A Low-Cost Intelligent Thermostat Incorporating a Learning

A Low-Cost Intelligent Thermostat Incorporating a Learning
A Low-Cost Intelligent Thermostat Incorporating a Learning Algorithm
Based Scheduler and Internet Accessible Control Interfaces
1
M.M. A. Rahman, D. Huizenga, 2E. Chiaburu
1
School of Engineering, Grand Valley State University, Grand Rapids, MI
2
Gentex Corporation, Zeeland, MI
Email: [email protected]
Abstract
This paper presents an internet connected programmable thermostat that automatically generates
an energy efficient heating and cooling schedule through the collection and analysis of
occupancy data. The implementation provides an easy to use color touch screen, an on-device
browser accessible web interface, and an Android smartphone interface. In this capstone project
work, the proposed system was developed and the proposed features were validated using a
commercial PIC32 development board. The novel features of the Intelligent Thermostat
represent a substantial evolution of thermostat technology. Currently, no commercial thermostat
is available that provides automatic scheduling capability, and no commercial thermostat is
available that serves a web connection feature from the device itself. The only announced
commercial thermostat with automatic scheduling capability is slated to cost substantially more
than the Intelligent Thermostat. The fully automated energy savings capabilities of the Intelligent
Thermostat will save the average homeowner thousands of dollars over the life of the hardware
and, with widespread deployment, could eliminate the emission of millions of tons of CO2
annually. Given the easily available energy savings with virtually no impact on a homeowner’s
quality of life, it is believed that automatic scheduling thermostats represent the future standard
in programmable thermostat technology.
Introduction
According to US EPA research, the average household in the United States spends about $2200
per year on energy [2]. Of this, approximately 50% is spent on heating and cooling. The US
DOE calculates that a savings of 1% can be achieved for each degree that a programmable
thermostat is set back during a period of at least eight hours. In total, a heating and cooling
schedule that meets US EPA guidelines will save the average US household about $180 per year
[3]. Prior to the implementation of this project, numerous Internet-connected programmable
thermostats were surveyed. Even the most advanced thermostats had shortcomings: (1) they
required the user to complete a tedious programming process to take advantage of energy saving
heating and cooling scheduling capabilities, (2) internet connectivity was achieved through
connection to the manufacturer’s server network, limiting future utility of web connection
features to the window during which the manufacture continued to maintain the server network,
(3) thermostats with modern color touch screen user interfaces and web connectivity retailed for
in excess of $300. These limitations have resulted in limited adoption of programmable
thermostats and limited utilization of programmability even in those cases where a
programmable thermostat is installed. In fact, a study conducted in Connecticut found that only
36% of the homes within the test group (who had been provided with programmable thermostats
free of charge) were using the automatic temperature setback capability of the thermostat [1].
The Intelligent Thermostat solves each of these problems with unique features that were not
Proceedings of the 2014 ASEE North Central Section Conference
Copyright © 2014, American Society Engineering Education
found in any announced or available commercial thermostat when this project was started: (1)
The Intelligent Thermostat completely automates schedule generation, and as a result, automates
energy savings (2) The Intelligent Thermostat hosts the web server for the Internet interface on
the device, eliminating dependency on the continued availability of third party servers, (3) The
materials cost for a custom hardware implementation of the Intelligent Thermostat would be
approximately $110 per unit (without considering volume discounts), (4) In addition, the
Intelligent Thermostat is designed to achieve parity with the standard scheduling and temperature
control features provided by more expensive units, (5) Using the EPA estimates, the cost
breakeven point for the Intelligent Thermostat occurs after less than 10 months of use. (6) Since
the Intelligent Thermostat is able to automatically tailor its schedule to the user, and provides
remote override capability for unscheduled occupancy, it is configured by default to use more
aggressive setbacks than the EPA standards. As a result, the Intelligent Thermostat should
exceed EPA energy savings, further accelerating breakeven.
Market Survey: Programmable Thermostats
A market survey was conducted to determine the state of the art for energy saving thermostats.
When this project was started, two thermostats stood out as providing more advanced features
and capabilities than those offered by typical programmable thermostats: the Honeywell Prestige
HD 2.0 with optional RedLINK Internet gatway, and the Ecobee EB-STAT-02. In addition, two
other thermostats are presented here: The Nest Learning Thermostat and the 3M Filtrete WiFi
Programmable Thermostat. The Nest (which was announced during development of the
Intelligent Thermostat) provides a learning capability similar to that of the Intelligent
Thermostat, while the 3M Filtrete WiFi Programmable Thermostat was the lowest cost webconnected thermostat found during the survey. Table 1 compares these thermostats to the
implemented Intelligent Thermostat.
Table 1 Commercial Thermostat Comparisons
Proceedings of the 2014 ASEE North Central Section Conference
Copyright © 2014, American Society Engineering Education
Block Diagram of Intelligent Thermostat
This advanced thermostat has four implementation areas: firmware, web interface, android app,
and hardware. A block diagram depicting the functional interfaces among these implementation
areas is presented in Fig. 1.
Figure 1 Intelligent Thermostat block diagram
Thermostat Firmware
The thermostat firmware provides many core features to the Intelligent Thermostat. It is responsible
for: controlling the HVAC system, maintaining the data stores that back the automatic schedule
generation algorithm, automatically generating the intelligent HVAC schedule, executing HVAC
schedules, serving the Web interface, executing the local touch screen User Interface, and executing
an HTTP based Remote Procedure Call application supporting multiple simultaneous clients.
Detailed information on the firmware’s feature set and the underlying implementation is available in
[17] and beyond the presentation scope of this paper.
Occupancy Sensor Network and Schedule Generation
The HVAC Intelligence Module is responsible for system mode transitions, collecting and
logging occupancy sensor network data, and automatic schedule generation based on occupancy
logs. The Intelligence Module supports a total of 8 system modes, which automatically reconfigure
the thermostat’s internal state to achieve a desired behavior. The available system modes are detailed
in the Table 2. The Intelligence Module’s automatic schedule generation functionality has four
primary stages: data combination (based on analysis group numbers), occupancy probability
generation, probability threshold application, and schedule windowing application. A flowchart
describing the automatic schedule generation algorithm is presented in Figure 2.
Proceedings of the 2014 ASEE North Central Section Conference
Copyright © 2014, American Society Engineering Education
Table 2 HVAC System Modes
Figure 2 Automatic schedule generation algorithm
Proceedings of the 2014 ASEE North Central Section Conference
Copyright © 2014, American Society Engineering Education
Web Interface and Android Application
The web interface is served by the thermostat firmware in response to HTTP requests to its root
directory. The web interface provides: set-point and mode management; graphical manual
schedule creation and editing; and thermostat configuration editing. Android OS is 100% open
platform and was used for remote access. The goal of the Android application was to allow users
easy access to all thermostat features from anywhere using their Android smartphone. The
application had to be easy to use, yet capable of modifying the thermostat’s more advanced
settings. Detailed information about these functions is available in [17].
Hardware Implementation and Results
During hardware selection, a survey was performed of the available microcontroller
architectures, supporting software toolsets, and available development boards. The PIC32MX7
architecture stood out in particular because it offered microcontrollers that provided sufficient
resources to implement the desired feature set, an integrated MAC, low cost, and a commercial
quality IDE with a generous trial period. After selection of a platform, a survey was conducted of
available development boards, and the MicroElektronica MultiMeda Board (MMB) for
PIC32MX7 was selected. This board offered the following features needed to implement the
Intelligent Thermostat demonstration:
· PIC32MX795F512L based design
· QVGA touchscreen
· Ethernet phy + RJ45 port
· LEDs for simulation of HVAC outputs
· Switch for simulation of occupancy input
· Integrated temperature sensor
· In-Circuit debugging and programming capability
· Additional peripherals for potential capabilities growth (SD card support, audio codec, onboard
EEPROM and flash, etc.)
Details on the development board components and implementation are available in [17].
Proposed intelligent thermostat with all desired features was successfully demonstrated on the
development board and some of the pictures taken from the demonstrated prototype are provided
in Fig. 3.
Conclusions
A low-cost, web connected and smartphone accessible, Intelligent Thermostat capable of
creating operating schedule based on learning algorithm was designed and implemented on a
PIC32development board. The novel features of this capstone project on the Intelligent
Thermostat represent a substantial evolution of thermostat technology. Currently, no commercial
thermostat is available that provides automatic scheduling capability and that serves a web
connection features from the device itself. The only announced commercial thermostat with
automatic scheduling capability is slated to cost substantially more than the Intelligent
Thermostat. The fully automated energy savings capabilities of the Intelligent Thermostat will
save the average homeowner thousands of dollars over the life of the hardware, and with
widespread deployment, could eliminate the emission of millions of tons of CO2 annually. Given
the easily available energy savings with virtually no impact on a homeowner’s quality of life, we
believe that the learning algorithm based automatic scheduling thermostats represent the future
standard in programmable thermostat technology.
Proceedings of the 2014 ASEE North Central Section Conference
Copyright © 2014, American Society Engineering Education
Figure 3(a) Configuration editor
Figure 3(c) Direct control view
Figure 3(b) Application welcome screen
Figure 3(d) Set temperature dialog
Proceedings of the 2014 ASEE North Central Section Conference
Copyright © 2014, American Society Engineering Education
Figure 3(e) Add schedule dialog
Figure 3(g) Add transition dialog
Figure 3(i) Temperature and humidity graph
Figure 3(f) Schedule list context menu
Figure 3(h) Edit transition dialog
Figure 3(j) HVAC and occupancy graph
Proceedings of the 2014 ASEE North Central Section Conference
Copyright © 2014, American Society Engineering Education
Bibliography
1. David Cross and David Judd. Consortium for Energy Efficiency. [Online].
http://www.cee1.org/eval/db_pdf_es/953es.pdf
2. U.S. Environmental Protection Agency. (2011, Dec.) Programmable Thermostats: ENERGY START. [Online].
http://www.energystar.gov/index.cfm?fuseaction=find_a_product.showProductGroup&pgw_code=TH
3. U.S. Department of Energy. (2011, Dec.) Energy Savers: Thermostats and Control Systems. [Online].
http://www.energysavers.gov/your_home/space_heating_cooling/index.cfm/mytopic=12720
4. (2011, Dec.) Prestige TH9321 Operating Manual. [Online].
http://customer.honeywell.com/techlit/pdf/PackedLit/69-2493EFS.pdf
5. Ecobee. (2011, Dec.) SMART THERMOSTAT User Manual. [Online].
http://www.ecobee.com/wp-content/themes/ecobee/etc/file/UMR5-041510web.pdf
6. Nest. (2011, Dec.) Nest Support. [Online]. http://support.nest.com/
7. 3M. (2011, Dec.) Operation Guide 3M-50. [Online].
http://www.radiothermostat.com/filtrete/guides/3M-50-Operation-13may10.pdf
8. (2011, Dec.) PexSupply.com. [Online]. http://www.pexsupply.com/Honeywell-THX9421R5013-Prestige-IAQ-20-HD-Comfort-Thermostat
9. (2011, Dec.) pexsupply.com. [Online]. http://www.pexsupply.com/Honeywell-THM6000R1002-RedLINKInternet-Gateway
10. (2011, Dec.) Amazon.com. [Online]. http://www.amazon.com/Ecobee-EB-STAT-02-Enabled-InternetThermostat/dp/B004150PJG/ref=sr_1_1?ie=UTF8&qid=1323131122&sr=8-1
11. (2011, Dec.) Nest Store. [Online]. http://store.nest.com/
12. (2011, Dec.) The Home Depot. [Online]. http://www.homedepot.com/
13. (2011, Dec.) Java, Eclipse, Android, and Web Programming Tutorials. [Online].
http://www.vogella.de/android.html
14. Rito Meier, Professional Android 2 Application Development, 2nd ed., Milan Shah, 46 Ed. Indianapolis, IN:
Wiley, 2011.
16. US EPA. (2011) Encouraging Energy Efficiency in New and Existing Homes.
17. (2012) E. chiaburu, et.al An Implementation of a Low Cost Intelligent Thermostat Incorporating a Learning
Algorithmic Scheduler and Internet Accessible Control Interfaces Capstone project report.
Proceedings of the 2014 ASEE North Central Section Conference
Copyright © 2014, American Society Engineering Education
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

advertisement