perimeter of each opening. The air mass
119
flow coefficient (𝐶𝐶𝑄𝑄 ) (kg/s at1 Pa) is calculated by multiplying the air mass flow
coefficient (kg/s. crack length at1 Pa) by the length of the crack (i.e. the perimeter of
the opening).
When these openings are open another form of the power law equation in form of
equation (5-4) is used:
2∆𝑃𝑃
𝑄𝑄 = 𝐶𝐶𝑑𝑑 𝐴𝐴�
(5-4)
𝜌𝜌
Where:
𝑄𝑄 = volume flow rate across the opening (m3 /s)
𝐶𝐶𝑑𝑑 = discharge coefficient (dimensionless); depends on the geometry of the opening
and the Reynolds number of the flow
A = surface area of the opening (m2 ); defined using an opening factor which defines
the fraction of total surface area of an opening which is opened
ΔP = pressure difference across the opening (Pa)
ρ = air density (kg/m3 )
The air mass flow rate (kg/s) is then calculated by multiplying the volume flow rate by
the air density. Bi-directional flows can be modelled for vertical openings when air is
simultaneously moving in two directions depending on stack effects and wind
conditions (US Department of Energy, 2013c).
EnergyPlus can also use AFN to model air flows through horizontal openings such
as staircase. Horizontal openings can produce two-way flow when forced and
buoyancy flows co- exists, however, AFN cannot model bi-directional flows at a
given time step (US Department of Energy, 2013c)
The input variables required for establishing the AFN were: wind pressure
coefficients (𝐶𝐶𝑝𝑝 ), air mass flow coefficient (𝐶𝐶𝑄𝑄 ) (kg/s at1 Pa) and flow exponent (n) for
120
each crack, air mass flow coefficient (𝐶𝐶𝑄𝑄 ) (kg/s. m crack length) and flow exponent (n)
for the doors and windows when they are closed and discharge coefficient (𝐶𝐶𝑑𝑑 ) for
each opening at each opening factor. These are discussed in more detail below.
•
Wind pressure coefficients (𝑪𝑪𝒑𝒑 )
AFN uses wind pressure coefficients (𝐶𝐶𝑝𝑝 ) to calculate the wind driven pressure on
the external surfaces of a building. Wind pressure coefficient values are required for
each wind direction at an interval (for example: every 45 degrees) on each external
surface. Sensitivity analysis by Cóstola et al. (2010) has shown 𝐶𝐶𝑝𝑝 as one of the
most influential input parameters on air change rate and thus several building
performance indicators such as energy consumption and thermal comfort (Cóstola,
Blocken & Hensen, 2009). The wind pressure coefficient is dependent on a number
of factors including building geometry, facade detailing, position on the facade, the
degree of exposure, wind speed and wind direction (Cóstola, Blocken & Hensen,
2009). Therefore, wind pressure coefficients are generally unknown, except in the
case of very simple structures or extremely well studied buildings, and must be
assumed which could significantly influence the accuracy of the air change rate
calculations (ASHRAE, 2009).
Wind pressure coefficients could be obtained from full scale measurements or wind
tunnel model tests of the specific site and building or via CFD (ASHRAE, 2009).
However, full scale experiments are very complex and expensive. Alternatively, there
are databases of 𝐶𝐶𝑝𝑝 values which could be used as secondary sources of data.
DesignBuilder is supplied with a database of wind pressure coefficients based on
data from Liddament (1986) which is also reported in CIBSE guide A (CIBSE, 2006a)
and is often used as a “good first level of approximation for basic design purposes”
(DesignBuilder, 2014). The 𝐶𝐶𝑝𝑝 data is for low rise buildings (i.e. buildings of 3 storeys
or less) with square surfaces (aspect ratio 1:1) and for 3 levels of site exposure to
wind: sheltered, normal and exposed. The data is given in 45° increments. In this
study, 𝐶𝐶𝑝𝑝 data was chosen from DesignBuilder’s database considering normal site
exposure. Figure 5-5 was adopted from CIBSE guide A (CIBSE, 2006) and shows
the definition of surfaces in determining wind pressure coefficients.
121
Figure 5-5: definition of surfaces in determining wind pressure coefficients (CIBSE,
2006)
Example of wind pressure coefficients over façade 1 and roof (front) for wind angels
in 45º increments were presented in Table 5-6 (DesignBuilder, 2014). They were
based on the slope of surfaces considering normal exposure of the site to wind and
aspect ratio 1:1.
Table 5-6: Wind pressure coefficients over façade 1 and roof (front) for wind angles
in 45º increments based on the slope of surfaces considering normal exposure of the
site to wind and aspect ratio 1:1 (DesignBuilder, 2014)
Wind angel to surface Vertical Slope<=10º Slope 11-30º Slope 31-89º
0º
0.4
-0.6
-0.35
0.3
45º
0.1
-0.5
-0.45
-0.5
90º
-0.3
-0.4
-0.55
-0.6
135º
-0.35
-0.5
-0.45
-0.5
180º
-0.2
-0.6
-0.35
-0.5
225º
-0.35
-0.5
-0.45
-0.5
270º
-0.3
-0.4
-0.55
-0.6
315º
-0.1
-0.5
-0.45
-0.5
122
•
Air mass flow coefficient (𝑪𝑪𝑸𝑸 ) (kg/s at 1 Pa) and flow exponent (n) for
each crack
AFN requires air mass flow coefficient (𝐶𝐶𝑄𝑄 ) (kg/s) at a reference condition
(temperature, pressure and humidity) for each crack in internal and external walls,
floor/ceiling and roof defined at 1 pa pressure difference across the crack. Gaps and
cracks in the building fabric cannot be accurately characterized by visual inspection
as the leakage paths are often obscured by internal finishes or external cladding and
are hard to follow (ATTMA, 2010). Although the air tightness of the test houses were
measured at 50 Pa, it was not possible to use these values directly in the model
when using AFN.
DesignBuilder uses a simplified approach which defines one crack for each surface
of the building. The characteristics of these cracks are defined in DesignBuilder
crack templates. There are five crack templates in DesignBuilder: Very poor, poor,
medium, good and excellent which can be selected according to the leakiness level
of the building under study. Since the air permeability test proven an indication of
poor air tightness of the test houses (see section 3.3.6), data corresponding to “poor”
crack template was chosen for the model. The crack templates has air mass flow
coefficient per square meter of each surface (kg/s.𝑚𝑚2 ) at1 Pa (Table 5-7) which
provides the air mass flow coefficient (𝐶𝐶𝑄𝑄 ) (kg/s) required in EnergyPlus by
multiplying the flow coefficient per square meter of the surface by the surface area
(Table 5-7). In addition, DesignBuilder’s crack templates have flow exponents (n)
(equation (5-3)) for internal and external walls, floor/ceiling and roof (Table 5-7).
123
Table 5-7: Crack characteristics according to DesignBuilder’s “poor” crack template
used in the model for walls, floors and the roof
Air mass flow
coefficient (𝑪𝑪𝑸𝑸 )
Flow exponent (n)
0.0002
0.7
Internal walls
0.005
0.75
Internal floors
0.002
0.7
External floors
0.001
1.0
Roof
0.00015
0.7
Building element
(Kg/s.𝒎𝒎𝟐𝟐 ) at 1Pa
External walls
•
Air mass flow coefficient (𝑪𝑪𝑸𝑸 ) (kg/s. m crack length) and flow exponent
(n) for the doors and windows when they are closed
DesignBuilder also provides the air mass flow coefficient (𝐶𝐶𝑄𝑄 ) (kg/s. m crack) at1 Pa
and flow exponent (n) for the cracks around the perimeter of these openings on the
same five point scale (Table 5-8).
Table 5-8: Crack characteristics according to DesignBuilder’s “poor” crack template
used in the model for the doors, windows and vents
Air mass flow
coefficient (𝑪𝑪𝑸𝑸 ) (Kg/s.
Flow exponent (n)
External windows
0.001
0.6
External doors
0.0018
0.66
Internal doors
0.02
0.6
External vents
0.01
0.66
Building element
m) at 1Pa
•
Discharge coefficients (𝑪𝑪𝒅𝒅 )
Discharge coefficient is difficult to determine and experimental values which has
found for discharge coefficient varies from 0.3 to 0.8 and without a clear
understanding of what causes these differences (International Energy Agency, 1992).
124
CONTAMW which is a multi-zone air flow and contaminant transport analysis
software developed by US department of commerce (Dols & Walton, 2002) suggests
a discharge coefficient of 0.6 for orifices and slightly higher for large openings in
buildings. ASHRAE (ASHRAE, 2009) propose the correlation based on inter zone
temperature differences as in equation (5-5) for the range of ΔTs from 0.5 to 40ºC:
Cd = 0.4 + 0.0045 ΔT
(5-5)
DesignBuilder’s help documentation notes that “given other uncertainties in natural
ventilation calculations (wind pressure coefficients, effective areas of real-world
openings and crack flows etc.), using a discharge coefficient between 0.60 and 0.65
should provide sufficient accuracy” (DesignBuilder, 2014). Discharge coefficient of
0.65 was selected for all the openings including the horizontal openings and both
opening factors.
5.4 Modelling the heating systems
This section describes the methods for modelling the heating systems for the coheating test (section 5.4.1) and the HT1 (section 5.4.2).
5.4.1 Modelling the heating system for the co-heating test
The co-heating test (see section 3.3.6) was modelled using electric convectors with
100% efficiency in every zone of the LMP1930 building envelope model (except the
unheated loft (attic) and subfloor zones). The average air temperature measured
during the co-heating test in each zone was used as the set-point temperature of that
zone in the model (Table 5-9).
125
Table 5-9: Measured average air temperature in different zones during the coheating test; theses temperatures were used as the set-point temperature of each
zone in the DTM when modelling the co-heating test
House 1
House 2
Set-point temperature
Set-point temperature
(°C)
(°C)
Living room
24.32
24.40
Dining room
24.64
24.10
Kitchen
25.15
24.29
Hallway ground floor
24.00
24.16
Hallway first floor
24.67
24.48
Bedroom 1
24.62
24.82
Bedroom 2
24.67
23.35
Unoccupied bedroom
24.83
24.43
Bathroom
23.53
24.20
24.50
24.82
Zone
Volumetric weighted
Average for the whole
house
Electricity used by circulation fans during the co-heating test was considered to end
up as heat in the zone, thus there was no need to model these separately.
5.4.2 Modelling the heating systems for the space heating trials
The gas powered central heating systems were modelled to simulate the HT1: one
with CC and the other one with ZC. Each heating system consisted of a gas fired
condensing combination boiler and 7 radiators as described in section 3.3.4
(Table 3-3). They were modelled for each house using DesignBuilder’s detailed
HVAC option.
The condensing combination boilers were modelled with nominal heat output of 30
kW and mean efficiency of 84.2% and 85.7% as measured during the HT1 (see
section 4.5). The normalized boiler efficiency curve of condensing combination
boilers was selected from DesignBuilder’s template library. The circulating hot water
flow temperature was set to maximum during the HT1 which is 88°C according to the
126
manufacturer’s data (Worcester Bosch Group, 2009). In DesignBuilder, the hot water
flow temperature in a wet heating system is controlled via a set-point manager which
controls the hot water flow temperature according to a schedule. This was set to be
always 88ºC.
Radiators were modelled using the water baseboard heater model of EnergyPlus
enabling both convection and radiation heat transfer. The water mass flow rate of
each radiator supplied from the primary system is calculated at each time step by
determining the impact of radiator on surrounding air via convection and to the
surfaces by radiation (US Department of Energy, 2012).
There will be water flow rate and therefore heat transfer from the radiator when all of
the three following criteria are met: firstly, the radiator unit is “on” at that time step;
secondly, there is any heat requirement remaining in the zone to be met according to
the zone’s set-point temperature and finally the boiler is “on” according to its
schedule.
The water baseboard heater model requires a number of inputs: rated average water
temperature (°C), rated water mass flow rate (kg/s) and rated capacity (W).
According to the radiators’ manufacturer data: rated average water temperature was
70°C and the rated water mass flow rate (kg/s) of each radiator was calculated using
equation (5-6):
𝑚𝑚 = 𝐻𝐻/(𝐶𝐶𝑝𝑝 ∗ (𝑡𝑡𝑓𝑓 − 𝑡𝑡𝑟𝑟 ))
(5-6)
Where:
𝑚𝑚= rated water mass flow rate (kg/s)
𝐻𝐻= rated capacity of radiator (W) selected from Table 3-3 according to the
manufacturer’s data
𝐶𝐶𝑝𝑝 = specific heat capacity of water and was approximated as 4187 J/kg.°C for the
purpose of calculating water flow in radiators
127
𝑡𝑡𝑓𝑓 = standard water flow temperature (°C) = 75°C
𝑡𝑡𝑟𝑟 = standard water return temperature (°C) = 65°C
The radiant fraction of the radiators is the portion of the power input transferred to
the occupants and surfaces as radiant heat and was considered to be 0.3 for all the
radiators according to Oughton & Hodkinson (2008).
A constant speed pump was modelled for the circulating hot water supply loop of
each house with a maximum loop flow rate of 0.00034 𝑚𝑚3 /𝑠𝑠 and minimum loop flow
rate of zero and a rated pump head of 6000 pa according to the specifications of the
central heating pumps in the houses. The control type of the pump was selected as
intermittent control. This enabled the modelled pump to shut down when no heating
was required. When there was heat demand, the pump selected a flow rate
somewhere between the maximum and minimum user defined flow rates in order to
meet the heating requirements. Rated energy consumption of the pumps was left as
“autosize” and default value of 0.9 was selected for the motor efficiency of the pumps
as the electricity consumption of the houses was not studied in this research.
All the pipes in the system were assumed to be adiabatic. There was no information
available regarding the pipe run in the houses and obtaining more information
required removing a large amount of the floor boards on the ground and first floors
which was not possible to do in this work.
The Programmable Room Thermostat (PRT) (see section 4.3) was modelled using
the boiler operation availability schedule of DesignBuilder’s circulating hot water loop
data. The radiators availability schedules were set to be always “on”.
The default control strategy of a wet heating system in a multi zone building model in
EnergyPlus and DesignBuilder is that each zone has its own room thermostat which
could be scheduled to assign set-point and set-back temperatures throughout a day.
However, this control strategy of the heating system is inherently different from the
control strategy in houses with either CC or ZC where boiler operation was controlled
by a PRT located in the hallway and set-point and set-back temperatures (only in ZC)
for each room are applied by TRVs (in CC) or PTRVs (in ZC). Currently, there is no
solution in DesignBuilder in order to better represent the control strategy in multi
128
zone houses with a PRT control over the boiler and overcome the problem
discussed. However, Energy Management System (EMS) which is an advanced
feature of EnergyPlus enables one to write custom programmes to describe specific
control algorithms in a language called EnergyPlus Runtime Language (ERL) (US
Department of Energy, 2013a). Such code could be added directly to the
EnergyPlus’s IDF file to override the existing default control. An ERL code was
initially written for this purpose which could be found in appendix A.2. The code was
written in order to shut down the hot water supply from the boiler at any time step
when the air temperature in the ground floor hallway (where PRT was located)
increased above its set-point temperature of 21 °C. However, it was found that
adding such code to better represent the control strategy requires accurate
predictions of the air temperature. As it will be discussed in sections 6.3 and 6.4, it
was not possible to accurately predict the hallway ground floor air temperature due
to complexities involved with modelling the air flow between the ground floor and first
floor hallways. Therefore, after running a number of simulations and compare the
predictions with the default control strategy, it was decided not to use the ERL code
as it could not increase the accuracy in this case when the air temperatures could
not be accurately predicted.
5.5 Modelling the occupancy
There was no occupancy during the co-heating test. All the internal doors in the
model were set 100% open while all the windows and external doors were set 100%
closed as was the case throughout the co-heating test. All window blinds were
modelled open for the whole simulation period as it was during the test.
Modelling the occupancy for the HT1 was also straightforward as the synthetic
occupancy presented was fully known. The electricity use measured in each zone
was used to model the lighting and equipment gains in the modelled zone. The fan
heaters used to represent heat gains in the kitchen were added as electric
equipment with 100% convective heat. The oil filled radiators were also added as
electric equipment but with a radiant fraction of 0.3. All the other lighting devices
were added as lights with 0.42 radiant and 0.18 visible fractions.
129
All the external doors and windows were closed for the whole simulation period and
the operation of the internal doors were set in the model according to their operation
in real test houses described in section 3.3.5. Operation schedule of the window
roller blinds were set according to their real schedule explained in section 3.3.5.
5.6 Weather file Construction
It is important that the weather parameters in the model represent the real weather
conditions at the test houses during the experimental period for comparing the model
predictions and the measured data from the experiments. The EnergyPlus weather
converter programme was used to create weather files for the test periods. Hourly
data derived from weather stations were: dry bulb temperature (ºC), dew point
temperature (ºC), relative humidity (%), atmospheric pressure (pa), direct normal
solar radiation (Wh/𝑚𝑚2 ), diffuse horizontal solar radiation (Wh/𝑚𝑚2 ), wind direction
(degree), wind speed (m/s), total sky cover (tenth) and snow depth (cm).
Weather parameters required were measured on site or sourced from either: the
Centre for Renewable Energy Systems Technology (CREST) weather station at
Loughborough University, 2 km from the test houses; Sutton Bonnington, 7.5 km
from the test houses; or Nottingham Watnall (26 km from the test houses). Sutton
Bonnington and Nottingham Watnall weather data for the period of experiments were
sourced via MIDAS Land Surface Observation database at the British Atmospheric
Data Centre (BADC) operated by the UK Meteorological Office (2012).
Hourly dry bulb temperature was measured outside the test houses during all tests
(see section 4.2). Hourly dew point temperature, wind speed, wind direction and
humidity were sourced from Sutton Bonnington weather station. Cloud cover and
atmospheric pressure data were sourced from Nottingham Watnall. Hourly Wind
speed in Knots and the amount of cloud cover in Oktas13 were converted to m/s and
tenths respectively. The following criteria were used to convert the amount of cloud
cover in Oktas to tenth (BADC, 2014):
13
Although cloud amount has been measured in eighths (or Oktas) since 1949 (BADC, 2014), EnergyPlus still
uses the old format of cloud cover data (i.e. tenths of coverage).
130
Table 5-10: Conversion factors of cloud cover from Oktas to tenth
Value in Oktas
0
1
2
3
4
5
6
7
8
Equivalent value in tenths
0
2
3
4
5
6
8
9
10
Direct Normal Radiation (DNR) is the amount of solar radiation in Wh/m2 received
directly from the solar disk on a surface perpendicular to the sun’s rays; Diffuse
Horizontal Radiation (DHR) is the amount of solar radiation in Wh/m2 received from
the sky (excluding the solar disk) on a horizontal surface, and the Global Horizontal
Radiation (GHR) is the total amount of direct and diffuse solar radiation in
Wh/m2 received on a horizontal surface.
Hourly GHR and DHR were measured at Centre for Renewable Energy Systems
Technology (CREST) at Loughborough University and used to derive DNR.
DNR can be calculated for each hour from GHR and DHR measurements using
equation (5-7):
𝐷𝐷𝐷𝐷𝐷𝐷 =
𝐺𝐺𝐺𝐺𝐺𝐺 − 𝐷𝐷𝐷𝐷𝐷𝐷
cos(𝜃𝜃𝑧𝑧 )
(5-7)
Where:
𝜃𝜃𝑧𝑧 = solar zenith angle and can be calculated using equation (5-8):
𝑐𝑐𝑐𝑐𝑐𝑐𝜃𝜃𝑧𝑧 = 𝑐𝑐𝑐𝑐𝑐𝑐∅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 + 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
(5-8)
Where:
∅ = latitude for the location where the test houses were located.
𝛿𝛿 = solar declination and can be calculated according to equation (5-9):
𝛿𝛿 = 23.45 sin �360 ∗
284 + 𝑛𝑛
�
365
(5-9)
131
Where:
n = day of the year.
𝜔𝜔 = solar hour angle which is the angular displacement of the sun east or west of the
local meridian due to rotation of the earth on its axis at 15º per hour; morning
negative, afternoon positive.
In this research DNR was automatically calculated using the weather converter
programme of EnergyPlus by inserting GHR and DHR. Snow depth was considered
as zero since there was no snow on the ground during the period of the experiments.
Table 5-11 summarizes the sources of weather data used in this study.
Table 5-11: Summary of hourly weather parameters, their units and sources of data
Parameter
unit
Source
Dry bulb temperature
ºC
Measured locally outside the test houses
Dew point temperature
ºC
Sutton Bonnington weather station
Relative humidity
%
Sutton Bonnington weather station
Global horizontal
W/𝑚𝑚2
Measured at Loughborough university
radiation
Direct normal radiation
Diffuse horizontal
W/𝑚𝑚
2
campus, CREST
Derived from global and direct normal
horizontal radiation using EnergyPlus
weather converter programme
Measured at Loughborough university
Wind direction
W/𝑚𝑚2
Degree
Sutton Bonnington weather station
Wind speed
Knots
Sutton Bonnington weather station
Total sky cover
Oktas
Nottingham Watnall weather station
Snow depth
cm
radiation
Atmospheric Pressure
Hecto
Pascals
campus, CREST
Considered as zero for the whole tests
period
Nottingham Watnall weather Station
These parameters were inserted into a CSV file which then was imported in
EnergyPlus weather convertor programme to generate the EPW file. Latitude,
longitude and elevation of the test houses were found using Google earth (2015) and
132
inserted in a separate “definition” (.def) file. This definition file should be saved with
the same name as the CSV file and is needed by the weather converter programme
for the conversion process.
5.7 Summary
This chapter described the dynamic thermal modelling tools, techniques and the
input parameters which were used to model the co-heating test and the space
heating trial 1 (HT1). This included modelling the building envelope of the test
houses, the air flow modelling strategies employed, the heating systems used during
the tests and their control strategies as well as occupancy profiles. It also describes
the method used to construct a weather file which was used to simulate the coheating test and the HT1 including the weather parameters used and their sources.
The results from modelling the co-heating test and the HT1 will be compared in
chapter 6.
133
6 Comparison of the DTM predictions and
measurements: DTM calibration
6.1 Introduction
In this chapter, the results from modelling the co-heating test and HT1 are compared
to the experimental results. Firstly, in section 6.2, the energy uses measured during
the co-heating test are compared to those predicted using each air flow modelling
strategy. Then in section 6.3 the measured and predicted energy uses and indoor air
temperatures of the LMP1930 test houses during the HT1 are compared. Section 6.4
describes the calibration procedure which was conducted to achieve a calibrated
model. Finally, section 6.5 provides a summary of this chapter.
6.2 Comparison for the co-heating test
In this section, the measured energy use of the houses during the co-heating test are
compared with the energy use predicted using the model with Scheduled Natural
Ventilation (SNV) (section 6.2.1) and Air Flow Networks (AFN) (section 6.2.2).
6.2.1 Model with SNV
The total hourly electricity consumption predicted by the model for each house was
compared to that measured during the 9 days of the co-heating test (Figure 6-1). The
comparison showed that the predictions have a similar trend to the measurements.
In both cases the electricity use decreases when the outdoor air temperature
increases and vice versa. A strong negative relationship between the amount of
hourly global horizontal solar radiation (W/𝑚𝑚2 ) and electricity use of the test houses
was observed (Figure 6-2). During the daytime, when the solar radiation was at its
peak, the energy consumption dropped to its minimum for that day. Generally, during
the days when the solar radiation was higher, the outdoor air temperature was also
higher and the energy consumption was lower compared to days when the solar
radiation was lower. During the night, when there was no solar gain, the
temperatures dropped and the amount of energy use was considerably increased.
134
In House 1, some discrepancies were found between the predictions and the
measurements of energy use during days 7 and 8 where the model underestimated
the energy use (Figure 6-1). The average wind speed during day 7 and day 8 were
4.6 m/s and 3.2 m/s, respectively, compared to the average wind speed of 2.3 m/s
for the rest of the co-heating period (Figure 6-3). Therefore, the discrepancies could
be explained as when the wind speed is higher, the rate of heat loss through
infiltration increases while the model assumed the same rate of infiltration regardless
of the wind speed.
For House 2, Figure 6-1 shows that the model slightly overestimates the energy use
for the whole period. This is in line with the results of the co-heating test in
section 3.3.6, where it was found that the total heat loss coefficient of the House 2
was 5.6% lower than House 1. By assuming the same construction for both houses
in the model, the predicted energy use of the house 2 was higher during the coheating test due to its lower total heat loss coefficient. It was important to model the
houses with the same construction as it was not clear which parts of the fabric are
responsible for the differences observed. It was unlikely that every part of the fabric
contributed the same to the whole house better thermal performance.
135
Figure 6-1: Whole house hourly electricity consumption measured in House 1 and 2
compared with the model prediction along with the hourly outdoor air temperature
(SNV)
136
House 1
10
300
7
200
6
5
150
4
100
3
2
50
)
250
8
Hourly global horizontal solar radiation (W/
Whole house hourly electricity consumption (kWh)
9
1
0
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
0
00:00
Day
Model
Measured
Hourly global horizontal solar radiation
House 2
10
300
250
)
8
7
200
6
5
150
4
100
3
2
50
Hourly global horizontal solar radiation (W/
Whole house hourly electricity consumption (kWh)
9
1
0
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
0
00:00
Day
Model
Measured
Hourly global horizontal solar radiation
Figure 6-2: Whole house hourly electricity consumption measured in House 1 and 2
compared with the model prediction along with the hourly global horizontal solar
radiation (SNV)
137
House 1
10
8
7
8
6
7
5
6
5
4
4
3
Wind speed (m/s)
Whole house hourly electricity consumption (kWh)
9
3
2
2
1
1
0
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
0
00:00
Day
Model
Measured
wind speed (m/s)
House 2
10
8
7
8
6
7
5
6
5
4
4
3
Wind speed (m/s)
Whole house hourly electricity consumption (kWh)
9
3
2
2
1
1
0
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
0
00:00
Day
Model
Measured
Wind speed (m/s)
Figure 6-3: Whole house hourly electricity consumption measured in House 1 and 2
compared with the model prediction along with the hourly wind speed (SNV)
138
The infiltration rate of each zone in the ground and first floors was calculated by
DesignBuilder as described in section 5.3.1 according to equation (5-2). The
underlying assumptions of the calculation method were reflected in the results.
Infiltration rate of the zones with more than one exposed surface including living
rooms, hallway ground floors, kitchens, bedroom 2, unoccupied bedrooms and
bathrooms were calculated as 1.3 ACH while this was calculated as 0.9 for the
zones with only one exposed surface including dining rooms, bedroom 1, hallway
first floors and WCs. The infiltration rates of the subfloors and the roof were 8.0 and
2.7 ACH, respectively, as they were explicitly defined.
The difference between daily electricity use predicted by the model and the
measured daily electricity use varied from -6% to +1% for House 1 and from -1% to
+8% for House 2 (Figure 6-4). On average, for the whole co-heating test period, the
difference between daily electricity consumption predicted and measured was 0.1%
and 4.8% in House 1 and 2, respectively.
139
Figure 6-4: Measured and predicted whole house daily electricity consumption in
House 1 and 2 during the co-heating test (SNV)
140
The ASHRAE acceptance criteria for the calibration of building simulation models
described in section 2.7.3 showed that the models of both houses met the
requirements for both criteria of MBE and CVRMSE (Table 6-1).
Table 6-1: MBE (%) and CVRMSE (%) calculated and their acceptable limit (coheating test with SNV)
House 1
House 2
Acceptable limit
MBE (%)
0.9%
4.9%
10%
CVRMSE (%)
5.4%
7.9%
30%
6.2.2 Model with AFN
The model was re-run using AFN instead of SNV. The predicted room by room
infiltration rate and the whole house infiltration rate was not comparable to the model
in SNV or measured results from the airtightness test. This was due to the different
methodology of AFN for calculating air flows compared to SNV. In AFN, for each
crack or opening in any exterior surface, the model predicts the air volume flow rate
from outdoors to the thermal zone associated with that specific crack or opening. In
addition, AFN reports the air volume flow rates in the reverse direction (i.e. from a
thermal zone to outdoors). AFN also reports the air volume flow rates from each
zone to its adjacent zones through interior surfaces (inter-zone air flow). These air
volume flow rates in AFN are not constant like SNV and they change from one time
step to another according to the variations in the wind and stack effects.
In total, there were more than 200 cracks and openings in the LMP1930 model.
Hourly air flows from outdoors to each zone (𝑚𝑚3 /ℎ𝑟𝑟) was calculated as the sum of
hourly air flows (𝑚𝑚3 /ℎ𝑟𝑟) in the direction of outdoors to indoors through all the cracks
and openings in all exterior surfaces of the zone. An average air infiltration rate (ach)
for the co-heating test period was achieved for each zone by averaging the hourly air
flows from outdoors to the zone divided by the volume of the zone. Similarly, an
average exfiltration rate (ach) for the co-heating test was calculated for each zone
considering the air flows in the reverse direction (i.e. from the thermal zones to
outdoors). The average infiltration and exfiltration of each zone of the LMP 1930 test
houses during the co-heating test period were reported in Table 6-2.
141
Table 6-2: Zone by zone average infiltration rate and exfiltration rate for the
LMP1930 test houses calculated by AFN
House 1
Zone
House 2
Average
Average
Average
Average
infiltration rate
exfiltration rate
infiltration rate
exfiltration rate
(ach)
(ach)
(ach)
(ach)
Living room
0.23
0.37
0.23
0.37
Dining room
0.55
0.05
0.56
0.04
Kitchen
1.65
0.25
1.2
0.4
Hallway
2.02
0.18
1.54
0.26
Hallway first floor
0.02
1.28
0
0.8
Bedroom 1
0
0.6
0
0.6
Bedroom 2
0
1.7
0
1.7
Unoccupied room
0.15
2.15
0
2.3
Bathroom
0.02
1.88
0
1.9
WC
0
3.5
0
2.1
Subfloor
21
0
21
0
Roof
0
2.1
0
2.1
As it can be seen from Table 6-2, the AFN predicted that the air was coming from
outdoors to inside the building mainly through the subfloor air bricks and the ground
floor cracks and openings. Average infiltration rates of near to zero for the rooms at
the first floor and the roof, show that the amount of air which flows from outdoors to
indoors through the first floor rooms and the roof is negligible. The air was mainly
escaping to outside through the cracks in the exterior surfaces of the first floor and
the roof.
The AFN predictions of how the air was flowing in the LMP1930 houses during the
co-heating test proved the significant effect of stack ventilation compared to wind
induced ventilation. The indoor air at temperatures of about 25°C maintained during
the co-heating test was considerably warmer and thus less dense than the colder
outdoor air. This causes a significant pressure difference during the whole period of
the co-heating test in which the air entering the building was continuously heated.
The warm, less dense air which was trying to rise and escape from the cracks at
higher levels of the building (i.e. first floor and the roof) was drawing the cold dense
air into the cracks at the lower levels (i.e. subfloor and the ground floor) (Figure 6-5).
142
Negative Pressure
Warm air which is less
dense escapes from the
first floor and the roof
Neutral Pressure
Cold dense air enters the
houses from the subfloor
and the ground floor
Positive Pressure
Figure 6-5: Schematic of the pressure distribution and the air flows in the LMP1930
test houses during the co-heating test
The energy use predictions by the AFN model were compared with the
measurements in the same way as the predictions from the model with SNV
(Figure 6-6). The AFN model underestimated the hourly electricity consumption
during the whole co-heating test period for both houses. The calculated MBE and
CVRMSE were higher than for the model with SNV (Table 6-3). However, the energy
use predictions of the model still met the ASHRAE calibration criteria for both houses.
143
House 1
10
14
12
8
10
7
6
8
5
6
4
3
Out door air temperature (ºC)
Whole house hourly electricity consumption (kWh)
9
4
2
2
1
0
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
0
00:00
Day
Model
Measured
Outdoor Air Temperature
House 2
9
14
12
7
10
6
5
8
4
6
3
4
Outdoor air temperature (ºC)
Whole house hourly electricity consumption (kWh)
8
2
2
1
0
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
00:00
0
00:00
Day
Model
Measured
Outdoor Air Temperature
Figure 6-6: Whole house hourly electricity consumption measured in House 1 and 2
compared with the model prediction along with the hourly outdoor air temperature
(AFN)
144
Table 6-3: MBE (%) and CVRMSE (%) calculated and their acceptable limit (Coheating test with AFN)
House 1
House 2
Acceptable limit
MBE (%)
9.0%
5.4%
10%
CVRMSE (%)
10.7%
8.0%
30%
Comparing the results of the models with SNV and AFN, it was concluded that in this
case, AFN was better able to represent wind pressures and the stack ventilation
effects. However, the magnitude of air flows and the overall building heat transfer
was better represented by SNV based on the energy demand results. It was not
possible to determine if this would also be the case for an intermittently heated
building as in the HT1. Therefore, both air flow modelling strategies were employed
to simulate the HT1 and the results compared.
6.3 Comparison for the Heating Trial 1
In this section, the measured and modelled energy demands (section 6.3.1) and
indoor air temperatures (section 6.3.2) of the houses during the HT1 are compared
and the potential reasons for any discrepancies are discussed.
6.3.1 Comparison of the energy demands
Daily boiler heat output measured during the HT1 was compared with model
predicted daily boiler heat output using both air flow modelling strategies
(Figure 6-7and Figure 6-8). In the house with ZC (Figure 6-7) the predicted daily
boiler heat outputs with either of the air flow modelling strategies were lower than the
measured daily boiler heat output for the majority of the days. For the whole HT1, the
model with SNV under-predicted the total boiler heat output in the house with ZC by
8% while the model with AFN under-predicted by 23%.
145
120
House with ZC
Measured daily boiler heat output
Simulated daily boiler heat output (AFN)
Simulated daily boiler heat output (Scheduled nat vent)
100
80
60
40
20
15/03/2014
14/03/2014
13/03/2014
12/03/2014
11/03/2014
10/03/2014
09/03/2014
08/03/2014
07/03/2014
06/03/2014
05/03/2014
04/03/2014
03/03/2014
02/03/2014
01/03/2014
28/02/2014
27/02/2014
26/02/2014
25/02/2014
24/02/2014
23/02/2014
22/02/2014
21/02/2014
20/02/2014
19/02/2014
18/02/2014
17/02/2014
16/02/2014
0
Figure 6-7: Measured and predicted daily boiler heat output during Heating Trial 1 in
house with ZC
In the CC house (Figure 6-8), model predictions were closer to the measured boiler
heat outputs. As for the house with ZC, the model with SNV predicted higher daily
boiler heat outputs than the model with AFN. The difference between the measured
and predicted boiler heat demand in the house with CC was 0.5% and 11% for the
models with SNV and AFN respectively.
146
Measured daily boiler heat output
Simulated daily boiler heat output (AFN)
Simulated daily boiler heat output (Scheduled nat vent)
House with CC
120
100
80
60
40
20
15/03/2014
14/03/2014
13/03/2014
12/03/2014
11/03/2014
10/03/2014
09/03/2014
08/03/2014
07/03/2014
06/03/2014
05/03/2014
04/03/2014
03/03/2014
02/03/2014
01/03/2014
28/02/2014
27/02/2014
26/02/2014
25/02/2014
24/02/2014
23/02/2014
22/02/2014
21/02/2014
20/02/2014
19/02/2014
18/02/2014
17/02/2014
16/02/2014
0
Figure 6-8: Measured and predicted daily boiler heat output during Heating Trial 1 in
house with CC
Hourly analysis of the predicted and measured boiler heat outputs showed that none
of the models could be considered calibrated according to ASHRAE hourly
calibration criteria (Table 6-4). Although MBE (%) calculated for both houses were
within the 10% limit for the model with SNV (8% and -0.4% for the house with ZC
and CC respectively), they exceeded the limit for both houses using the model with
AFN (23% and 11% for the ZC and CC house respectively). CVRMSE (%) calculated
for both houses were above the 30% accepted limit using SNV and AFN.
Table 6-4: MBE (%) and CVRMSE (%) calculated for each house and their
acceptable limit using each air flow modelling strategy
ZC House
CC House
Acceptable
SNV
AFN
MBE (%)
8%
23%
10%
CVRMSE (%)
35%
45%
30%
MBE (%)
-0.4%
11%
10%
CVRMSE (%)
39%
44%
30%
limit
147
Lower energy consumption in rooms with a radiator was predicted by most building
energy simulation programs tested by Lomas et al. (1997) in the International Energy
Agency (IEA) report.
The energy savings in boiler heat output of ZC predicted by the DTM with SNV and
AFN were 26% and 21% respectively. These were considerably higher than the
measured 14.5% and therefore further work was needed to understand the
differences and calibrate the model. This was addressed in section 6.4.
6.3.2 Comparison of the indoor air temperatures
Measured and predicted indoor air temperatures were compared for each room of
the house with ZC (Figure 6-9) and the house with CC (Figure 6-10). Outdoor air
temperatures, and global horizontal solar radiation for the south facing rooms, were
added to the plots to aid understanding. The heating on hours, heating off hours,
occupied and unoccupied hours, set-point and set-back temperatures were indicated
on each plot. These plots were inspected visually to identify repeating patterns of
discrepancies between the measured and predicted air temperatures.
The plots presented here were for the three consecutive days; two weekdays
(Thursday 27 and Friday 28 February 2014), and one weekend day (Saturday 1
March 2014) to include different heating schedules used at weekdays and weekends.
The weekdays represent days with higher (Thursday) and lower (Friday) levels of
solar radiation: the average daily global horizontal solar radiation for weekday 1 and
weekday 2 were 116 and 52 W/𝑚𝑚2 respectively compared to the average of 95 W/𝑚𝑚2
for the whole HT1. Daily average outdoor temperature, global horizontal solar
radiation and wind speed for the selected days and the whole HT1 period were
presented in Table 6-5.
148
Table 6-5: Weather parameters for the selected days and the whole HT1
Whole
Weather parameter
Thursday
Friday
Saturday
test (28
days)
average outdoor temperature
(°C)
average global horizontal solar
radiation (W/𝑚𝑚2 )
average wind speed (m/s)
5.2
2.5
2.5
6.2
116
52
94
95
4.6
2.8
1.5
3.8
149
Measured Indoor Air Temperature
Simulated Indoor Air Temperature (Scheduled Nat.Vent.)
Simulated Indoor Air Temperature (Air Flow Network)
Outdoor Air Temperature
Global horizontal solar radiation
Set-point
Set-back
Occupied hours
'Heating on' hours
Living room
Indoor air temperature (ºC)
Indoor air temperature (ºC)
Weekday 1, Thursday 27/02/2014 Weekday 2, Friday 28/02/2014 Weekend, Saturday 01/03/2014
25
W/𝑚𝑚^2
700
20
600
500
15
400
10
300
5
200
0
-5
25
100
Hour
0
Dining room
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Hallway - ground floor
600
20
500
15
400
10
300
5
200
0
100
Indoor air temperature (ºC)
-5
25
W/𝑚𝑚^2
700
0
Kitchen
20
15
10
5
0
-5
150
Indoor air temperature (ºC)
Indoor air temperature (ºC)
25
Unoccupied bedroom
20
15
10
5
0
-5
25
W/𝑚𝑚^2
700
600
500
400
300
200
100
0
Bedroom 1
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Bedroom 2
600
20
500
15
400
10
300
5
200
0
100
Indoor air temperature (ºC)
-5
25
W/𝑚𝑚^2
700
0
Bathroom
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Hallway - first floor
20
15
10
5
0
-5
Figure 6-9: predicted and measured indoor air temperatures of the house with ZC
along with measured outdoor air temperatures and global horizontal radiation; 27
Feb to 1 March 2014
151
Measured Indoor Air Temperature
Simulated Indoor Air Temperature (Scheduled Nat.Vent.)
Simulated Indoor Air Temperature (Air Flow Network)
Outdoor Air Temperature
Global horizontal solar radiation
Set-point
Occupied hours
'Heating on' hours
Indoor air temperature (ºC)
Indoor air temperature (ºC)
Living room
Weekday 1, Thursday 27/02/2014 Weekday 2, Friday 28/02/2014 Weekend, Saturday 01/03/2014
W/𝑚𝑚^2
25
700
600
20
500
15
400
10
300
5
200
0
-5
25
100
Time
0
Dining room
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Hallway - ground floor
600
20
500
15
400
10
300
5
200
0
100
Indoor air temperature (ºC)
-5
25
W/𝑚𝑚^2
700
0
Kitchen
20
15
10
5
0
-5
152
Indoor air temperature (ºC)
Indoor air temperature (ºC)
25
Unoccupied bedroom
20
15
10
5
0
-5
25
W/𝑚𝑚^2
700
600
500
400
300
200
100
0
Bedroom 1
20
15
10
5
0
-5
Indoor air temperature
(ºC)
25
Bedroom 2
20
600
15
400
10
5
200
0
Indoor air temperature (ºC)
-5
25
W/𝑚𝑚^2
800
0
Bathroom
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Hallway - first floor
20
15
10
5
0
-5
Figure 6-10: predicted and measured indoor air temperatures of the house with CC
along with measured outdoor air temperatures and global horizontal radiation; 27
Feb to 1 March 2014
153
During the “heating on” hours, indoor air temperatures predicted using SNV and AFN
followed a similar pattern to those measured in each room and in both houses. Air
temperatures rise when the heating comes on, and with CC continued to increase
until the set-point temperature was achieved (Figure 6-10). However, with ZC, when
the room was not scheduled to be occupied, the air temperature only increased until
the set-back temperature of the room was achieved (Figure 6-9). When the room
was scheduled to be occupied, the room air temperature increased to its set-point
temperature (Figure 6-9). This demonstrates that the heating schedules in the model
were similar to those in the real test houses. However, discrepancies were observed
between the predicted and measured hourly air temperatures which were persistent
throughout the test period. These differences could be divided into two categories:
differences during the” heating on” hours and differences when the heating was off.
•
Differences during “heating on” hours
When scheduled to be occupied, living room air temperatures measured in both
houses exceeded the nominal set-point of 21°C (Figure 6-9 and Figure 6-10) as the
PTRVs and the TRVs did not maintain the room air temperatures accurately.
Temperatures of up to 25°C were recorded during the evening hours when the door
was closed and there was high level of internal heat gains. High temperatures were
also recorded when there was high level of solar radiation. The measured radiator
surface temperatures indicated that their heat output continued even when the
rooms were above their set-point temperatures (see Figure 4-7).
Similarly, temperatures achieved in the dining rooms of both houses were slightly
higher than the nominal set-point temperatures assumed in the model when they
were scheduled to be occupied and with internal heat gains and the doors closed.
During the rest of the heating on hours, predicted dining room air temperatures were
relatively close to those measured.
The ability of TRVs to maintain a set-point temperature was found to vary between
rooms. The measured and predicted air temperatures in bedroom 1 and 2 were
similar in the house with CC, while slightly different in the house with ZC.
154
The unoccupied bedrooms were only heated when their air temperature dropped
below 12°C. The doors were also always closed and the model using AFN better
predicted air temperatures.
Detailed operation of TRVs and PTRVs cannot easily be modelled in Design Builder
or EnergyPlus. This could be a significant source of inaccuracy as this difference
would affect the rate of heat transfer to adjacent rooms and outdoors as well as the
accuracy of the predicted air temperatures. Therefore, better predicted air
temperatures might be achieved during the heating on hours by changing the
nominal set-points in the model to an average of the measured air temperature for
each room.
Solar gains also played a role in the differences observed between the predicted and
measured air temperatures. The measured air temperatures in the south facing
rooms with large windows (i.e. living room and bedroom 2) were higher than those
predicted during sunny days (Figure 6-9 and Figure 6-10). Since the glazing and
wooden frame area of the windows were accurately measured and inserted in the
model as described in section 5.2.1, differences observed could be attributed to one
or more than one of the following reasons:
a) Solar transmittance of the glazing might be assumed low in the model.
b) Solar absorptance of the floor might be assumed low in the model.
(EnergyPlus assumes that all direct normal solar radiation entering a zone
falls on the floor (US Department of Energy, 2012)).
c) Ground reflectance values which are used to calculate the ground reflected
solar radiation might be assumed low in the model.
d) Differences between the amounts of solar radiation measured at weather
station compared to actual on site solar radiation.
e) Errors involved in measuring air temperature under high solar radiation using
a thin layer of aluminium foil to protect the temperature sensor from direct
solar radiation.
155
There were two unheated rooms in each house: kitchen and first floor hallway 14.
Predicted air temperatures in both of these rooms were found to be lower than the
measured air temperatures during the heating on hours. In the two houses, the boiler
was located in the kitchens and the pipes were uninsulated (Figure 6-11). The
additional heat gains from the boiler casing and its associated pipe work were not
included in the model. The heat loss from a boiler casing and associated pipe work
and fittings could be considered to be about 2% of the boiler’s rated output (Vesma,
2014). This would result in 600 W additional heat gains in the kitchens when the
heating was on.
Boiler
Temperature
sensor
Figure 6-11: Boiler and its uninsulated pipe work and the position of temperature
sensor on a tripod in the kitchen of House 2
The lower predicted air temperature of the hallway first floor compared to the
measured air temperatures was believed to be due to the difficulties in modelling
14
WCs were ignored in this analysis due to its relatively small floor area and the fact that their air temperatures
were not measured during the HT1
156
natural convection through the staircase which connected the ground floor and first
floor hallways of each house. Both air flow modelling strategies could only very
poorly represent the air flows through large horizontal openings. According to
DesignBuilder’s help document, “the air flow between two floors connected by large
horizontal openings (i.e. holes) could be only modelled “very approximately” when
using the AFN”. According to EnergyPlus input output reference (US Department of
Energy, 2013c), the AFN model is unable to model bi-directional flows through large
horizontal openings at a given time step.
Inaccuracies of modelling natural convection through staircase would also cause
inaccuracies in predicted air temperatures of the ground floor hallway. The measured
hallway air temperature of the house with ZC was below 21°C for most of the heating
period (Figure 6-9) which was lower than model prediction. This could be explained
as a large proportion of the heat emitted from the radiator in the ground floor hallway
was transferred to the hallway first floor and its adjacent bedrooms (as well as colder
rooms in the ground floor). In the house with CC (Figure 6-10), the hallway ground
floor air temperature predicted and measured during the heating on hours were very
close to the nominal set-point temperature of 21°C. This can be explained as in the
house with CC, since the first floor rooms were also heated during the heating on
hours, the rate of heat loss from the ground floor hallway to the first floor hallway was
considerably lower than the house with ZC.
One alternative method could be to consider the ground floor and first floor hallways
as a single zone. However, according to EnergyPlus documentation (US Department
of Energy, 2013c) AFN cannot model the air temperature stratification within a
thermal zone which is the case if the hallway would have been considered as a
single zone. Another alternative is to increase the air flow from the ground floor to
the first floor by increasing discharge coefficient of the hole connected the two floors
when using AFN or to increase the amount of air mixing between the ground floor
and first floor when using SNV air flow modelling.
•
Differences during “heating off” hours
When the heating turned off, the predicted air temperatures fell at a faster rate than
was measured. This was true in all of the rooms, for both houses and regardless of
157
the air flow modelling method. The first potential reason could be higher fabric heat
loss or higher infiltration heat loss assumed in the model. However, the model
showed a reasonable prediction of the overall heat loss due to fabric and infiltration
when modelling the co-heating test and overall, the predicted energy use was lower
than the measured energy use in both houses regardless of the air flow modelling
strategy. Therefore, reducing the fabric or ventilation heat loss would not improve the
model.
Figure 6-9 and Figure 6-10 show that predicted rates of heat up are also higher than
measured. This could potentially be due to lower thermal mass in the model than in
the real building. These fast heat up and cool down rates in rooms heated with
radiators were similar to the findings of others. Zhai & Chen (2005) used
experimental data from IEA annex 21/task 12 15 reported by Lomas et al. (1997) to
simulate natural convection in a room with an oil-filled radiator controlled via a PID
controller. They found that the difference between the predicted and measured air
temperatures of the room with a radiator were significant during the heat up and cool
down. Similar findings were reported by Beausoleil-Morrison (2000). Figure 6-12
which is adopted from Zhai and Chen (2005) shows the predicted and measured
mean air temperature for the IEA test room with radiator for their study (a) and study
by Beausoleil-Morrison (2000) (b).
15
International Energy Agency (IEA) annex 21/task 12 was conducted for the purpose of empirical
validation of building energy simulation programs
158
Figure 6-12: predicted and measured mean air temperature over a single day for the
IEA test room with radiator for (a) study by Zhai and Chen (2005) and (b) study by
Beausoleil-Morrison (2000) (Figure was reproduced from Zhai and Chen (2005)).
Zhai and Chen (2005) argue that the higher rates of air temperature change
predicted in the model is because of the dynamic behaviours of the radiators: the
time delay as the water warms or cools when the heater is switched on or off cannot
be represented. This causes the air to heat up much faster in the model than it does
in reality.
159
To sum up, ten input parameters were identified as those which could have
potentially influenced the discrepancies observed between the predicted and
measured indoor air temperatures and will be investigated further:
1. Set-point temperatures of the rooms in which the heating was controlled by
TRV or PTRV
2. Unaccounted heat gains in the kitchens from the boiler casing and pipe work
3. The amount of air flow between the ground floor and first floor hallways
4. Hallway zoning strategy
5. Ground reflectance
6. Solar transmittance of the glazing
7. Solar absorptance of the floor materials
8. Building fabric heat loss
9. Infiltration heat loss
10. Thermal mass
6.4 Model calibration
Indoor air temperatures and boiler heat outputs measured during the HT1 were used
to calibrate the model. The calibration procedure consisted of three steps:
1. Sensitivity analysis was conducted to evaluate the effects of the 10 parameters
proposed in section 6.3.2 on improving the model’s predictions of energy use
and indoor air temperatures.
2. The parameters which had the potential to improve the predictions of both
energy and indoor air temperature were adjusted in the base case model to
generate a refined model.
3. The refined model was assessed against the acceptance criteria for hourly
calibration of building energy simulation models according to ASHRAE
Guideline 14. In addition, the hourly indoor air temperatures measured and
predicted were plotted and inspected visually as an additional check in order to
identify any discrepancies.
The calibration procedure was applied to two versions of the base case model: one
using SNV, and one using AFN to model the air flows through the houses. Ten
160
variants for the LMP1930 test house model were constructed. For each variant, all of
the model inputs were exactly the same as the base case model except for the one
parameter being studied. This parameter was altered, within a reasonable range, to
investigate if it improved the accuracy of energy use and indoor air temperature
predictions. The ten variants are described below:
Variant 1 was constructed to evaluate the effects of changing the set-point
temperatures in the rooms in which the heating was controlled by TRV or PTRV: In
ZC house, the nominal set-point temperatures assumed in the base case model was
replaced with the average air temperature measured during the occupied hours in
each room (Table 6-6). In CC house, the average air temperatures measured in
each room during the heating on hours, except the first hour of each heating period
(warm up periods) were replaced the nominal set-point temperatures.
Table 6-6: Nominal and new set-point temperatures which were applied for variant 1
Room
Set-point temperature ZC (°C)
Set-point temperature CC (°C)
Nominal
New
Nominal
New
Living room
21.0
22.1
21.0
23.0
Dining room
19.0
20.3
19.0
20.1
Bedroom 1
19.0
19.9
19.0
19.5
Bedroom 2
19.0
19.0
19.0
19.0
Bathroom
21.0
19.7
21.0
18.7
Variant 2 was constructed to evaluate the effects of adding a heat emitter to the
kitchen of the two houses in order to represent the kitchen heat gains from the boiler
casing and its associated pipe work. A radiator was added with a rated capacity of
600 W and was scheduled to be always on when the heating was on.
Variant 3 was constructed to evaluate the effect of increasing the air flow between
the ground floor and the first floor hallways. The discharge coefficient of the opening
was changed from 0.65 to 0.72 (10% increase) for the version of the model which
used AFN to model the air flows. For the version of the model which used SNV, the
design flow rate between the ground floor and first floor hallway increased by 10%.
161
Variant 4 was constructed to evaluate the effect of modelling the ground floor and
first floor hallways as a single zone instead of two separate zones.
Variant 5 was constructed to evaluate the effects of higher ground reflectance by
increasing the monthly ground reflectance of the model from 0.2 to its maximum
value of 1.0.
Variant 6 was constructed to evaluate the effects of higher solar transmittance of the
glazing by increasing it by 10% from 0.837 to 0.92.
Variant 7 was constructed to evaluate the effects of higher solar absorptance of the
floor materials (i.e. carpet and timber floor) by increasing each of them by 10%.
Variant 8 was constructed to evaluate the effects of lower building fabric heat loss.
The conductivities of the two layers of external walls were reduced by 30% each.
This resulted in a reduction of 17% in the U-values of the external walls.
Variant 9 was constructed to evaluate the effect of lower infiltration heat loss. For
the version of the model using AFN, the poor crack template was replaced by the
medium crack template (Table 6-7 and Table 6-8).
Table 6-7: New crack characteristics according to DesignBuilder’s “medium” crack
template used in the variant 9 for walls, floors and the roof
Air mass flow
coefficient (𝑪𝑪𝑸𝑸 )
Flow exponent (n)
0.0001
0.7
Internal walls
0.003
0.75
Internal floors
0.0009
0.7
External floors
0.0007
1.0
Roof
0.0001
0.7
Building element
External walls
(Kg/s.𝒎𝒎𝟐𝟐 ) at 1Pa
162
Table 6-8: New crack characteristics according to DesignBuilder’s “medium” crack
template used in the variant 9 for the doors, windows and vents
Air mass flow
coefficient (𝑪𝑪𝑸𝑸 ) (Kg/s.
Flow exponent (n)
External windows
0.00014
0.65
External doors
0.0014
0.65
Internal doors
0.02
0.6
External vents
0.008
0.66
Building element
m) at 1Pa
For the version of the model which used SNV, the infiltration design flow rate of each
room was decreased by 10%.
Variant 10 was generated in order to investigate the effects of assuming higher
thermal mass in the model on the predictions of energy use and indoor air
temperature. The “Temperature Capacity Multiplier” object of EnergyPlus was used
to increase the thermal capacitance of the air in every zone. It was used in previous
studies (Huchuk, Brien & Cruickshank, 2012), to account for the thermal mass of
room contents. However, the value used was not mentioned in their paper. The
object was also used by German et al. (2014) for calibrating a model in which the
temperatures responded too quickly to outdoor environmental changes. A
“Temperature Capacity Multiplier” value of 15 was found in their study to improve the
rate of change of indoor air temperature. In this study, it was found that a reasonable
rate of air temperature change in the model could be achieved by using a
“Temperature Capacity Multiplier” of 10 for the heavily instrumented houses.
For all ten variant models MBE (%) and CVRMSE (%) for the hourly boiler heat
output were calculated. In addition, the difference between the measured and
predicted volumetrically weighted whole house average air temperatures (ΔTavg
(°C)) was calculated (Table 6-9). For each case, the three indices were compared to
the base case model: where a variant improved the prediction it was indicated by a
tick mark and where it was not improved it was indicated by a cross mark.
163
For the model with SNV, none of the 10 variants improved the predictions of both
energy and indoor air temperature. While this model could closely predict the energy
use in the co-heating test, where all the zones were heated to very similar
temperatures, it failed when the rooms were heated to different temperatures and the
effects of natural convection were significant.
For the model with AFN, three variants improved the predictions of both energy and
indoor air temperatures: 1, 2 and 10.
164
Table 6-9: MBE (%), CVRMSE (%) and ΔTavg (°C) calculated for each case and
each house using AFN and SNV
Model with AFN
ZC house
Model with SNV
CC house
ZC house
CC house
Variant
MBE
(%)
CVR
MSE
(%)
ΔTavg
MBE
(°C)
(%)
CVR
MSE
(%)
ΔTavg
MBE
(°C)
(%)
=model improved
Base
case
1
2
3
4
5
6
7
8
9
10
CVR
MSE
(%)
ΔTavg
MBE
(°C)
(%)
CVR
MSE
(%)
ΔTavg
(°C)
=model not improved
23
45
1.8
11
44
1.3
8
35
2.2
-0.4
39
1.8
20
43
1.6
8
42
1.0
6
35
2.1
-4
40
1.4












16
40
1.6
10
43
1.2
13
38
2.3
2
42
1.6












20
44
1.8
12
45
1.3
8
35
2.2
-0.4
39
1.8












20
44
1.9
12
44
1.2
10
40
2.2
3
47
1.6












28
51
1.5
17
48
1.2
14
39
2.0
6
41
1.6












23
46
1.7
12
44
1.3
9
36
2.2
0.0
40
1.8












23
45
1.8
11
44
1.3
8
35
2.2
0.4
39
1.8












26
48
1.7
15
46
1.2
11
37
2.2
4
40
1.7












34
56
1.44
25
53
1.07
11
36
2.18
2
39
1.71












14
28
1.39
8
33
1.06
3
30
1.87
8
29
1.48












165
Therefore, variants 1, 2 and 10 were combined using AFN to construct the refined
model which improved the predictions of both energy and indoor air temperature
(Table 6-10).
Table 6-10: Comparison of MBE (%) and CVRMSE (%) and ΔTavg (°C) between the
base case model and the refined model
Base case model
Refined model
MBE
CVRMSE
ΔTavg
(%)
(%)
(°C)
ZC
23
45
1.8
CC
11
44
1.3
House
CVRMSE
ΔTavg
(%)
(°C)
3.8
22
0.9
3.9
28
0.5
MBE (%)
These results met the acceptance criteria for hourly calibration of building energy
simulation models according to the ASHRAE Guideline 14: MBE calculated for the
house with ZC and CC house were reduced to 3.8% and 2.9% respectively for the
refined model which both were below the 10% limit outlined by ASHRAE guideline
14; CVRMSE (%) of houses with ZC and CC were also reduced to 22% and 28%
respectively which were both below the 30% acceptance limit. As can be seen from
Figure 6-13 and Figure 6-14, the heating demand predictions were similar to those
measured with a total difference of only 3.9% for both houses. The refined model
predicted a reduction of 14.5% in heat demand for the house with ZC compared to
the house with CC during the HT1, which is in exact agreement with the measured
percentage of savings.
166
100
Measured daily boiler heat output: ZC
Predicted daily boiler heat output: ZC
90
Daily boiler heat output (kWh)
80
70
60
50
40
30
20
10
0
Figure 6-13: predicted daily boiler heat output against measured boiler heat output
for the 28 days of HT: ZC
100
Measured daily boiler heat output: CC
Predicted daily boiler heat output: CC
90
Daily boiler heat output(kWh)
80
70
60
50
40
30
20
10
0
Figure 6-14: predicted daily boiler heat output against measured boiler heat output
for the 28 days of HT: CC
167
There was a reasonable agreement between the measured and predicted indoor air
temperatures of each room (Figure 6-15 and Figure 6-16). The volumetrically
weighted whole house average air temperatures predicted were 0.5 °C and 0.9°C
lower than those measured for the house with CC and ZC, respectively. In addition,
MBE (%) and CVRMSE (%) of the predicted volumetrically weighted whole house
average air temperatures before and after calibration were calculated for both
houses. MBE and CVRMSE for the house with CC were reduced from 7.3% to 3.1%
and 9.8% to 6% respectively. MBE and CVRMSE for the house with ZC were
reduced from 10% to 5.6% and from 11% to 6.9% respectively.
The co-heating test model was re-run in order to test the implications of adding
thermal mass to the energy use of the houses for the version of the model using
AFN. It was found that MBE of House 1 and House 2 were reduced from 9.0 % and
5.4% respectively, to 3.8% and 0.15%, respectively. CVRMSE of the two houses
were also reduced from 10.7% and 8.0% to 7.0% and 5.8% respectively. This gave
further confidence in the revised model.
168
Set-point
Measured Indoor Air Temperature
Simulated Indoor Air Temperature (Air Flow Network)
Outdoor Air Temperature
Living room
Set-back
Occupied hours
'Heating on' hours
Indoor air temperature (ºC)
Indoor air temperature (ºC)
Weekday 1, Thursday 27/02/2014 Weekday 2, Friday 28/02/2014 Weekend, Saturday 01/03/2014
25
20
15
10
5
0
-5
25
Hour
Dining room
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Hallway - ground floor
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Kitchen
20
15
10
5
0
-5
169
Indoor air temperature (ºC)
Indoor air temperature (ºC)
25
Unoccupied bedroom
20
15
10
5
0
-5
25
Bedroom 2
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Bedroom 1
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Bathroom
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Hallway - first floor
20
15
10
5
0
-5
Figure 6-15: Indoor air temperatures measured and predicted by the refined model
for the ZC house along with measured outdoor air temperatures; 27 Feb to 1 March
2014
170
Set-point
Measured Indoor Air Temperature
Simulated Indoor Air Temperature (Air Flow Network)
Outdoor Air Temperature
Indoor air temperature (ºC)
Indoor air temperature (ºC)
Weekday 1, Thursday 27/02/2014
Living room
Occupied hours
'Heating on' hours
Weekday 2, Friday 28/02/2014 Weekend, Saturday 01/03/2014
25
20
15
10
5
0
-5
25
Hour
Dining room
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Hallway - ground floor
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Kitchen
20
15
10
5
0
-5
171
Indoor air temperature (ºC)
Indoor air temperature (ºC)
25
Unoccupied bedroom
20
15
10
5
0
-5
25
Bedroom 2
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Bedroom 1
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Bathroom
20
15
10
5
0
Indoor air temperature (ºC)
-5
25
Hallway - first floor
20
15
10
5
0
-5
Figure 6-16: Indoor air temperatures measured and predicted by the refined model
for the CC house along with measured outdoor air temperatures; 27 Feb to 1 March
2014
172
6.5 Summary
This chapter compared the energy use and indoor air temperatures measured at the
LMP1930 test houses during the co-heating test and HT1 with those predicted by
DTM using two different air flow modelling strategies: Scheduled Natural Ventilation
(SNV) and Air Flow Network (AFN). The base case model could reasonably predict
the energy use of both houses for the co-heating test using both air flow modelling
strategies. However, the predictions were better using SNV compared to the AFN.
The base case model was not able to reasonably predict the energy use and indoor
air temperatures of the test houses for the case of the HT1 using either of the two
airflow modelling strategies. Differences between the measured and predicted
results were investigated and potential parameters which could have contributed to
the differences observed were identified. Sensitivity analysis was then conducted for
these parameters and the parameters which could improve the predictions of energy
use and indoor air temperatures were identified.
Based on the results of the sensitivity analysis, a refined model was calibrated
against the ASHRAE guidelines for hourly calibration of building simulation programs.
The model could be considered calibrated only when using AFN and it did not meet
the calibration criteria when using SNV. The calibrated model could closely predict
the energy savings of ZC measured during the HT1. The model will be used in
chapter 7 to predict the energy savings of ZC which could be achieved in homes in
different UK regions or in better insulated homes.
173
7 Potential savings in other UK locations and
better insulated houses
7.1 Introduction
This chapter discusses the implication of the findings for the annual energy savings
potential of ZC in different UK houses. Firstly, in section 7.2, the empirical results
from the heating trials were evaluated for houses built and occupied in a similar way
to the test houses, but located in different regions of the UK. Then, in section 7.3, the
evaluations in different locations were repeated using the calibrated DTM of the test
houses constructed as described in chapters 5 and 6. Section 7.3 also explores any
difference between the predictions of the empirical model and DTM model and
discusses the potential reasons for the discrepancies observed. In section 7.4, the
calibrated DTM is used and the potential savings of ZC in better insulated homes are
investigated. Finally, section 7.5 provides a summary of the findings in this chapter.
7.2 Evaluation of the empirical results for different UK
locations
7.2.1 Annual heating fuel and cost savings in different UK locations
To extend the measured gas consumptions with CC and ZC to annual values, and to
make an initial estimate of the effect of the weather in different parts of the UK, the
results of the space heating trials were normalised and then evaluated using a
Heating Degree Days (HDD) method.
Firstly, the base temperature (Tbase) to be used for calculating the HDD was
determined using the experimental results and then the relationship between the
weekly HDD and the measured gas consumption was determined. This linear
relationship was then used to estimate the weekly, and so annual, gas consumption
for UK regions with different HDD.
174
7.2.2 Relationship between measured gas use and weather
conditions
The measured weekly gas consumption (WGC) during the trials was strongly
correlated with the weekly average outdoor air temperature (T_wao) for both ZC and
CC (see Figure 7-1, 𝑅𝑅𝑍𝑍𝑍𝑍 2 = 0.72 𝑎𝑎𝑎𝑎𝑎𝑎 𝑅𝑅𝐶𝐶𝐶𝐶 2 = 0.78). The linear relationship for the two
control strategies was similar, but subtlety and importantly different. The regression
lines indicate that for any average weekly ambient temperature below 13.4ºC, ZC will
use less gas than CC. During the heating season, say September to April, the
weekly average ambient is virtually always below 13.4ºC in all regions of the UK. It is
also evident that the energy saved by ZC increases as the weekly average ambient
temperature falls.
The base temperatures of the houses, i.e. the external temperature at which no heat
is needed, is the intercept with the x-axis of best fit line; this was 18.2ºC for ZC and
17.3ºC for CC (Figure 7-1). However, the difference in intercepts is perhaps due to
the limited range of weekly ambient temperatures, to which the two systems were
exposed, leading to poor definition of the x-axis intercepts as reflected in Figure 7-1
by wide 95% confidence intervals for both systems at the x-axis intercept. Thus, the
same base temperature of 17.8ºC, which is the mean value of 17.3ºC and 18.2ºC,
was selected as the base temperature for houses with both ZC and CC. However,
the sensitivity of energy consumption predictions to the HDD base temperature was
investigated using a lower base temperature of 15.5ºC and a higher base of 20ºC
and this will be presented later.
175
1200
1100
1000
Weekly gas consumption (kWh)
900
CC
WGC = -50.0*(T_wao) + 866
R² = 0.78
800
ZC
95% conf int ZC
700
CC
95% conf int CC
600
Linear (ZC)
Linear (CC)
500
400
ZC
WGC = -40.4*(T_wao) + 737.3
R² = 0.72
300
200
100
0
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14
Weekly average outdoor air temperature (ºC)
15
16
17
18
19
20
21
22
Figure 7-1: Weekly gas consumption of the houses with ZC and CC against weekly
average outdoor air temperature for 8 weeks of monitoring, best fit lines and 95%
confidence intervals
The base temperatures for CC and ZC were used to calculate the HDD during the
heating trials (equation (7-1)).
𝑑𝑑𝑑𝑑𝑑𝑑 7
𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊 𝐻𝐻𝐻𝐻𝐻𝐻 = �
𝑑𝑑𝑑𝑑𝑑𝑑 1
(𝑇𝑇𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 − 𝑇𝑇𝑜𝑜𝑜𝑜𝑜𝑜 )𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 ((𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇−𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇)>0)
60 ∗ 24
(7-1)
Where:
𝑇𝑇𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 = the base temperature for CC and ZC houses (i.e. 17.8ºC for this analysis)
𝑇𝑇𝑜𝑜𝑜𝑜𝑜𝑜 = outdoor air temperature (ºC) measured outside the test houses
The subscript shows that only positive differences are summed and if (𝑇𝑇𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 −
𝑇𝑇𝑜𝑜𝑜𝑜𝑜𝑜 )𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 < 0 , then it is set to 0 for that minute in equation (7-1).
176
Weekly HDD were used in preference to daily HDD because different heating
patterns were used for weekdays and weekends. The weekly gas consumption was
then plotted against the weekly HDD for each control configuration. Least squares
regression analysis was used to determine the equation of the performance line.
There was a strong correlation between the 8 measured weekly gas consumption
measurements and the weekly HDD for both ZC and CC (Figure 7-2, 𝑅𝑅𝑍𝑍𝑍𝑍 2 = 0.73
and 𝑅𝑅𝐶𝐶𝐶𝐶 2 = 0.79). If the regression was forced through the origin, the correlation
remained strong and the change in gas consumption per unit change in HDD was
very similar (ZC - 6.03kWh/HDD, 𝑅𝑅𝑍𝑍𝑍𝑍 2 = 0.73; CC - 6.85kWh/HDD , 𝑅𝑅𝑐𝑐𝑐𝑐 2 = 0.79).
700
CC
WGC = 7.2HDD - 27.6
R² = 0.79
600
ZC
WGC = 5.8HDD + 15.7
R² = 0.73
WGC (kWh)
500
400
CC
ZC
300
Linear (CC)
Linear (ZC)
200
100
0
0
10
20
30
40
50
60
70
80
90
100
Weekly HDD
Figure 7-2: Measured weekly gas consumption plotted against calculated weekly
HDD for the houses with ZC and CC
7.2.3 Effect of different UK locations
The performance lines (as in Figure 7-2 and not forced through the origin) were used
to estimate the likely gas consumption for ZC and CC as if houses were built and
occupied in a similar way to those measured, but were located in different regions of
the UK. The HDD were calculated for seven UK regions using the base
177
temperatures of 17.8ºC, for the heating months, October to April. To achieve this,
“typical weather year” data from the International Weather for Energy Calculations
(IWEC) (ASHRAE, 2001) were used for each region: London, the East of England,
the West Midlands, Yorkshire, the Northwest, Northern Ireland and Scotland.
The calculated energy use for heating with each system shows that, regardless of
the location, for the particular house and occupancy tested, ZC saves 11.8-12.5% of
annual gas consumption for heating compared to CC (Table 7-1).
In order to explore the sensitivity of the results to different base temperatures, the
calculations were repeated with a lower base temperature of 15.5ºC, as this is often
used by convention for UK homes (CIBSE, 2006b) and also with 20.0oC, which,
given the set-point temperature of 21oC would seem to be a plausible maximum
value. The relationship between weekly gas consumption and weekly HDD was
determined with these new base temperatures and the energy use recalculated. The
regression coefficients with the new base temperature of 20oC were very similar to
those achieved with a base temperature of 17.8oC. However, for the base
temperature of 15.5oC the regression coefficients were much poorer (𝑅𝑅𝑍𝑍𝑍𝑍 2 = 0.55,
𝑅𝑅𝐶𝐶𝐶𝐶 2 = 0.63). However, it can be seen that the energy savings of ZC is not very
sensitive to the base temperature selected (Table 7-1).
To estimate the impact on annual space heating costs, the Department of Energy
and Climate Change (DECC, 2012b) energy & emissions projections central
scenario for residential gas prices was used (Figure 7-3).
5.8
Residential gas
price
5.7
Pence/kWh
5.6
5.5
5.4
5.3
5.2
5.1
2014
2016
2018
2020
2022
2024
2026
2028
Year
Figure 7-3: projected residential gas prices between 2014 and 2028 (DECC, 2012b)
178
A discounted cash flow analysis was conducted, using a modest discount rate of 5%,
to calculate the Net Present Value (NPV) after 15 years (assumed lifespan of the
system) of upgrading a same size house with conventional heating controls to zonal
heating control in each of the 7 regions. The zonal heating kit is a recently developed
commercial system and therefore the life span of the system is not exactly known,
however, a typical normal TRV has a life span of 15 years and therefore a life span
of 15 years was assumed for the programmable TRVs as well. The cost of batteries
with a life span of two years was included in the total price of the system. The
Internal Rate of Return (IRR), which stays the same regardless of the discount rate,
was also calculated for each region as it is an indication of the discount rate
necessary to pay back the investment within the 15 years. Two ZC systems with
different capital costs were considered for the calculation of NPV: a ‘Luxury type 1’
ZC system with a touch screen central controller (which costs £1200 including
installation costs) and a ‘basic type 2’ ZC system with no central controller in which
PTRVs need to be programmed individually by the household (which costs £120).
The calculations show that, 15 years after upgrading to the Luxury ZC system,
houses in Scotland will have a positive NPV while the houses in all other regions will
have a slightly negative NPV with the houses in more Southern regions having larger
negative NPVs (Table 7-1). This indicates that ZC is a more profitable energy
efficiency measure for the homes in the colder more northerly parts of the UK. The
IRR calculations show that discount rates of up to about 6% is imaginable for the
house in Scotland, whereas the upgrade to luxury ZC would only be financially
worthwhile in London at discount rates of below 3.5% (Table 7-1). In contrast, if
households buy the basic ZC system, which is 10 times cheaper than the luxury
system, they can save about £1000 (present value) after 15 years, regardless of the
location of their house (Table 7-1).
Calculations using the base temperature of 15.5oC and 20oC show that the NPV and
IRR are sensitive to the base temperature selected. This is due to the fact that NPV
and IRR are dependent on the actual kWh of gas saved when using ZC rather than
the percentages of gas savings. It was found that considering a base temperature
lower than 17.8oC, results in lower annual space heating energy use for both
systems, thus lower kWh gas saved by ZC and correspondingly lower NPV and IRR
179
while using a higher base temperature results in exactly opposite results. However,
irrespective of the HDD base, ZC was found to be a more cost effective measure in
Northern regions of the UK based on the empirical approach discussed.
Table 7-1: Estimated gas use for heating the test house, with the same occupancy,
in seven different regions of the UK, using either ZC or CC and, the NPV, IRR or
financial savings, for both a basic and a luxury ZC systems
Annual
Annual
Region
heating
heating
(Weather
energy use
energy use
station)
CC
1
ZC
1
Reduction in
heating
energy use
(%)
NPV after
15 years:
Luxury
system
2
(£)
NPV after
IRR
Luxury
system
15 years:
Basic
3
system
(%)
2
(kWh)
(kWh)
(£)
London
15685
13839
11.8%
-£109
3.4%
£971
(Gatwick)
14884, 15950
13217, 14053
11.2% , 11.9%
-£214, -£79
1.8%, 3.9%
£866, £1001
East of
15696
13848
11.8%
-£108
3.4%
£972
14875,15963
13210, 14064
11.2%, 11.9%
-£216, -£77
1.8%, 3.9%
£864, £1003
Northwest
15805
13936
11.8%
-£95
3.6%
£985
(Aughton)
14973, 16073
13286, 14152
11.3%,11.9%
-£203, -£65
2.0%, 4.1%
£877, £1015
England
(Hemsby)
West
16354
14379
12.0%
-£33
4.5%
£1,047
15460, 16623
13667, 14596
11.6%, 12.2%
-£140, -£2
2.9%, 5.0%
£940, £1078
Ireland
16374
14395
12.1%
-£30
4.6%
£1,050
(Belfast)
15471, 16642
13675, 14611
11.6% , 12.2%
-£139, £0
3.0%, 5.0%
£941, £1080
Yorkshire
16507
14503
12.1%
-£15
4.8%
£1065
(Finningley)
15604, 16774
13780, 14718
11.7%, 12.2%
-£121, £15
3.2%, 5.2%
£959, £1095
Midlands
(Birmingham)
Scotland
17346
15180
12.5%
£80
6.1%
£1,160
(Aberdeen)
16334,17616
14349 ,15397
12.1% ,12.6%
-£27, £111
4.6%, 6.6%
£1053, £1191
Calculated based on HDD base temperature of 17.8ºC in large regular fonts; Calculated based on
15.5ºC and 20.0ºC in small italic font.
1
2
For a typical weather year with heating months being October to April.
Based on Department of Energy and Climate Change (DECC, 2012b) energy & emissions
projections central scenario for residential gas prices and discount rate of 5%
3
Based on the life span of 15 years for TRVs
180
7.3 Evaluation of the DTM results for different UK locations
and comparison with empirical evaluation
A calibrated DTM of the LMP1930 test houses which could reasonably predict the
potential savings from ZC compared to CC during the HT1 was constructed as
discussed in chapters 5 and 6. In this section, this model was used to investigate the
effects of weather in different regions of the UK on potential annual space heating
energy savings of ZC. These results were then compared with the results of the
empirical approach described in section 7.2.
The typical weather year data used in the empirical work were also used with the
DTM for the same seven UK regions: London, the East of England, the West
Midlands, Yorkshire, the Northwest, Northern Ireland and Scotland. The same
heating season was also considered: from 1st of October to end of April. The heating
season averages of air temperature, wind speed and global horizontal radiation in
each region are presented in Table 7-2. Each simulation took more than 1 hour to
complete (computer used: HP ProBook 6460b, 2.5 GHz processor, 4.0 GB RAM)
due to the complexity of the model and AFN calculations.
Table 7-2: Average air temperature, wind speed and global horizontal radiation in
each region during the heating season
Average global
Average air
Average wind
temperature (°C)
speed (m/s)
London
6.74
3.3
East of England
6.73
5.8
65.0
Northwest
6.65
4.5
59.7
West Midlands
6.3
4.0
66.7
Ireland
6.28
5.1
55.8
Yorkshire
6.2
4.5
59.5
Scotland
5.65
5.1
54.0
Region
horizontal solar
radiation (Wh/𝑚𝑚2 )
64.9
Annual gas use of the LMP1930 test houses predicted by the DTM was compared to
annual gas use estimated by the empirical model (Table 7-3). For all the regions, the
181
annual gas use of both houses with CC and ZC predicted by the DTM was more
than the annual gas use estimated by the empirical model. The difference between
the annual gas use of the house with CC predicted by DTM and empirical model
varied from 8.3% in London to 15.0% in the East of England. For the house with ZC,
this difference varied from 6.4% in London to 16.9% in Scotland.
The differences found between the predicted energy use by the empirical model and
DTM could be explained due to their different methodology for estimating the annual
gas use. The HDD method used in the empirical model took into account only the
outdoor air temperature for predicting the annual energy use. Therefore, as it can be
seen in Table 7-2 and Table 7-3, as the average air temperature decreased from
region to region, the estimated annual gas use by the empirical model increased for
both houses. In case when two regions had very similar average air temperatures
(for example London and the East of England) (Table 7-2), the gas use predictions
by the empirical model was also very similar (Table 7-3). However, this was not the
case for the DTM. For example, although London and the East of England had about
the same average air temperature during the heating season (Table 7-2), DTM
predicted 7.3% more energy use for the house with CC and 9.6% for the house with
ZC for the East of England compared to the houses in London. Since their average
global horizontal solar radiations were also very similar (Table 7-2) for these two
locations, this could be explained due to considerably higher average wind speed in
the East of England (5.8 m/s) compared to London (3.3 m/s).
Another example could be observed when comparing the North West and the West
Midlands regions. The air temperature was on average colder in the West Midlands
by 0.35°C (Table 7-2). However, the average wind speed was lower by 11% and the
average global horizontal solar radiation was higher by 10% compared to the
Northwest. Since, the empirical model only considered the outdoor air temperature; it
predicted higher annual gas use for house with CC in the West Midlands compared
to the house with CC located in the Northwest. However, DTM which took into
account the effects of wind speed and solar radiation predicted slightly higher annual
gas use in the house with CC in the slightly warmer but windier and less sunny
region; The North West.
182
As discussed, the effects of wind speed and solar radiation on the estimation of
annual gas use were not considered in the empirical model. Therefore, the DTM
have the advantage to take into account the variations of the solar radiation and the
wind speed from region to region and could potentially provide more robust
estimations.
Although the absolute amount of gas use predicted by the empirical model and DTM
were up to 17% different, the predicted percentage of energy savings from applying
ZC was closely matched between the two approaches for all the regions. While the
empirical approach predicted the energy savings to vary from 11.8% to 12.5%
among different regions of the UK, the DTM model predicts that to vary from 10.7%
to 13.6%. The largest difference between the predicted percentage savings from ZC
by empirical model and DTM was 1.8 pp which was found for the warmest and
coldest regions (i.e. London and Scotland). This is remarkably close prediction
especially when considering the differences between the two methodologies and the
uncertainties involved in both models.
As discussed in section 7.2.3, the empirical model predicted that as we move
towards the more northerly regions of the UK, the percentages of savings slightly
increases. However, the difference between the percentages of savings from ZC in
the warmest and coldest region (i.e. London and Scotland) was below 0.3 pp. DTM
did not show such trend. In contrast, percentages of savings often predicted lower in
more northerly regions of the UK. For example, the percentage savings in London
were 2.9 pp higher than in Scotland.
These differences between the two model predictions prevent any conclusions been
drawn on the effect of UK location on the potential savings. However, the effect of
UK location was found to be small by either the empirical model or the DTM. More
importantly, both models showed that ZC could save more than 10% of annual gas
use in a typical un-furbished 1930s house regardless of the UK location.
183
Table 7-3: Total annual gas use for house with ZC and CC and annual percentages
of savings by ZC in different regions of the UK predicted by DTM and Empirical
Model (EM) and their differences
Region
Annual gas use (KWh) CC
Annual gas use (KWh) ZC
% diff
% diff
Empirical
DTM
(DTMEM)
London
DTM
(DTMEM)
Empirical
DTM
1
15685
17106
8.3%
13839
14781
6.4%
11.8%
13.6%
15696
18468
15%
13848
16351
15.3%
11.8%
11.5%
15805
17994
12.1%
13936
15566
10.5%
11.8%
13.5%
16354
17985
10.0%
14379
15745
9.5%
12.1%
12.5%
Ireland
16374
19227
14.8%
14395
16980
15.2%
12.1%
11.7%
Yorkshire
16507
18468
11.9%
14503
16398
13.1%
12.1%
11.2%
Scotland
17346
19870
14.6%
15180
17741
16.9%
12.5%
10.7%
East of
England
Northwest
West
Midlands
1
Empirical
1
% Savings from ZC
Percentage difference between the predictions of energy use by DTM and Empirical Model (EM)
The cost analysis conducted using the same approach as discussed in section 7.2.3,
but based on the energy savings predicted by DTM suggests that ZC is a cost
effective retrofit measure across all the UK regions particularly when the basic
system is employed (Table 7-4). The highest NPV after 15 years was found in the
Northwest (£235 for the luxury and £1315 for the basic system) and the lowest was
found in the Yorkshire (£24 for the luxury and £1104 for the basic system). In
contrast with the empirical approach, DTM did not show clear relationship that
suggests if the houses in the South or the North could be more financially benefited
from installing the system.
184
Table 7-4: NPV, IRR or financial savings for both a basic and a luxury ZC systems
calculated for seven different regions of the UK based on modelling results for the
un-furbished houses
Region
NPV after 15 years:
(Weather station)
London
(Gatwick)
East of England
(Hemsby)
Northwest
(Aughton)
West Midlands
(Birmingham)
Ireland
(Belfast)
Yorkshire
(Finningley)
Scotland
(Aberdeen)
Luxury system
2
IRR Luxury
system
3
NPV after 15 years:
Basic system
(£)
(%)
(£)
£174
7.4%
£1254
£51
5.73%
£1131
£235
8.3%
£1315
£124
6.7%
£1204
£128
6.8%
£1208
£24
5.3%
£1104
£59
5.8%
£1139
2
7.4 Implications for better insulated homes
To explore how savings might change in a better insulated house, the building
envelope of the LMP1930 house was upgraded in the DTM. The following changes
were made to the model:
•
The air gap between the two layers of the external walls was filled with XPS
Polystyrene (Table 7-5). This reduced the U-value of the external walls from
•
1.666 to 0.392 W/𝑚𝑚2 𝐾𝐾.
300 mm of mineral wool insulation was added to the roof construction, on top
of the first floor ceiling (Table 7-5). This reduced the U-value of the ceiling
(calculated by DesignBuilder) from 3.1 to 0.13 W/𝑚𝑚2 𝐾𝐾.
185
•
All the windows were replaced by double glazed windows with 6 mm clear
glass sheets and 13 mm air between the glass sheets. This reduced the U-
•
value of the windows from 5.9 to 2.67 W/𝑚𝑚2 𝐾𝐾.
Changing the windows would also improve the air tightness. As described in
section 5.3.2, when using AFN, the length of the cracks are fixed (i.e. around
the perimeter of the windows) and could not be changed without changing the
sizes of the windows. However, the air flow coefficient (kg/s. m crack at 1 pa)
could be changed to reflect the lower air leakage from the new double glazed
windows. Therefore, the flow coefficients of the cracks around the windows
which were adopted from DesignBuilder’s “poor” template for the un-furbished
model were changed to those for DesignBuilder’s “good” template. This meant
that the flow coefficients were changed from 0.001 to 0.00006 kg/s. m crack
at1 pa.
Table 7-5: Thermal properties of the insulating materials used in the refurbished
model
Material
Conductivity
(W/m. K)
Density
(kg/𝑚𝑚3 )
Specific heat
capacity (J/kg. K)
XPS Polystyrene
0.034
35
1400
mineral wool; stone wool rolls
0.04
30
840
The revised DTM predicted reduced annual gas use in all the regions as expected.
The annual gas use of the house with CC was reduced by between 42% and 47%
across different regions (Table 7-6). Similarly, the annual gas use of the house with
ZC was reduced by between 42% and 46% (Table 7-6). The percentage of savings
from refurbishment for houses with CC and ZC were very similar for each region. For
both houses, the savings were higher in London and the West Midlands (45 to 47%)
and lower in Scotland (42%).
To test the reliability of model predictions, the results were compared with those from
another modelling tool: the Standard assessment procedure (SAP) (BRE, 2014). The
house with CC was modelled in London and Scotland before and after refurbishment
as SAP does not enable the modelling of ZC. The SAP model predicted 50% and 46%
of savings after refurbishment for the house in London and Scotland respectively.
186
These were slightly higher than the predictions by the DTM; though remarkably close
and in the same direction (i.e. higher savings after refurbishment in London
compared to Scotland). Previous research by Yilmaz et al. (2014) had also shown
that SAP tends to overestimate the percentages of savings which could be achieved
by applying different refurbishment measures compared to EnergyPlus. This result
adds confidence to the findings from the DTM.
The percentage savings of gas use from applying ZC predicted by DTM was found to
be lower in the better insulated house compared to the un-furbished house in all the
regions (Table 7-6). However, the percentage of savings from ZC was reduced more
in the warmer regions (for example 2pp in London) compared to the colder regions
(for example 0.2pp in Scotland) after refurbishment. The percentages of savings
from applying ZC in the better insulated house were found to range from 9.3% in the
East of England to 11.8% in the Northwest. The results from the model showed that
considerable amounts of energy which is used for space heating could be saved
even in refurbished (better insulated) UK houses and in all regions; although to a
less extent compared to un-furbished houses.
Table 7-6: Annual gas use and percentages of savings from refurbishment for ZC
and CC houses for different regions of the UK along with percentage of savings from
ZC after refurbishment and its differences compared to the savings in un-furbished
house
Annual
Region
gas use
(KWh)
CC
London
% Savings
from
refurbishment
Annual
gas use
(KWh)
ZC
pp difference
% Savings
%
in savings
from
Savings
from ZC
refurbishment
from ZC
compared to
un-furbished
9028
47
7981
46
11.6
-2.0
10306
44
9351
43
9.3
-2.2
10026
44
8842
43
11.8
-1.7
9788
46
8672
45
11.4
-1.1
Ireland
10791
44
9756
43
9.6
-2.1
Yorkshire
10300
44
9235
44
10.3
-0.9
Scotland
11551
42
10334
42
10.5
-0.2
East of
England
Northwest
West
Midlands
187
The cost analysis which was conducted similar to the case of un-furbished houses
suggests that a luxury ZC system would not be a cost effective retrofit measure for
homes when refurbished similar to this study (Table 7-7). NPV after 15 years for all
the seven regions were negative; ranging from -£481 in Scotland to -£635 in the East
of England. IRR was also negative for all the regions for the luxury system which
shows that the investment is not profitable. However, a basic ZC system would still
be a cost effective measure across all UK regions even after refurbishment as it was
confirmed by positive NPV across all the regions.
Table 7-7: NPV, IRR or financial savings for both a basic and a luxury ZC system
calculated for seven different regions of the UK based on DTM results for better
insulated houses
Region
(Weather station)
London
(Gatwick)
East of England
(Hemsby)
Northwest
(Aughton)
West Midlands
(Birmingham)
Ireland
(Belfast)
Yorkshire
(Finningley)
Scotland
(Aberdeen)
NPV after 15 years:
Luxury system
2
IRR Luxury
system
3
NPV after 15 years:
Basic system
(£)
(%)
(£)
-£581
-4.5%
£499
-£635
-5.6%
£445
-£500
-3.0%
£580
-£540
-3.7%
£540
-£588
-4.7%
£492
-£570
-4.3%
£510
-£481
-2.6%
£599
2
7.5 Summary
In this chapter, two models were used:
(a) An empirical model which was developed using the HDD method based on
the data measured over the 8-week period of the space heating trials; and
188
(b) A calibrated DTM which was created as described in chapters 5 and 6.
They were used to predict the annual energy savings which could be achieved by
applying ZC instead of CC in houses built and occupied in similar way to the
LMP1930 but located in different UK regions.
The empirical model predicted that:
•
ZC could save 11.8% to12.5% of the annual space heating gas use compared
to the CC regardless of the geographical location.
•
The amount of savings is likely to be more in Northern regions of the UK.
The DTM model predicted that:
•
ZC could save 10.7% to13.6% of the annual space heating gas use compared
to the CC regardless of the geographical location.
•
There is no clear relationship between the potential energy savings of ZC and
the geographical location of the house.
The differences between the predictions of DTM and empirical model were
considered to be due to their different level of details incorporated in the
methodology of the two models.
The DTM was also used to predict the savings for better insulated homes with cavity
and loft (attic) insulation and double glazing instead of single glazing located in
different regions. DTM predicted that savings from ZC would be slightly (between 0.2
to 2.2 percentage points) lower in a better insulated house across all the regions.
It was found that ZC is a profitable energy efficiency measure for both un-furbished
and refurbished UK homes across all the regions when a cheap basic system is
employed.
189
8 Discussion and future work
8.1 Introduction
Saving energy in the residential sector and in particular heating energy is essential to
achieve the UK’s 2050 carbon emissions reduction target. In recent years,
development and deployment of new space heating control strategies, which could
enable households to more efficiently control the delivery of heat, has commanded
the attention by the academics, industry and the government. In the UK, two-zone
space heating control has become mandatory for new homes, and the effect of time
and temperature zone control has been considered in the UK government’s
Standard Assessment Procedure (SAP) for the energy performance assessment of
dwellings. Zonal space heating control (ZC) using programmable TRVs is one of
such emerging systems and allows households with low pressure wet central heating
systems to heat only the occupied spaces of their house instead of all the spaces
and therefore potentially save energy. A range of such products is currently available
on the UK market. The systems are easy to retrofit making them a valuable energy
efficiency measure provided the claimed energy savings can be realised in practice.
Prior to this research, there was no peer reviewed published literature to indicate
how much energy ZC might save in UK homes. Without such information,
households could only rely on the claims of the manufacturers which could be
misleading. A reliable and repeatable method has therefore been developed to
measure the energy saving potential of a ZC system compared to a conventional
control (CC) system. The results from the measurement campaign are discussed in
section 8.2. A Dynamic Thermal Model (DTM) was then used to predict the savings
in the same house. The model was calibrated using the measured data. The findings
from the DTM analysis are discussed in section 8.3. The potential for energy savings
with ZC was then assessed for different UK houses using an empirical model based
on the measured data and the DTM and their predictions were compared. The
results are discussed in section 8.4. Finally, section 8.5 provides a summary of this
chapter.
190
8.2 Measuring the energy savings potential of ZC in a UK
home
To the best knowledge of the author, this is the first study that directly measured the
impacts of ZC on energy use and indoor air temperatures in UK houses. The sideby-side comparison method adopted for the space heating trials is a powerful
technique by which the effects of home energy efficiency measures on building
energy use and thermal comfort can be independently assessed whilst controlling for
the effects of the other influential factors, such as the outdoor weather, occupant
behaviour and heating system characteristics. The method enables relatively small
differences in energy demand caused, for example, by energy efficiency measures,
to be identified. Although this method was used in the late 1970s and 1980s, for
example in a couple of studies by the UK Building Research Establishment (BRE)
(Rayment et al. 1983 and Rayment & Morgan 1984), the method has rarely been
used since. A literature review showed that lack of such comparisons was one of the
main factors which limited the availability of consistent evidence on the energy
savings potential of new space heating controls (Munton et al. 2014). The lack of
recent studies is believed to be because paired full-size test facilities are not widely
available; they can be expensive to construct or buy, the creation of synthetic
occupancy regimens is expensive and time consuming, and the need to match the
buildings can take time and effort. Pairs of old un-furbished homes, as used in the
trials reported here, are very hard to find and secure for research purposes.
Much effort was put into matching the two existing, un-furbished, 1930 houses, at
Loughborough (LMP1930) by using the same heating systems and synthetic
occupancy equipment and profiles, minimizing the effects of different morning and
afternoon solar gains and by switching the space heating control strategies between
the two space heating trials. However, although the houses showed remarkably
close thermal performance during the characterisation tests, they cannot be
considered to be 100% matched due to factors which could not be controlled such as
the wind effects on the East and West facades and small inherent differences in their
constructions.
191
The need to record occupants’ behaviour when measuring the energy saving
potential of heating controls was encouraged by recent studies (Munton et al. 2014).
However, synthetic occupancy can eliminate the variability in the behaviour of people,
which can dominate patterns of domestic energy demand. It also allows measures
that are intrusive or potentially damaging to property or occupants. Examples of such
disruptive measures in this study were using wired thermistors in every room,
installing heat flow meters to measure boiler heat outputs, and insulating the
windows in the East and West facades. However, health and safety concerns may
constrain the behaviours that are simulated. For example, turning on and off gas
ovens and hobs, the automatic opening and closing of doors can pose dangers when
researchers are working in the house and the operation of outside windows and
doors can compromise security.
A number of assumptions were made in undertaking the experiments which place
caveats on the generality of the results. First of all, a single occupancy profile was
considered based on the time use data (ONS, 2002). However, the way occupants
behave in their houses can be very different from this. For example, it was assumed
that the occupants close the doors of the living room, dining room and bedrooms
when they are ‘occupied’. This is perhaps the best scenario for saving energy with
ZC while maintaining comfort as it minimizes the heat transfer from occupied rooms
to other rooms. In reality, the occupants might not wish to change their internal door
opening habits, even if they know it is the best way to get the most benefit from ZC.
The effect of different internal door opening behaviours on the energy savings by ZC
is a useful area for future research.
The trials assumed a household with two working adults and two children, occupying
all the rooms except one, who heat their home intermittently. It was found that ZC is
likely to provide the greatest benefits with intermittent heating rather than continuous
heating. This suggests that, if a house is occupied by a household that spends most
of its time in a heated house, then ZC would save less energy. However, if that
household tended to occupy only one or two rooms, rather than the whole house,
then this could increase the energy savings from ZC. Future work is needed to
consolidates the findings of this study and further investigate the effects of occupants’
space use on the energy savings potential of ZC.
192
In this study, houses could achieve adequate fresh air by infiltration through the
leaky fabric and so window opening was not mimicked. In practice, however, people
may choose to open windows or trickle vents even in winter, for example at night in
occupied bedrooms. The additional heat loss may extend the time needed to
achieve comfort temperatures after the heating has switched on, thus reducing the
benefits of ZC. In addition, it would further reduce the (already low) night time
bedroom air temperatures when the heating is off and so could cause thermal
discomfort.
The already complex, expensive and time consuming instrumentation curtailed the
use of equipment for the detailed assessment of thermal comfort. Thus, indoor air
temperature was taken as a proxy for thermal comfort. However, thermal comfort is
better assessed using operative temperature, which combines air temperature and
mean radiant temperature (MRT) (CIBSE, 2008). Although the difference between
MRT and air temperature is usually small in well insulated homes, it is likely to be
greater in thermally massive buildings which are intermittently heated. Further work
is needed to better understand thermal comfort implications of ZC in different types
of homes.
The forgoing discussion has indicated where there is scope for further useful work in
LMP1930 or similar test houses to explore different occupancy schedules, heating
regimes and thermal comfort measures. There are, however, matters that might
more usefully be explored in other facilities or by other types of study. For example,
this study only examined the potential savings from a house with a heating system
that already complied with the building regulations. If houses have poorly controlled
heating systems, i.e. no TRVs, or even no thermostat (PRT), then applying ZC could
save considerably more energy. Moreover, this study used type 2 ZC systems in
which the boiler operation is controlled using a master room thermostat. The
consequences of the boiler control mechanism used by type 1 ZC systems (in which
each PTRV can call for heat) on the energy savings and boiler efficiency needs
further investigation.
The 11.8% gas savings achieved by ZC compared to CC in the LMP1930 were
based on data collected over an 8-week period and were only reliable for houses of
the same size, type, thermal mass and thermal efficiency and under the same
193
weather conditions. The space heating trials did not measure the annual gas savings,
or savings in refurbished houses or those located in different UK regions. Conducting
longer or larger field trials was not possible in this work. However, even in large field
trials, results are limited only to the homes and households from which the data are
gathered. Therefore, dynamic thermal modelling was employed to explore the
performance of ZC more thoroughly.
8.3 Dynamic thermal modelling and calibration of a UK
home with ZC
This is believed to be the first study in which a DTM has been used to simulate zonal
space heating control with actual measured data being used to calibrate the model.
DTM allow the performance of energy efficiency measures to be investigated.
However, a large number of inputs are required to construct a model, and these are
often very difficult to measure and unavailable even for well characterised buildings.
Input parameters are assumed by the modeller and simplifications are inevitable.
The documentations provided with DTMs provide guidance on the values that can be
used, or assumptions that can be made by modellers. However, the guidance is
often very general, insufficient or unsuitable for a particular building. Hence, the
modeller’s art is to make the “best guess” for the missing parameters in absence of
any rigorous measured evidence. The inaccuracies of the assumed parameters are
a major contributor to the inaccuracies of the DTM’s predictions and the differences
between the predicted and measured performance of buildings, known as the
“performance gap”.
A number of assumptions were made when constructing DTMs to simulate the coheating test and space heating trials which their potential implications on the results
should be carefully considered. For example, the party wall cavity was modelled as a
partition wall. However, there is evidence in literature (Lowe et al., 2007) which
shows significant heat losses from air movement through the party wall cavities.
Since the overall heat loss from each house matched the measured heat loss in the
co-heating test, this would suggest that the model over-predicted the heat loss by
other means (e.g. conduction through external walls or infiltration) to compensate for
the unaccounted heat loss through the party wall cavity. In addition, this would
suggest that the model under-predicted the heat loss through the rooms adjacent to
194
the party wall and as a result over-predicted the heat loss from other rooms to match
the predicted overall heat loss to those measured. However, in this work, it was not
possible to measure the heat loss from individual rooms and future work is needed to
support more evidence.
A futher example of model simplifications is that chimneys and chimney breasts were
not explicitly modelled. Although, passive air vents located at original fire places
were sealed to minimise the air flows through chimney breasts, still air could have
been escaping through small cracks. This could result in higher heat loss via
infiltration in rooms which had a chimney compared to the predicted infiltration heat
loss when chimney breasts were not considered. On the other hand, chimneys are
non-insulated ventilated cavities which could potentially have insulating effect and
therefore reduce the heat loss through the party wall. Reduced fabric heat loss and
increased infiltration heat loss would have cancelling effect considering the overall
heat loss of the room. In addition, the model without chimneys did not consider the
thermal mass of the bricks used in construction of the chimneys which is believed to
be relatively small compared to the rest of the house. To accurately model existing
buildings, reconciliation of the model predictions with the known, measured,
performance, which is known as model calibration, is essential. Such calibration
provides greater confidence in a model’s predictions. Using test house facilities with
synthetic occupancy instead of real occupied homes greatly assisted the process of
model construction and calibration. It eliminated the uncertainties in model inputs
related to occupancy, which determine the operation of doors, windows and window
blinds as well as the time, location and magnitude of the equipment heat gains. In
addition, the characteristics of the heating system components and their operation,
including the heating regimes and nominal set-point and set-back temperatures,
were fully known. In this research it was also possible to undertake whole house
characterisation tests and use these to calibrate the DTM’s representation of the
buildings’ envelope. This is not practical in occupied houses as it needs the houses
to be vacated for a long period.
Modelling the performance of the houses when they were subject to a co-heating
test, in which all the spaces of the house are continuously heated to the same
temperature, was significantly easier than modelling multi-zone, intermittently used,
195
wet heating systems. The base case model created in EnergyPlus showed
reasonable predictions of the energy use in both houses when the co-heating test
was simulated. However, when Heating Trial 1 (HT1) was simulated using the
calibrated building envelope model, the model could not predict the energy use and
indoor air temperatures with reasonable accuracy. The predicted energy saving of
ZC was considerably higher than the measured savings. This clearly shows the risks
involved in trusting predictions of complex models without rigorous calibration of the
model using actual measured data.
Achieving a good model of the intermittently heated multi-zone house controlled by
ZC was difficult. One of the main challenges was to model the air flows in the
building. Both simplified, and a more detailed air flow modelling strategy was tested
and each had its own advantages and disadvantages. A Scheduled Natural
Ventilation (SNV) method which included defining the infiltration rate to each room,
could not model the wind and buoyancy driven air flows but it did allow the use of the
measured whole house air tightness value in the model. The simplified assumption
of SNV that equal amounts of air are exchanged between zones was found to be a
good approximation when all the zones were heated to the same temperature.
However, in this study, SNV was found to be unsuitable for modelling the air flows in
the house with ZC. For this case, inter-zone heat transfer via natural convection was
better represented using an Air Flow Network (AFN). Although AFN provides a more
detailed approach, it requires a large number of model inputs particularly envelope
leakage and wind pressure coefficients which are difficult to measure even in test
facilities. A standard blower door test provides no information regarding the
distribution of air leakage paths, the ventilation rates in individual zones or inter-zone
air flows. These can, in principle, be determined using tracer gas techniques or by
conducting a number of air tightness tests using more than one fan (Liddament,
1996). Multi-tracer gas techniques have also been previously used to determine the
air exchange between zones. However, according to Liddament (1996),
“Measurements using more than three tracers are rare and the practical maximum is
probably restricted to five. This limits the number of zones in which measurements
can be made”. Increasing the number of zones in a ZC house would cause the
instrumentation and computer controlled feedback and injection system required for
196
these methods to become extremely complex and bulky16. Given these difficulties, it
was not possible to calibrate the AFN. In fact, except for a very limited number of
validation projects, reconciliation of measurements with an AFN model has not been
done with “any degree of scientific rigor” (Armstrong, Hadley, Stenner, et al., 2001).
This could place caveats on accuracy of the room-by-room infiltration rate
predictions as well as inter-zone air flows. However, it should be noted that although
the predicted heat loss via infiltration or heat losses from individual rooms could not
be tested, the overall heat loss from the houses was in good agreement with the coheating test. Developing improved methods which could measure the air flows in the
buildings is essential for rigorous calibration of multi-zone dynamic thermal models.
It was also difficult to reliably model the thermal effects of intermittent heating.
Intermittent heating requires prediction of heat up and cool down rates, which are
highly dependent on accurate modelling of a building’s thermal inertia as well as
other parameters such as heating power and internal heat gains. Thermal inertia is a
measure of the responsiveness of materials to variations in temperatures and
includes the mass of the building envelope as well as partitions, furniture, equipment,
etc. inside the building (Pupeikis, Burlingis & Stankevičius, 2010). These parameters
are difficult to measure and accurately account in thermal models.
The radiator model in EnergyPlus is not able to model the dynamic behaviour of
radiators (i.e. the time delay as the water warms or cools when the radiator is
switched on or off) which resulted in much higher rates of heat up and cool down
being predicted by the model than were measured. Because of this modelling error,
it was difficult to accurately predict the air temperatures of the rooms during the
periods of rapid changes in the load (i.e. when the heating was switched on or off).
Others such as Booten & Tabares-Velasco (2012) have made similar observations.
Future work is needed in this area. Accurate prediction of indoor air temperatures
during the heat up periods would be particularly beneficial for the studies looking at
thermal comfort in intermittently heated buildings. Without this ability, DTMs might
not be able to realistically predict occupants’ thermal comfort during the early hours
of occupancy.
16
Liddament (1996) suggest the maximum number of zones that can be injected with gas is approximately ten.
197
Another challenge was to realistically model the performance of the Programmable
Thermostatic Radiator Valves (PTRVs). Currently the operation of the PTRVs cannot
be modelled in DTMs. The nominal set-point and set-back temperatures used in the
base case model failed to realistically represent the variations observed in the room
air temperatures during the heating hours. This was partially because the PTRVs
were unable to maintain the nominal set-point temperatures in most of the rooms.
This was worse when a room had closed doors or when high levels of internal heat
gains were present in the room. It was due to the poor sensation of the air
temperature of the room by the temperature sensors located on the PTRV heads
which would be influenced by the heat from the radiators or other heat sources.
Therefore, in this study, the mean air temperature measured during the occupied
hours was used as the set-point temperature of each room for the purpose of model
calibration. However, some discrepancies inevitably remained between the predicted
and achieved air temperatures during the heating hours.
Modelling the operation of PTRVs is important for accurate predictions of energy
demand and indoor air temperatures. Future research should be focused on
implementing realistic TRV and PTRV operation in DTM tools. In addition,
manufacturers should produce PTRVs which are able to receive temperature
information from an external temperature sensor which is located in a position that
better represents the mean room air temperature. Meanwhile, studies which aim to
measure the energy savings which could be achieved by efficient space heating
control systems could benefit from using a heating system in which the nominal setpoint temperature is accurately achieved and maintained, perhaps using electrical
heating. Electrical heating would also allow the heat input to each zone to be
accurately and easily measured which would be beneficial for validation purposes.
Limitations and underlying assumptions in DTM tools also caused difficulties for
reconciliation of the measured and predicted energy use and indoor air temperatures.
For example, the AFN poorly represented the natural convective heat transfer via air
flow through horizontal openings such as staircases. This was important as it did not
allow accurate predictions of the air temperatures in the ground floor and first floor
hallways. The ground floor hallway is where the master thermostat which controls
the boiler operation is often located. Accurate prediction of the air temperature is
198
essential for modelling heating systems with a master thermostat (such as type 2 ZC
systems) as used in this study. Without accurate prediction of the air temperature of
the zone with the master thermostat, EnergyPlus’s Energy Management systems
(EMS) could not improve the accuracy of the model predictions. Future work is
needed to develop DTM tools which can accurately model the air flows through
horizontal openings and allow heating systems with a master thermostat to be
accurately modelled.
The core assumption of the heat balance equation in the multi-zone thermal models
such as EnergyPlus is that zone air is well mixed with a uniform temperature
distribution. This simplified assumption cannot reflect reality well because the room
air temperature will vary throughout the room due to the various heat gain and
stratification effects. In this study, the measured air temperature in the volumetric
centre of each room was assumed to reasonably represent the room mean air
temperature. However, using more than one temperature sensor in each room could
have given more confidence in this assumption. Therefore, the comparison which
was made between the measured and predicted room air temperatures should be
only considered approximate. In recent years a number of advanced numerical
models such as zonal models (Megri & Haghighat, 2007) have been developed and
in very limited cases they were integrated into multi-zone DTMs in order to increase
the accuracy of air temperature predictions within a zone. In addition, in a limited
number of studies, computational fluid dynamics which is a more complex and
computationally intensive method for simulating fluid flow, has been employed and
integrated with DTM tools for this purpose (Negriio 1998, Beausoleil-Morrison 2000,
Bartak et al. 2002 and Tan & Glicksman 2005). However, more work is needed in
this area.
Differences in the weather file used in the model and the actual weather conditions
during the tests also contributed to the discrepancies between the model predictions
and the measured data. Except for the outdoor air temperatures, none of the input
parameters used in the weather file was measured on site. Data was collected from
three different weather stations which were between 2 to 26 km away from the
houses. In particular, on site measurements of solar radiation could have been
beneficial as discrepancies were observed between predicted and measured indoor
199
air temperatures during the hours of high solar radiation. Global horizontal radiation
can be measured on site using a pyranometer (Kotti, Argiriou, & Kazantzidis, 2014).
However, it is more difficult to measure direct horizontal radiation which is measured
using a pyranometer positioned horizontally on support equipped with an adjustable
device such as a shadowband or shade disk that blocks the direct component from
the sensor (Kotti et al., 2014). Future calibration studies should be designed to
collect as many of the weather parameters as possible on site with particular
attention to solar radiation data.
Of course, the precision of all the measurements made in the houses depends on
the accuracy of the monitoring devices (as indicated in chapter 4) and the
measurement methods adopted. This would also contribute to a part of the
discrepancies between predictions and measurements.
Plotting room-by-room hourly air temperatures and inspecting the discrepancies
between the measured and predicted values proved to be a useful method for
identifying potential reasons for discrepancies between the measured and predicted
performance. Combining this method with sensitivity analysis, which is a well
established technique, would form a powerful procedure to assist with the calibration
of multi-zone dynamic thermal models.
Despite the difficulties in calibration, a DTM of test houses could reasonably predict
the energy use and indoor air temperatures during the first heating trial. The model
predicted a very similar ZC gas savings to that actually measured. The model was
validated according to the criteria recommended in ASHRAE guideline 14 (ASHRAE,
2002) for the hourly calibration of building simulation models. The model was then
used to predict the savings in different regions of the UK and for a better insulated
home.
8.4 Predicting the energy savings potential of ZC in
different UK houses
Two different models were employed to predict the annual gas savings of ZC
compared to CC in houses in different UK regions. Each model has its own
advantages and disadvantages. The empirical model which was based on a Heating
200
Degree Day (HDD) analysis had substantial benefits over other simplified methods
that use mean outdoor temperatures to calculate energy demand such as BSEN ISO
13790 (BSI, 2004) since the “HDD method accounts for fluctuations in outdoor
temperature and can capture extreme conditions in a way that mean temperature
methods cannot” (CIBSE, 2006b). The model developed was based on relationships
found between the weekly gas use and the average outdoor air temperatures of
each house during both space heating trials. Therefore, the empirical model’s
predictions did not directly take into account other influential factors such as solar
radiation and wind. DTM is significantly more detailed and allows the effects of wind
and solar radiation to be accounted in model predictions. However, DTM has its own
limitations as discussed in section 8.3.
Employing two models for evaluation of the results was a powerful technique which
allowed inter-model comparison in order to find confidence in the model predictions.
The empirical model predicted that the energy savings by ZC would be greater in
colder regions. It predicted that the annual gas savings of ZC varies from 11.8% in
the warmest UK region (i.e. London) to 12.5% in the coldest region (i.e. Scotland). In
contrast, the DTM did not show a trend for higher gas savings in colder regions. In
fact, it showed lower savings in Scotland (10.7%) compared to London (13.6%).
Since the results from the two models were not in agreement, this study was not able
to conclude whether ZC would be more suitable for colder climates or warmer
climates. Both models were based on data collected during a short winter period
which did not include many warm days. This increased the uncertainty in the models
used to extrapolate the measurements to warmer periods of the year and to other
locations. Further trials, in milder weather conditions are needed to further
investigate the effect of weather on the potential savings of ZC.
The evaluation using the DTM showed that the energy savings which could be
achieved by ZC in a better insulated home would be slightly lower than for poorly
insulated homes. It was estimated that ZC could save between 9.3 to 11.8% of
annual gas use in a better insulated home across the UK regions. Findings were in
line with previous forecasts by Utley & Shorrock (2008) that argued savings from
heating certain spaces instead of the whole house could be higher for a house with
poor levels of insulation while it would be lower for a well-insulated house where heat
201
transfer from the heated spaces can often achieve the comfort temperatures
throughout the house. There was a tendency for more reduction in the potential
savings of ZC after refurbishment of the houses in warmer regions compared to
houses in colder regions. For example, the annual gas saving of a house with ZC in
London was estimated to reduce from 13.6% to 11.6% after refurbishment while it
was only reduced from 10.7% to 10.5% for the house in Scotland. However, the
reduction after refurbishment was small across all the regions (between 0.2 to 2.2
pp). More work using a DTM is needed to investigate the effects of different
interventions on the potential energy savings of ZC.
Despite the fact that the empirical model and DTM were not in agreement regarding
the effect of different UK region on the energy savings, both models predicted that
ZC is able to save between 10-14% of annual gas use regardless of the UK location
for the particular house and occupancy tested. In addition, the percentage of savings
would not drop below 9% in any region even after the house was refurbished. This
clearly shows that retrofitting of ZC to existing houses in the UK offers an opportunity
for reducing energy demand for space heating. It is also much easier, cheaper,
faster and non-disruptive for the households (but less energy efficient) than other
retrofit measures such as external wall insulation, double glazing etc. The cost
analysis also shows that upgrading to ZC could be a good investment for homes in
the UK, especially when purchasing the cheaper basic system. However, the
cheaper system does not have a user friendly interface with a touch screen central
controller. This might influence how much households actually get involved with the
control of their heating system and could shrink the potential cost savings of
installing such systems. Large field trials are essential to investigate the occupants’
interaction with ZC systems.
8.5 Summary
In this chapter, the results from the experimental, dynamic thermal modelling and
evaluation campaigns have been discussed. The advantages of using test house
facilities with synthetic occupancy rather than real occupied homes have been
presented. On the other hand, this approach limits the generality of the results to
other houses in other locations with different fabric energy efficiency. Areas for future
202
work in similar test houses to further develop our understanding of the potential
energy savings of ZC have been outlined.
The results of comparing predicted and measured energy use and indoor air
temperatures during the heating trials have been discussed and the importance of
model calibration prior to wider scale evaluation was argued. The difficulty of
creating reliable multi-zone DTMs of houses with ZC have been presented and some
limitations of current dynamic thermal modelling tools that could be be addressed in
future work have been noted.
Finally, the strengths and weaknesses of the empirical and predictive evaluation
techniques used in this study have been discussed. The results predicted by both
techniques, for houses in different UK locations have been compared and the
reasons for any discrepancies explored.
203
9 Conclusions
9.1 Introduction
In this thesis, the potential energy savings from using zonal space heating controls
instead of conventional space heating controls in a UK home have been investigated
and quantified. This was achieved by completing the three objectives. Firstly, a pair
of test houses were instrumented and shown to be well matched in thermal
performance using a side-by-side co-heating test. The houses were then automated
to replicate the impacts of an occupant family (two adults and two school aged
children). Over a winter period, the energy use and indoor air temperatures of the
two houses were measured when the space heating had Conventional Control (CC)
in one house and ZC in the other house. The control strategies were swapped half
way through the test in order to avoid any differences between the thermal
performances of the two houses. Then, a dynamic thermal model (DTM) of the same
houses with the same occupancy pattern was constructed and calibrated against the
measured data. Finally, the results from the experimental work and the DTM were
evaluated and the potential energy savings of ZC in different UK climates or in better
insulated homes was investigated. This chapter summarises and concludes the main
findings from each of the three components of this study and provides
recommendations based on this research.
9.2 Measuring the energy savings potential of ZC in a UK
home
Zonal control heating was compared with conventional control in a matched pair of
1930s -era UK semi-detached houses with synthetic occupancy over an 8-week
winter test period (16 February to 21 April 2015; including 9 days in which the test
was stopped due to equipment failure and swapping the control strategies). It was
found that:
•
Daily boiler heat output of the house with ZC was lower than in the house with
CC on every single day of the tests. On average, over the test period, ZC,
compared to CC, provided a 14.1% reduction in measured boiler heat output.
204
•
ZC reduced the average daily boiler efficiency by 2.4 percentage points.
•
The resultant effect was that ZC produced an 11.8% saving in gas
consumption over the 8-week monitoring period, compared with CC.
•
The average air temperature in all of the rooms, and on average for the whole
house, was lower with ZC than with CC during: the whole day, the period
when the heating system was on, and the period when the heating was off.
There was little or no reduction in the average air temperature in rooms while
they were occupied and the occupants were awake, although during sleeping
hours bedroom temperatures were up to 1.8ºC cooler on average with ZC.
The average air temperatures of bedrooms in both houses during the sleeping
period were below the air temperatures recommended by CIBSE.
•
The average gas saving of ZC was found to be higher during the intermittently
heated weekdays rather than the weekends when the houses were heated for
longer periods.
•
The PTRVs did not maintain their nominal set-point temperatures in most of
the rooms as the average air temperature measured during the occupied
period when the heating was on was different than the nominal set-point
temperatures.
9.3 Dynamic thermal modelling and calibration of a UK
home with ZC
A DTM of the test houses was constructed and the co-heating test and heating trial
were simulated using two different air flow modelling strategies: Scheduled Natural
Ventilation (SNV); and an Air Flow Network (AFN). Comparing the predicted energy
use and indoor air temperatures with those measured during the tests revealed that:
•
Both air flow modelling strategies were able to reasonably predict the energy
use of the test houses under the co-heating test. However, for this case
study, the simple SNV strategy provided energy use predictions which were
closer to the measured energy use compared to when AFN was used.
However, this does not provide definitive evidence on which of the two air
205
flow modelling strategies would be more accurate or appropriate considering
the assumptions and limitations incorporated in each approach.
•
For the case of the heating trial, the energy use and indoor air temperatures
predicted by the DTM prior to calibration were poor using either of the two air
flow modelling strategies.
•
Achieving a well calibrated DTM of an intermittently heated multi-zone house
with a wet central heating system controlled by ZC was very difficult. This
was due to: difficulties in accurately modelling air flows in the houses;
limitations of the current dynamic thermal modelling tools such as difficulties
in modelling the PTRV operation; underlying assumptions within the DTM
regarding the fully mixed air temperature in a zone; inaccuracies of the
measurements; and the availability of important model inputs.
•
Hourly comparison of the measured and predicted indoor air temperatures
and sensitivity analysis were found to be useful techniques for the calibration
of the multi zone DTMs.
9.4 Predicting the energy savings potential of ZC in
different UK houses
The potential savings from ZC for houses in different UK regions were calculated
using an empirical heating degree day (HDD) method and also using the calibrated
DTM. The empirical model suggested that:
•
Regardless of geographic location, ZC, in houses built and occupied in a
similar way to the test houses, could save about 11.8% to12.5% of the annual
space heating energy, compared to CC.
•
ZC is potentially a more cost-effective measure in Northern regions of the UK,
compared with Southern regions. However, the financial costs and benefits of
upgrading from CC to ZC are subject to many uncertainties.
The calibrated DTM suggested that:
206
•
Regardless of geographic location, ZC, in houses built and occupied in a
similar way to the test houses, could save about 10.7% to13.6% of the annual
space heating energy, compared to CC.
•
There is no clear relationship between the potential energy savings of ZC and
the geographical location of the house.
•
The DTM was also used to predict the savings for the houses after installing
double glazed windows and insulating the cavity wall and the loft (attic) space.
The DTM predicted that savings from ZC would be between 0.2 to 2.2
percentage points lower after refurbishment across all the regions. This was in
agreement with the forecasts of previous studies.
The differences between the predictions of DTM and empirical model were believed
to be because:
•
The simplified HDD method employed in the empirical model only took into
account the outdoor air temperature as the factor which determined the gas
use while the more detailed DTM considered other influential parameters such
as solar radiation and wind speed.
•
Development of the empirical model and validation of the DTM model were
based on data collected during a short winter period which did not include
many warm days. This increased the uncertainty when extrapolating to
warmer periods of the year and to other locations.
9.5 Overall conclusions and recommendations for future
work
Annual gas savings of ZC compared to a house heated conventionally is in the range
of 10-14% for a typical un-insulated 1930s UK family home. ZC is likely to save more
energy in un-insulated and intermittently heated homes compared to refurbished,
continuously heated homes. ZC could be considered as a cost effective energy
efficiency measure for UK homes in all regions particularly when cheaper ZC
systems are employed. Further studies in the Loughborough matched pair homes
are suggested to enable the effects of different occupancy and heating schedules on
207
energy savings to be investigated. Further work, using a dynamic thermal model
calibrated against data which is measured for long period including warmer periods,
will enable the energy saving potential of zonal control to be explored more fully.
208
References
Agarwal, Y., Balaji, B., Dutta, S., Gupta, R.K., et al. (2011) Duty-cycling buildings
aggressively: The next frontier in HVAC control. Proceedings of the 10th ACM/IEEE
International Conference on Information Processing in Sensor Networks. p.246–257.
Allard, F. & Utsumi, Y. (1992) Airflow through large openings. Energy and Buildings.
Vol.18, p.133–145.
Allen, E.A. & Pinney, A.A. (1990) Standard dwellings for modelling details of
dimensions, construction and occupancy schedules.UK, Garston, Building Research
Establishment, Building Environmental Performance Analysis Club (BEPAC),
Technical note 90/2.
Allinson, D. (2007) Evaluation of aerial thermography to discriminate loft insulation in
residential housing. PhD thesis, The University of Nottingham. Available from:
http://etheses.nottingham.ac.uk/284/.
Andrews, D., Lowry, G. & Whitehead, B. (2012) Assessing the life cycle impacts of
thermostatic and programmable thermostatic radiator valves in a UK household.
Proceesings of the 2nd LCA Conference, 6-7 November 2012, Lille, France.
Armstrong, P.R., Hadley, D.L., Stenner, R.D. & Janus, M.C. (2001) Whole-building
airflow network characterization by a many-pressure-states (MPS) technique.
ASHRAE Transactions. Vol.107, Part 2, p.645–657.
ASHRAE (2001) International Weather for Energy Calculations (IWEC Weather Files)
User’s Manual and CD-ROM. American Society of Heating, Refrigerating and AirConditioning Engineers, Atlanta, GA.
ASHRAE (2002) ASHRAE guideline 14-2002: measurement of energy and demand
savings. Atlanta, GA, American Society of Heating, Refrigerating and Airconditioning Engineers.
ASHRAE (2009) 2009 ASHRAE handbook: fundamentals. SI. Atlanta, GA, American
Society of Heating, Refrigerating and Air-conditioning Engineers.
209
ATTMA (2010) Technical Standard L1: measuring air permeability of building
envelopes (dwellings).The Air Tightness Testing & Measurement Association, The
British Institute of Non-Destructive Testing. Available online at: <
http://www.attma.org/wp-content/uploads/2013/10/ATTMA-TSL1-Issue-1.pdf>
[Accessed on 28th May 2014]
BADC (2014) WMO Meteorological codes. British Atmospheric Data Centre.
Available online at:< https://badc.nerc.ac.uk/data/surface/code.html#oktas>
Badiei, A., Firth, S. & Fouchal, F. (2014) The role of programmable TRVs for space
heating energy demand reduction in UK homes. Proceedings of Building Simulation
and Optimization Conference. 23-24 June 2014, UCL, London, UK.
Barakat, S.A. (1987) Inter-zone convective heat transfer in buildings : A Review.
Transactions of the ASME. Vol. 109, p.71–78.
Bartak, M., Beausoleil-Morrison, I., Clarke, J. a., Denev, J., et al. (2002) Integrating
CFD and building simulation. Building and Environment. Vol.37, p.865–871.
Beattie, D.H. (1966) Domestic hot water supplies and central heating. Corsby
Lockwood &Son Ltd, London, UK.
Beausoleil-Morrison, I. (2000) The adaptive coupling of heat and air flow modelling
within dynamic whole-building simulation. PhD thesis, University of Strathclyde,
Glasgow, UK.
Boardman, B., Darby, S., Killip, G., Hinnells, M., Jardine, C.N., Palmer, J. & Sinden,
G. (2005) 40% House project. Environmental Change Institute. [Online]. Available
from: <http://www.eci.ox.ac.uk/research/energy/downloads/40house/40house.pdf>
[Accessed: 2nd September 2015]
Booten, C. & Tabares-Velasco, P. (2012) Using EnergyPlus to Simulate the Dynamic
Response of a Residential Building to Advanced Cooling Strategies. 2nd
International Conference on Building Energy and Environment . August 1-4 2012,
Colorado, US. Available from: http://www.nrel.gov/docs/fy12osti/55583.pdf.
210
BRE (2013) Energy Follow-Up Survey 2011. Report 4: Main heating systems.
Department of Energy & Climate Change. Available online at: <
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/27477
2/4_Main_heating_systems.pdf> [Accessed on: 2nd September 2014]
BRE (2014) SAP2012 - The Government’ s Standard Assessment Proceedure for
Energy Rating of Dwellings. BRE
BRECSU (2000) Good Practice Guide 284: Domestic central heating and hot water:
systems with gas and oil-fired boilers. Building Research Energy Conservation
Support Unit.
BRECSU (2001) Good Practice Guide 302: Controls for domestic central heating
and hot water – guidance for specifiers and installers. Building Research Energy
Conservation Support Unit.
Brown (2011) A BSRIA guide: Underfloor heating and cooling. BSRIA.
BSI (1999) BS 7478:1991- Guide to Selection and use of thermostatic radiator
valves. The British Standards Institution.
BSI (2004) BS EN ISO 13790. Thermal performance of buildings: Calculation of
building energy demand for heating. British Standards Institution, London, UK.
BSI (2006) BS EN 215:2004- Thermostatic radiator valves - Requirements and test
methods. The British Standards Institution.
BSI (2013) BS EN 12831: Heating systems in buildings — Method for calculation of
the design heat load. The British Standards Institution.
BSMW Products Ltd (2011) A guide to radiator sizing and heat output. BSMW
Products Limited. Available online at: <
http://www.pipecasings.co.uk/downloads/download%20files/Radiator%20Output%20
Calculator.pdf> [Accessed on: 3rd September 2015]
Burch, D.M. (1980) Infrared audits of roof heat loss. ASHRAE Transactions, Denver,
Colorado. ASHRAE Vol.86, Part 2, p.209-225.
211
Buswell, R., Marini, D., Webb, L. & Thomson, M. (2013) Determining heat use in
residential buildings using high resolution gas and domestic hot water monitoring.
Proceedings of 13th conference of international building performance simulation
association, August 26-28, Chambery, France.
Butler, D & Dengel, A. (2013) Review of co-heating test methodology. NHBC
Foundation. Available online at: <
http://www.nhbcfoundation.org/Researchpublications/Reviewofcoheatingtestmethodo
logies/ttabi/591/language/en-US/Default.aspx> [Accessed on: 28 May 2014]
Calleja Rodríguez, G., Carrillo Andrés, A., Domínguez Muñoz, F., Cejudo López,
J.M., et al. (2013) Uncertainties and sensitivity analysis in building energy simulation
using macroparameters. Energy and Buildings. Vol.67, p.79–87. Available from:
doi:10.1016/j.enbuild.2013.08.009.
Capon, R. & Hacker, J. (2009) Modelling climate change adaptation measures to
reduce overheating risk in existing dwellings. Proceedings of the Eleventh
International IBPSA Conference Glasgow. 2009 pp. 1276–1283.
Celotex (2015) The Celotex Handy Guide. Celotex Saint-Gobain, Ipswich, UK.
Available online at:< http://www.celotex.co.uk/assets/handy-guide-2015.pdf>
[Accessed on 2nd September 2015]
CIBSE (2000) Testing buildings for air leakage: TM23. The Chartered Institution of
Building Services Engineers London
CIBSE (2005) Heating, ventilation, air conditioning and refrigeration: CIBSE guide B.
The Chartered Institution of Building Services Engineers London
CIBSE (2006a) Degree days: Theory and application, TM41. Chartered Institution of
Building Services Engineers, London, UK.
CIBSE (2006b) Environmental design: CIBSE guide A. The Chartered Institution of
Building Services Engineers London
212
CIBSE (2008) CIBSE concise handbook. The Chartered Institution of Building
Services Engineers London
CIBSE (2009) Building control systems: CIBSE guide H. The Chartered Institution of
Building Services Engineers London
Clarke, J.A. (2011) Energy simulation in building design. 2nd Edition. Routledge, USA.
Coakley, D., Raftery, P. & Keane, M. (2014) A review of methods to match building
energy simulation models to measured data. Renewable and Sustainable Energy
Reviews. Vol.37, p.123–141.
Collins, K. (1986) Low indoor temperatures and morbidity in the elderly. Age and
Ageing. Vol. 15(4), p.212-20.
Committee on Climate Change (2008) Building a low-carbon economy – the UK ’ s
contribution to tackling climate change. TSO (The Stationary Office). Available from:
<https://www.theccc.org.uk/publication/building-a-low-carbon-economy-the-ukscontribution-to-tackling-climate-change-2/> [Accessed: 2nd September 2015]
Constable, G. Shaw, D. (2011) Evaluation of the Enistic and Plogg solutions for
socket level monitoring of electrical power consumption. Welsh Video Network.
Available online at:
<http://www.wvn.ac.uk/en/media/Plogg%20Enistic%20Evaluation.pdf> [Accessed on:
28 May 2014]
Consumer focus (2012) Consumers and domestic heating controls : a literature
review. Consumer focus. Available from:
<http://www.consumerfocus.org.uk/files/2012/01/Consumers-and-domestic-heatingcontrols-a-literature-review.pdf> [Accessed: 2nd September 2015
Cóstola, D., Blocken, B. & Hensen, J.L.M. (2009) Overview of pressure coefficient
data in building energy simulation and airflow network programs. Building and
Environment. Vol.44 (10), p.2027–2036. Available from:
doi:10.1016/j.buildenv.2009.02.006 [Accessed: 16 October 2014].
213
Cóstola, D., Blocken, B., Ohba, M. & Hensen, J.L.M. (2010) Uncertainty in airflow
rate calculations due to the use of surface-averaged pressure coefficients. Energy
and Buildings. Vol.42 (6), p.881–888. Available from:
doi:10.1016/j.enbuild.2009.12.010 [Accessed: 14 October 2014].
Crawley, D.B., Hand, J.W., Kummert, M. & Griffith, B.T. (2008) Contrasting the
capabilities of building energy performance simulation programs. Building and
Environment. Vol.43 (4), p.661–673. Available from:
doi:10.1016/j.buildenv.2006.10.027 [Accessed: 9 July 2014].
Crawley, D.B., Lawrie, L.K., Winkelmann, F.C., Buhl, W.F., et al. (2001) EnergyPlus:
creating a new-generation building energy simulation program. Energy and Buildings.
Vol.33 (4), p.319–331. Available from: doi:10.1016/S0378-7788(00)00114-6.
Danfoss (no date) Domestic heating controls are easy. Part 4- Motorized valves.
Available online at: http://danfossrandall.co.uk/PCMPDF/Introducing%20Motorised%20Valves%20Part%204.pdf
[Accessed 1st September 2015]
DECC (2011a) Energy consumption in the United Kingdom : 2011-Domestic energy
consumption in the UK since 1970. Department of Energy and Climate Change.
Available online at: <
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/44913
4/ECUK_Chapter_3_-_Domestic_factsheet.pdf> Note: Version 2011 is no longer
available, this link will now allow download of version 2015 (2nd of September 2015).
DECC (2011b) The Green Deal and Energy Company Obligation: consultation
document. Department of Energy and Climate Change
DECC (2011c) Statistical Release: UK climate change sustainable development
indicator: 2010 greenhouse gas emissions, provisional figures and 2009 greenhouse
gas emissions, final figures by fuel type and end-user. Department of Energy and
Climate Change. Available from:
<http://www.decc.gov.uk/assets/decc/statistics/climate_change/1515-statrelease-
214
ghg-emissions-31032011.pdf [Accessed on 30 March 2012]. Note: The link is no
longer available.
DECC (2012a) Statistical Release: 2011 UK Greenhouse gas emissions, provisional
figures and 2010 UK greenhouse gas emissions, final figures by fuel type and end
user. Department of Energy and Climate Change. Available from:
<https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/4089
38/Provisional_UK_greenhouse_gas_emissions_national_statistics_2011.pdf >
[Accessed: 2nd September 2015]
DECC (2012b) Updated energy & emissions projections 2012. Department of Energy
and Climate Change.
DECC (2014a) Statistical Release: 2013 UK Greenhouse gas emissions, provisional
figures and 2012 UK greenhouse gas emissions, final figures by fuel type and end
user. Department of Energy and Climate Change. Available from:
<https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/2959
68/20140327_2013_UK_Greenhouse_Gas_Emissions_Provisional_Figures.pdf>
DECC (2014b) Estimated average calorific values of fuels (DUKES A.1-A.3).
Department of Energy and Climate Change. Available online at: <
https://www.gov.uk/government/publications/dukes-calorific-values> [Accessed on:
28 May 2014]
DECC (2014c) Warm Front Scheme. Department of Energy and Climate Change.
Availbale online at: < https://www.gov.uk/warm-front-scheme>
Dentz, J. & Ansanelli, E. (2015) Thermostatic Radiator Valve Evaluation. U.S.
Department of Energy. Available online at: <
http://apps1.eere.energy.gov/buildings/publications/pdfs/building_america/thermostat
ic_radiator_valve.pdf> [Accessed: 2nd September 2015]
Department for Communities and Local Government (2001) English House Condition
Survey: Supporting Tables. HM Government archive. Available online at: <
http://discover.ukdataservice.ac.uk/catalogue?sn=6102> [Accessed on: 28 May 2014]
215
Department for Communities and Local Government (2006) Code for Sustainable
Homes - A step-change in sustainable home building practice. London, Department
for Communities and Local Government. Available from:
http://www.planningportal.gov.uk/uploads/code_for_sust_homes.pdf. [Accessed: 2nd
September 2015]
Department for Communities and Local Government (2011) Domestic Building
Services Compliance Guide. London, Department for Communities and Local
Government. Available online at: <
http://www.planningportal.gov.uk/uploads/br/domestic_building_compliance_guide_2
010.pdf> [Accessed: 2nd September 2015]
Department for Communities and Local Government (2012) English housing survey
2010: homes statistics. HM Government. Available online at:<
https://www.gov.uk/government/publications/english-housing-survey-homesstatistics-2010> [Accessed on: 28 May 2014]
DesignBuilder (2014) DesignBuilder user manual Version 3.4. UK: DesignBuilder
software limited, available online at:<
http://www.DesignBuilder.co.uk/helpv3.4/index.htm>
Dietz, R.N., Goodrich, R.W., Cote, E. a. & Wieser, R.F. (1986) Detailed Description
and Performance of a Passive Perfluorocarbon Tracer System for Building
Ventilation and Air Exchange Measurements. Measured Air Leakage of Buildings,
ASTM STP 904. p.203–264.
Doherty, C.H. (1967) Practices in the domestic field. Proceedings of heating and
ventilation for a human environment conference, Vol.182, Part 3E. The Institution of
Mechanical Engineers.
Dols, S.W. & Walton, G.N. (2002) CONTAMW 2.0 User Manual. U.S. Department of
Commerce.
Edwards, R., Hartless R. & Gaze, A. (1990) Measurements of subfloor ventilation
rates- comparison with BREVENT predictions. Proceedings of the 11th AIVC
Conference: The role of ventilation, 18-21 September 1990, Belgirate, Italy.
216
Energy Saving Trust (2008a) Central heating system specifications ( CHeSS ) Year
2008. London, Energy Saving Trust.
Energy Saving Trust (2008b) Domestic heating by gas : boiler systems – guidance
for installers and specifiers. London, Energy Saving Trust.
Enica Ltd (2014) OPTI-PULSE LED flash or pulse data logger with repeated pulse
output. Available online at:
<http://www.enica.co.uk/documents/optipulsebrochure.pdf> [Accessed on: 28 May
2014]
Environmental Change Institute (2009) UKDCM2 model. Environmental Change
Institute. University of Oxford, UK. Available online at:
<http://www.eci.ox.ac.uk/research/energy/bmtmodelling.php>
Enviros Consulting Ltd (2008) The potential for behavioural and demand-side
management measures to save electricity, gas and carbon in the domestic sector,
and resulting supply side implications. London, Enviros Consulting Ltd. Available
online at:
http://www.niam.scarp.se/download/18.6579ab6011d9b20740f8000648631/1350483
507484/decc2008+potential+behavioural+demand+side+management+-saveenergy-implications.pdf. [Accessed: 2nd September 2015]
Erickson, V.L., Achleitner, S. & Cerpa, A.E. (2013) POEM: Power-efficient
Occupancy Energy Management System. Proceedings of the 12th international
conference on Information processing in sensor networks - IPSN ’13. Available from:
doi:10.1145/2461381.2461407.
Eurotronic (2011) Sparmatic Comet programmable energy-saving radiator
thermostat: user manual and guarantee. Eurotronic Technology GmbH. Available
online at: <
http://www.eurotronic.org/fileadmin/user_upload/eurotronic.org/Download-BALWB/comet-BAL-en.pdf> [Accessed on: 4th September 2015]
Firth, S.K., Lomas, K.J. & Wright, a. J. (2010) Targeting household energy-efficiency
measures using sensitivity analysis. Building Research & Information. Vol.38 (1),
217
p.25–41. Available from: doi:10.1080/09613210903236706 [Accessed: 23 November
2012].
German, A., Hoeschele, M., Springer, D. & Davis Energy Group (2014) Maximizing
the Benefits of Residential Pre-Cooling. Proceedings of 2014 ACEEE Summer study
on Energy Efficiency in Buildings. p. 71–83.
Google earth (2015) Version 7.1.5.1557. 207 Ashby Road Loughborough.
52°46’16.19” N 1°13’26.36” W, elev 53m, Eye alt 137m. Bluesky image. Available
online at: <https://www.google.com/earth/>
Google maps (2015) 207 Ashby Road Loughborough. Bluesky map data. Retrieved
from:
<https://www.google.co.uk/maps/place/207+Ashby+Rd,+Loughborough,+Leicestersh
ire+LE11+3AD/@52.7711681,1.2242554,58m/data=!3m1!1e3!4m2!3m1!1s0x4879e0c5f961a1a3:0xc590ef61657d3
ea3> [Accessed on: 2nd September, 2015]
Groot, E. De, Spiekman, M. & Opstelten, I. (2008) 361 : Dutch Research into User
Behaviour in Relation to Energy Use of Residences. PLEA 2008-25th Conference on
Passive and Low energy Architecture.
Gu, L. (2007) Airflow network modelling in EnergyPlus. Proceedings of Building
Simulation 2007. P.964–971.
Guerra Santin, O. & Itard, L. (2010) Occupants’ behaviour: determinants and effects
on residential heating consumption. Building Research and Information. Vol.38 (3),
p.318-338. Available from: doi:10.1080/09613211003661074.
Hacker, J.N., De Saulles, T.P., Minson, A.J. & Holmes, M.J. (2008) Embodied and
operational carbon dioxide emissions from housing: A case study on the effects of
thermal mass and climate change. Energy and Buildings. Vol.40 (3), p.375–384.
Available from: doi:10.1016/j.enbuild.2007.03.005.
Hall, F. & Greeno, R. (2013) Building Services Handbook. 7th Edition. Routledge, UK.
218
Hartless, R. & White, M.K. (1994) Measuring Subfloor Ventilation Rates.
Proceedings of the 15th AIVC Conference: The role of ventilation. 27-30 September
1994, Buxton, Great Britain.
Hartless, R. (2004) Developing a risk assessment framework for landfill gas:
Calculating the probability of a landfill gas explosion. Paper Reference Pa04072,
Building Research Establishment, Watford, UK.
Hartley, A. (2006) West Midlands Health Issues: Fuel poverty. Birmingham, West
Midlands Public Health Observatory. Available online at: <
http://www.wmpho.org.uk/resources/Fuel_Poverty_Short.pdf> [Accessed on: 28 May
2014]
Hartmann, M. (no date) The present development of thermostatic control in individual
rooms. Danfoss. Available from:
http://heating.danfoss.com/pcmpdf/the_present_development.pdf [Accessed: 2nd
September 2015]
Heat Genius (2013) User guide: The Genius Hub. Heat Genius Ltd. Available online
at: < https://www.heatgenius.co.uk/downloads/Heat_Genius_-_User_Guide.pdf>
[Accessed: 2nd September 2015]
Heating and Hot Water Task Force (2010) Heating and Hot Water Pathways to 2020.
Heating and Hot Water Task Force. Available online at: <
http://www.hotwater.org.uk/uploads/559534E788187.pdf> [Accessed: 2nd September
2015]
Hindman, B. (2006) What is an iButton device? Maxim integrated. Available online at:
<http://pdfserv.maximintegrated.com/en/an/AN3808.pdf> [Accessed on: 28 May
2014]
HM Government (2013) The Building Regulations 2010, Conservation of fuel and
power Approved Document L1A (2013 version). NBS, part of RIBA Enterprise Ltd.
Holland, R. (2010) The Experimental Application of an Occupant Tracking
Technology in Domestic Post-Occupancy Evaluation. PhD thesis, University of
219
Nottingham. Available from: http://eprints.nottingham.ac.uk/10235/. [Accessed: 2nd
September 2015]
Honeywell (2014) Valves for heating system control. Honeywell International Inc.
Available online at:
< http://www.honeywelluk.com/documents/Literature/pdf/Flow%20Solutions.pdf>
[Accessed on: 4th September 2015]
Honeywell (2015) Evohome WiFi multi-zone system. Honeywell International Inc.
Available online at: <
http://www.honeywelluk.com/documents/All/pdf/Honeywell_Evohome%20Brochure.p
df> [Accessed on: 4th September 2015]
Huchuk, B., Brien, W.O. & Cruickshank, C. a (2012) Preliminary Results of Model
Predictive Control of Shading Systems (WIP). Proceedings of the sim’AUD 13:
Symposium on Simulation for Architecture & Urban Design.
Huebner, G.M. McMichael, M. Shipworth, D. Shipworth, M. Durand-Daubin, M. &
Summerfield, A. (2013) Heating patterns in English homes: Comparing results from a
national survey against common model assumptions. Building and Environment.
Vol.70, p.298-305.
Humphreys, M.A. (1979) The influence of season and ambient temperature on
human clothing behaviour. Proceedings of Indoor Climate, Danish Building Research,
Copenhagen. p.699-713.
Hunt, D.R.G. & Gidman, M.I. (1982) A national field survey of house temperatures.
Building and Environment. Vol.17 (2), p.107–124. Available from: doi:10.1016/03601323(82)90048-8.
I’anson, S.J., Irwin, C. & Howarth, A.T. (1982) Air Flow Measurement Using Three
Tracer Gases. Building and Environment. Vol.17 (4), p.245–252.
International Energy Agency (1992) Annex 20 : Air Flow Patterns within Buildings subtask2: Air flows between zones- Air flow through large openings in buildings.
Energy conservation in buildings and community systems Programme. EPFL,
220
Lusanne. Available online at: <
http://www.ecbcs.org/docs/annex_20_air_flow_through_large.pdf> [Accessed on 2nd
September 2015]
Invensys Controls (no date) RT212 Thermostatic Radiator Valve. Invensys Controls.
Available online at:
<http://www.draytoncontrols.co.uk/WorkArea/DownloadAsset.aspx?id=12746.>
[Accessed:1st September 2015]
Jeeninga, H., Uyterlinde, M. & Uitzinger, J. (2001) Energy use of energy efficient
residences (in Dutch: Energieverbruik van energiezuinige woningen), ECN rapport
ECN-C-01-072, Petten.
Johnston, D. (2003) A physically-based energy and carbon dioxide emissions model
of the UK housing stock. PhD thesis, Leeds Metropolitan University.
Kane, T. (2013) Indoor temperatures in UK dwellings: investigating heating practices
using field survey data. PhD thesis, Loughborough University. Available from:
https://dspace.lboro.ac.uk/dspace-jspui/handle/2134/12563.
Kavgic, M., Mavrogianni, a., Mumovic, D., Summerfield, a., Stevanovic, Z. &
Djurovic-Petrovic, M. (2010) A review of bottom-up building stock models for energy
consumption in the residential sector. Building and Environment. Vol.45 (7), p.1683–
1697. Available from: doi:10.1016/j.buildenv.2010.01.021.
Kotti, M.C., Argiriou, a. a. & Kazantzidis, a. (2014) Estimation of direct normal
irradiance from measured global and corrected diffuse horizontal irradiance. Energy.
Vol. 70, p. 382–392. Available from: doi:10.1016/j.energy.2014.04.012 [Accessed: 28
January 2015].
Liao, Z., Swainson, M. & Dexter, a. L. (2005) On the control of heating systems in
the UK. Building and Environment. Vol.40 (3), p.343–351. Available from:
doi:10.1016/j.buildenv.2004.05.014 [Accessed: 3 December 2012].
Liddament, M.W. (1986) Air infiltration calculation techniques- an application guide.
Coventry, air infiltration and ventilation centre.
221
Liddament, M.W. (1996) A guide to energy efficient ventilation. Oscar Faber plc on
behalf of the International Energy Agency, St Albans, UK.
Lomas, K. J. & Eppel, H. (1992) Sensitivity analysis techniques for building thermal
simulation programs. Energy and Buildings. Vol.19, p.21-44.
Lomas, K.J., Eppel, H., Martin, C.J. & Bloomfield, D.P. (1997) Empirical validation of
building energy simulation programs. Energy and Buildings. Vol.26 (3), p.253–275.
Available from: doi:10.1016/S0378-7788(97)00007-8.
Lowe, R.J., Wingfield, J., Bell, M. & Bell, J.M. (2007) Evidence for heat losses via
party wall cavities in masonry construction. Building Services Engineering Research
& Technology. Vol.28 (2), p.161-181.
Lstiburek, B.J.W. (2008) Insight New Light in Crawlspaces. ASHRAE Journal of
Building Sciences. Available online at: <
http://buildingscience.com/documents/insights/bsi-009-new-light-in-crawlspaces>
Lu, J., Sookoor, T., Srinivasan, V., Gao, G., et al. (2010) The Smart Thermostat :
Using Occupancy Sensors to Save Energy in Homes. Proceedings of the SenSys
2010, November 3-5 2010, Zurich Switzerland. p. 211–224.
Lutzenhiser, L., Cesafsky, L., Chappells, H., Gossard, M., et al. (2009) Behavioral
Assumptions Underlying California Residential Sector Energy Efficiency Programs.
California, CIEE Behaviour and Energy Program. Available from: <
http://www1.eere.energy.gov/buildings/betterbuildings/neighborhoods/pdfs/ba_ee_re
s_wp.pdf [Accessed on 2nd September 2015]
Megri, A.C. & Haghighat, F. (2007) Zonal Modeling for Simulating Indoor
Environment of Buildings: Review, Recent Developments, and Applications.
HVAC&R Research. Vol.13 (6), p.887–905. Available from:
doi:10.1080/10789669.2007.10391461.
Meier, A., Peffer, T., Pritoni, M., Aragon, C. & Perry, D. (2012) Facilitating energy
savings through enhanced usability of thermostats. Proceedings of ECEEE 2011
summer study, energy efficiency first: the foundation of low carbon society. p.1431222
1438. Available from: <
http://proceedings.eceee.org/visabstrakt.php?event=1&doc=6-094-11> [Accessed:
2nd September 2015]
Meyers, R.J., Williams, E.D. & Matthews, H.S. (2010) Scoping the potential of
monitoring and control technologies to reduce energy use in homes. Energy and
Buildings. Vol.42 (5), p.563–569. Available from: doi:10.1016/j.enbuild.2009.10.026
[Accessed: 13 November 2012].
Mitchell, P. (2008) Central heating: Installation, Maintenance and Repair. Brailsford
Press, Great Britain.
Monetti, V., Fabrizio, E. & Filippi, M. (2015) Impact of low investment strategies for
space heating control: Application of thermostatic radiators valves to an old
residential building. Energy and Buildings. Vol.95, p.202–210. Available from:
doi:10.1016/j.enbuild.2015.01.001.
Munton, A.G., Wright, A., Mallaburn, P. & Boait, P. (2014) How heating controls
affect domestic energy demand : A Rapid Evidence Assessment. Department of
Energy and Climate Change. Available online at:
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/27755
2/FINALHow_heating_controls_affect_domestic_energy_demand__A_Rapid_Evidence_Assessment.pdf [Accessed: 13 November 2012].
Natarajan, S. & Levermore, G.J. (2007) Predicting future UK housing stock and
carbon emissions. Energy Policy. Vol.35 (11), p.5719–5727. Available from:
doi:10.1016/j.enpol.2007.05.034.
National Measurement Office (2014) Gas and electricity meter regulations. HM
Government. Available online at: < https://www.gov.uk/gas-and-electricity-meterregulations> [Accessed on: 28 May 2014]
Negriio, C.O.R. (1998) Integration of computational fluid dynamics with building
thermal and mass flow simulation. Energy and Buildings. Vol.27, p.155–165.
223
Nevius, M.J. & Pigg, S. (2000) Programmable Thermostats that Go Berserk ? Taking
a Social Perspective on Space Heating in Wisconsin. Proceedings of the 2000
ACEEE summer study on energy efficiency in buildings. p.233–244.
Neymark and Judkoff (2002) International Energy Agency Building Energy
Simulation Test and Diagnostic Method for Heating, Ventilation, and Air-Conditioning
Equipment Models (HVACBESTEST- Volume 1: Cases E100-E200) National
Renewable Energy Laboratory, Golden co, USA.
Office for National Statistics (2011) Census 2011. Office for National Statistics.
Available from: http://www.ons.gov.uk/ons/guide-method/census/2011/censusdata/index.html. [Accessed on 2nd September 2015]
Office for National Statistics (ONS) (2002) Time use survey, 2000. Office for National
Statistics. Available online at: <
http://discover.ukdataservice.ac.uk/catalogue?sn=4504> [Accessed on: 28 May 2014]
Office of Public Sector Information (2008) Climate Change Act 2008. The stationary
office limited, UK. Available online at:
<http://www.legislation.gov.uk/ukpga/2008/27/pdfs/ukpga_20080027_en.pdf>
Office of Public Sector Information (2009) The Climate Change Act 2008 (2020
Traget, Credit Limit and Definitions) Order 2009. The stationary office limited, UK.
Osz, K. (2014) Understanding Heating Practices : The Example of TRV Use.
Proceedings of BEHAVE2014-Behaviour and Energy Efficiency Conference, 3-4
September 2014, The University of Oxford. Available online at: <
http://behaveconference.com/wpcontent/uploads/2014/08/F_Katalin_Osz_Loughborough_University.pdf>
Oughton, D. & Hodkinson, S. (2008) Faber & Kell’s Heating and Air-Conditioning of
Buildings. 10th Edition. Oxford, Butterworth-Heinemann.
Palmer, J. & Cooper, I. (2011) Great Britain’s housing energy fact file. Department of
Energy and Climate Change. Available online at: <
224
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/48195/
3224-great-britains-housing-energy-fact-file-2011.pdf>
Peeters, L., de Dear, R., Hensen, J., and D’haeseleer, W., (2009) Thermal comfort in
residential buildings: Comfort values and scales for building energy simulation.
Applied Energy. Vol.85(5) p.772-80. DOI:
<http://dx.doi.org/10.1016/j.apenergy.2008.07.011>
Peffer, T., Pritoni, M., Meier, A., Aragon, C., et al. (2011) How people use
thermostats in homes: A review. Building and Environment. Vol.46 (12), 2529–2541.
Available from: doi:10.1016/j.buildenv.2011.06.002 [Accessed: 13 November 2012].
Porritt, S.M. (2012) Adapting UK Dwellings for Heat Waves. PhD Thesis, De
Montford University, Leicester.
Pupeikis, D., Burlingis, A. & Stankevičius, V. (2010) Required additional heating
power of building during intermitted heating. Journal of Civil Engineering and
Management. Vol.16 (1), pp.141–148. Available from: doi:10.3846/jcem.2010.13.
Race, G.L. (2005) Understanding controls-CIBSE knowledge series: KS4. The
Chartered Institution of Building Services Engineers London
Rayment, R. & Morgan, K. (1984) Comparing Conventional and Electronic RoomStats. Batiment International, Building Research and Practice. Vol.12 (2), p.114–121.
Available from: doi:10.1080/09613218408551206.
Rayment, R., Cunliffe, A.R. & Morgan, K. (1983) Basic characteristics of low-cost
houses in order to reduce the energy consumption for heating- efficiency of heating
plants and controls. Commission of the European Communities. Available from:
doi:10.1016/S0172-2190(79)90045-0.
Reddy, T. Agami (2006) Literature review on calibration of building energy simulation
programs: uses, problems, procedures, uncertainty and tools. ASHRAE Transactions.
Vol. 112, Issue 1, p226.
225
Roberts, S. (2008) Altering existing buildings in the UK. Energy Policy. Vol.36 (12),
p.4482–4486. Available from: doi:10.1016/j.enpol.2008.09.023 [Accessed: 21 May
2013].
Rock, I. (2005) The 1930s House Manual. Haynes Publishing, Sparkford, Somerset,
UK: Number ISBN 1 84425 214 0.
Royapoor, M. & Roskilly, T. (2015) Building model calibration using energy and
environmental data. Energy and Buildings. Vol.94, p.109–120. Available from:
doi:10.1016/j.enbuild.2015.02.050.
Salus controls (2013) Salus programmable thermostatic radiator valve control head:
Instruction manual. Salus Controls plc. Available online at: < http://www.salustech.com/products/accessories/thermostatic-radiator-valve/_c1_84_ph60/>
[Accessed on: 4th September 2015]
Sanders, C., Haig, J. & Reideout, N. (2006) Airtightness of Ceilings: Energy loss and
condensation risk (IP4/06). BRE.
Scott, J., Brush, A.J.B., Krumm, J., Meyers, B., et al. (2011) PreHeat : Controlling
Home Heating Using Occupancy Prediction. Proceedings of UbiComp’11,
September 17–21, 2011, Beijing, China.
Shipworth, M., Firth, S.K., Gentry, M.I., Wright, A.J., et al. (2010) Central heating
thermostat settings and timing: building demographics. Building Research &
Information. Vol.38 (1), p.50–69. Available from: doi:10.1080/09613210903263007
[Accessed: 21 November 2012].
Shorrock, L. & Dunster, J. (1997) The physically-based model BREHOMES and its
use in deriving scenarios for the energy use and carbon dioxide emissions of the UK
housing stock. Energy Policy. Vol.25 (12), p.1027–1037. Available from:
doi:10.1016/S0301-4215(97)00130-4.
Sontex SA (2014a). The multi-functional integrator Supercal 531. Sontex SA.
Available online at: <http://www.sontex.ch/data/FT-531-e-03.pdf > [Accessed on: 28
May 2014]
226
Sontex SA (2014b) Superstatic 440: Static fluidic oscillator heat meter. Sontex SA.
Available online at: <http://www.sontex.ch/data/Superstatic440%20e_12%20Web.pdf>
[Accessed on: 28 May 2014]
Stephen, R.K. (2000) Airtightness in UK Dwellings: Information Paper IP 1/00. BRE.
Straube, J. (2008) Air flow control in buildings. Building Science Digest 014. Building
Sciences Press. Available online at: <
http://buildingscience.com/documents/digests/bsd-014-air-flow-control-in-buildings>
[Accessed on: 7th September 2015]
Swan, L.G. & Ugursal, V.I. (2009) Modeling of end-use energy consumption in the
residential sector: A review of modeling techniques. Renewable and Sustainable
Energy Reviews. Vol.13 (8), p.1819–1835. Available from:
doi:10.1016/j.rser.2008.09.033.
Tachibana, D. (2010) Residential Customer Characteristics Survey 2009. The City of
Seattle, Seattle City Light Department. Available from: <
http://www.seattle.gov/light/Conserve/Reports/Evaluation_15.pdf> [Accessed: 2nd
September 2015]
TACMA (2010) Guidance on how to comply with the 2010 Building Regulations Part
L. Version 2.0. 10 December 2010, TACMA. Available online at:
< http://www.beama.org.uk/asset/CD897A84-EDB1-4FA7-BB1E56C226D4AE45/>
[Accessed: 2nd September 2015]
Tahersima, F., Stoustrup, J. & Rasmussen, H. (2013) An analytical solution for
stability-performance dilemma of hydronic radiators. Energy and Buildings. Vol.64,
p.439–446. Available from: doi:10.1016/j.enbuild.2013.05.023.
Tan, G. & Glicksman, L.R. (2005) Application of integrating multi-zone model with
CFD simulation to natural ventilation prediction. Energy and Buildings. Vol.37 (10),
p.1049–1057. Available from: doi:10.1016/j.enbuild.2004.12.009.
Taylor, S., Allinson, D., Firth, S. & Lomas, K. (2013) Dynamic energy modelling of
UK housing: evaluation of alternative approaches. Proceedings of 13th Conference
227
of International Building Performance Simulation Association. August 26-28 2013,
Chambery, France. P.745–752.
The Carbon Trust (2011) Heating control: maximising comfort, minimising energy
consumption. The Carbon Trust, London, UK. Available online at: <
https://www.carbontrust.com/media/10361/ctg065_heating_control.pdf>
The Committee on Climate Change (2013) Review of potential for carbon savings
from residential energy efficiency. The Committee on Climate Change. Available
online at: < https://www.theccc.org.uk/wp-content/uploads/2013/12/Review-ofpotential-for-carbon-savings-from-residential-energy-efficiency-Final-report-A160114.pdf> [Accessed: 2nd September 2015]
The Office of the Deputy Prime Minister (2005) Guide to the condensing boiler
installation assessment procedure for dwellings. The Office of the Deputy Prime
Minister, London, UK. Available online at: <
https://www.gov.je/SiteCollectionDocuments/Planning%20and%20building/ID%20As
sessmentGuidetoCondensingBoilerInstallationinDwellings%2020100910%20mm.pdf>
[Accessed: 2nd September 2015]
HSE (2012) Managing the risks from hot water and surfaces in health and social
care. The Health and Safety Executive. Available online at:<
http://www.hse.gov.uk/pubns/hsis6.pdf> [Accessed on: 3rd September 2015]
UK Meteorological Office (2012) MIDAS Land Surface Stations data (1853-current).
British Atmospheric Data Centre. Available online at:
<http://badc.nerc.ac.uk/data/ukmo-midas> [Accessed on: 28 May 2014]
University of the West of England (2009) Evolution of Building Elements. The
construction website. Available online at: <
http://fet.uwe.ac.uk/conweb/house_ages/elements/index.htm > [Accessed on 2nd
September 2015]
US Department of Energy (2001) Residential Energy Consumption Survey (RECS).
Energy Information Administration, Available online at:
<http://www.eia.gov/consumption/residential/>
228
US Department of Energy (2012) EnergyPlus Engineering Reference. US
Department of Energy. Available online at: <
http://apps1.eere.energy.gov/buildings/energyplus/pdfs/engineeringreference.pdf>
[Accessed: 2nd September 2015]
US Department of Energy (2013a) Application Guide for EMS: Energy Management
System User Guide. US Department of Energy. Available online at:<
http://apps1.eere.energy.gov/buildings/energyplus/pdfs/ems_application_guide.pdf>
[Accessed: 2nd September 2015]
US Department of Energy (2013b) Auxiliary EnergyPlus Programs: extra programs
for EnergyPlus. US Department of Energy. Available online at: <
http://apps1.eere.energy.gov/buildings/energyplus/pdfs/auxiliaryprograms.pdf>
[Accessed: 2nd September 2015]
US Department of Energy (2013c) Input Output Reference: The Encyclopedic
Reference to EnergyPlus Input and Output. US Department of Energy. Available
online at:
http://apps1.eere.energy.gov/buildings/energyplus/pdfs/inputoutputreference.pdf
[Accessed: 2nd September 2015]
Utley, J.I. & Shorrock, L.D. (2008) Domestic energy fact file 2008. BRE.
Vera control Ltd (2014) Vera3. Vera control Ltd. Available online at:
<http://getvera.com/controllers/vera3/>.
Vesma, V. (2014) Energy saving through boiler water treatment : limits of plausibility.
Degree Days Direct Limited, Gloucestershire, UK. Available online at: <
http://www.vesma.com/downloads/op_treatment_savings.pdf> [Accessed: 2nd
September 2015]
Wang, Z.C. & Di, H.F. (2002) The state of the art and problems in heat metering and
billing, District Heating. Vol.3 (2002) 1–5 (in Chinese).
Watkins, D.E. (2011) Heating Services in Buildings. John Wiley & Sons, West
Sussex, UK.
229
Wei, S., Jones, R. & De Wilde, P. (2014) Driving factors for occupant-controlled
space heating in residential buildings. Energy and Buildings. Vol.70, p.36–44.
Available from: doi:10.1016/j.enbuild.2013.11.001.
Weker, P. & Mineur, J.M. (1980) A performance index for thermostatic radiator
valves. Applied Energy.Vol.6 (3), p.203–215. Available from: doi:10.1016/03062619(80)90034-3.
Westphal, F.S. & Lamberts, R. (2005) Building Simulation Calibration Using
Sensitivity Analysis. Proceedings of Building Simulation 2005. pp. 1331–1338.
Wingfield, J., Johnston, D., Miles-Shenton, D. & Bell, M. (2010) Whole house heat
loss test method (Co heating). Leeds Metropolitan University. Available online at: <
http://www.leedsbeckett.ac.uk/as/cebe/projects/coheating_test_protocol.pdf>
[Accessed on: 28 May 2014]
Worcester Bosch Group (2009) Engineer’s service booklet: Floor mounted and wall
hung RSF gas fired condensing boilers- GREENSTAR CDI range. Worcester Bosch
Group, Worcester, UK.
Wright, L. (1964). Home fires burning. Routledge & Kegan Paul Ltd, London. UK
Xu, B., Fu, L. & Di, H. (2008) Dynamic simulation of space heating systems with
radiators controlled by TRVs in buildings. Energy and Buildings. Vol.40 (9), p.1755–
1764. Available from: doi:10.1016/j.enbuild.2008.03.004 [Accessed: 5 December
2014].
Yilmaz, S., Allinson, D., Taylor, S. & Lomas, P.K. (2014) A parametric analysis of the
energy savings predicted by different thermal models: implications for the UK Green
Deal. Proceedings of Building Simulation and Optimization Conference (BS014), 2324 June 2014, UCL.
Yoon, J., Lee, E.J. & Claridge, D.E. (2003) Calibrating procedure for energy
performance simulation of a commercial building. Journal of Solar Energy
Engineering. Vol.125, pp. 251-257.
230
Young, B., Shiret, A., Hayton, J. & Griffiths W. (2013) Design of low-temperature
domestic heating systems. Watford, IHS BRE press. Available online at:
<http://www.brebookshop.com/samples/327255.pdf> [Accessed on: 3rd September
2015]
Zhai, Z.J. & Chen, Q.Y. (2005) Performance of coupled building energy and CFD
simulations. Energy and Buildings. Vol.37 (4), p.333–344. Available from:
doi:10.1016/j.enbuild.2004.07.001.
231
A.1 Appendix 1: Blower door test reports
A.1.1 House 1
Figure A-1: Blower door test report for House 1
232
A.1.2 House 2
Figure A-2: Blower door test report for House 2
233
A.2. Appendix 2: EMS Code for boiler control
! Boiler thermostatic control of 207 house
EnergyManagementSystem:Sensor,
Hallway_Air_Temperature207,
GroundFloor:Hallway207,
!- Name
!-
Output:Variable Index Key Name
Zone Mean Air Temperature;
!-
Output:Variable Name
EnergyManagementSystem:Actuator,
Actuator_Loop,
HW LoopZC,
!- Name
!- Actuated
Component Unique Name
Plant Loop Overall,
!- Actuated
Component Type
On/Off Supervisory;
!- Actuated
Component Control Type
EnergyManagementSystem:Actuator,
PumpFlowOverride,
!- Name
HW LoopZC Supply Pump,
!- Actuated
Component Unique Name
Pump,
!- Actuated
Component Type
Pump Mass Flow Rate;
!- Actuated Component
Control Type
EnergyManagementSystem:GlobalVariable,
PumpFlowOverrideReport;
EnergyManagementSystem:OutputVariable,
EMS Boiler Flow Override On [On/Off],
!- Name
234
PumpFlowOverrideReport,
!- EMS Variable
Name
Averaged,
!- Type of Data in
Variable
SystemTimeStep;
!- Update Frequency
EnergyManagementSystem:ProgramCallingManager,
HW LoopZC OnOff Management,
!-
Management type
InsideHVACSystemIterationLoop,
BoilerControl;
!- Calling time
!- Program
EnergyManagementSystem:Program,
BoilerControl,
IF (Hallway_Air_Temperature207 > 21.0),
!- Name
!- Conditional
statement
SET Actuator_Loop = 0.0,
SET PumpFlowOverride = 0.0,
SET PumpFlowOverrideReport = 1.0,
ELSE,
SET Actuator_Loop = Null,
SET PumpFlowOverride = Null,
SET PumpFlowOverrideReport = 0.0,
ENDIF;
Output:Variable,
*,
EMS Boiler Flow Override On,
!- Output
variable name
Hourly;
! Boiler thermostatic control of 209 house
EnergyManagementSystem:Sensor,
Hallway_Air_Temperature209,
!- Name
235
GroundFloor:Hallway209,
!-
Output:Variable Index Key Name
Zone Mean Air Temperature;
!-
Output:Variable Name
EnergyManagementSystem:Actuator,
Actuator_Loop1,
HW LoopCC,
!- Name
!- Actuated
Component Unique Name
Plant Loop Overall,
!- Actuated
Component Type
On/Off Supervisory;
!- Actuated
Component Control Type
EnergyManagementSystem:Actuator,
PumpFlowOverride1,
!- Name
HW LoopCC Supply Pump,
!- Actuated
Component Unique Name
Pump,
!- Actuated
Component Type
Pump Mass Flow Rate;
!- Actuated
Component Control Type
EnergyManagementSystem:GlobalVariable,
PumpFlowOverrideReport1;
EnergyManagementSystem:OutputVariable,
EMS Boiler1 Flow Override On [On/Off],
!- Name
PumpFlowOverrideReport1,
!- EMS Variable
Name
Averaged,
!- Type of Data in
Variable
SystemTimeStep;
!- Update Frequency
236
EnergyManagementSystem:ProgramCallingManager,
HW LoopCC OnOff Management,
!-
Management type
InsideHVACSystemIterationLoop,
BoilerControl1;
!- Calling time
!- Program
EnergyManagementSystem:Program,
BoilerControl1,
IF (Hallway_Air_Temperature209 > 21.0),
!- Name
!- Conditional
statement
SET Actuator_Loop1 = 0.0,
SET PumpFlowOverride1 = 0.0,
SET PumpFlowOverrideReport1 = 1.0,
ELSE,
SET Actuator_Loop1 = Null,
SET PumpFlowOverride1 = Null,
SET PumpFlowOverrideReport1 = 0.0,
ENDIF;
Output:Variable,
*,
EMS Boiler1 Flow Override On,
!- Output
variable name
Hourly;
237
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising