UHVnet 2011 uhvnet Fourth UHVnet Colloquium January 18th – 19th 2011 Winchester, UK This colloquium is supported by: UHVnet 2011 uhvnet Welcome to the fourth UHVnet colloquium hosted by the University of Southampton on the 18 th and 19th January 2011. Previous UHVnet events have been held at Cardiff University, Glasgow Caledonian University and the University of Manchester. This meeting will take place at Winchester Guildhall and consist of a registration and poster session on the Tuesday evening to encourage a relaxed discussion of current work between early career researchers. The second day of the event will consist of oral presentations covering the following four topic areas; High Voltage Plant, Condition Monitoring, Materials and Theories, Methods and Models. Each topic will have an opening presentation by a leading researcher in the field followed by 5 oral presentations by early career researchers and postgraduates. UHVnet is an informal grouping of universities and was set up in 2005 to further interests of high voltage research within the United Kingdom. The university members are Cardiff University, Glasgow Caledonian University, University of Liverpool, University of Leicester, University of Manchester, University of Southampton, University of Strathclyde and the University of Surrey. These universities are supported by a steering group which includes industrial representation from the Areva T&D Technology Centre, PPA Energy, National Grid and Narec. Specific objectives of the group include raising awareness of the researcher capabilities of group members to UK high-voltage related industry, particularly manufacturers and electricity supply companies and lobbying research funding organisations for ear-marked high-voltage related programs. We would be delighted to receive any feedback about this event as we are keen to further improve our communication with both UK and overseas stakeholders. Future events will be listed on our website and we hope to see you again. UHVnet 2011 Table of Contents Timetable.............................................................................................................................................................. iii UHVnet Members ................................................................................................................................................ iv 2011 Organising Committee ................................................................................................................................ v Transport .............................................................................................................................................................. vi Technical Programme Oral Session 1: High Voltage Plant .................................................................................................................... 1 Characterisation of Earth Electrodes including Experimental Tests on Large-Scale A1.1 2 Systems ....................................................................................................................................................... The Variation in Tangential Electric Fields for Different Bushing Shed Profiles in a A1.2 3 Polluted Environment .................................................................................................................................. A1.3 Environmental Friendly Compact High Voltage Substations ..................................................................... 4 A1.4 A Cascaded Flying Capacitor Multilevel Converter for HVDC and FACTS ............................................. 5 A1.5 Thermal Performance of High Voltage Power Cables ................................................................................ 6 Detection and Location of Underground Power Cable using Magnetic Field A1.6 7 Technologies ............................................................................................................................................... Oral Session 2: Condition Monitoring ................................................................................................................ 8 B1.1 Condition Monitoring for High Voltage Equipment ................................................................................... 9 Automated Phase-agnostic Time Domain Analysis of RF Partial Discharge Pulse Data B1.2 10 for Low-power Wireless Sensing Applications........................................................................................... B1.3 Partial Discharge Analysis of Defective Three-phase Cable ...................................................................... 11 B1.4 Optimum Coil Design for Inductive Energy Harvesting in Substations .................................................... 12 Study of Signal Processing Techniques used for Denoising Partial Discharge Signals in B1.5 13 Cables ......................................................................................................................................................... B1.6 Instrumentation and Condition Monitoring of Composite Cross Arms ...................................................... 14 Oral Session 3: Materials .................................................................................................................................... 15 C1.1 Towards Recyclable Insulation Materials for High Voltage Cables ........................................................... 16 C1.2 Five-Electrode Inclined-plane Tests of Textured Silicone Rubber Samples ............................................... 17 C1.3 A Raman Microprobe Study of Corona Ageing in a Controlled Atmosphere ............................................ 18 FTIR Spectrum of Layered PET Insulation with Artificial Voids Subjected to Electrical C1.4 19 Stressing ...................................................................................................................................................... C1.5 Dielectric Behaviour of Alkyl Esters of Seed-based Oil ............................................................................. 20 C1.6 The Influence of Spherical Cavity Surface Charge Distribution on PD Events .......................................... 21 Oral Session 4: Theories, Methods and Models ................................................................................................. 22 D1.1 Stochastic and Deterministic models for Electrical Tree Growth ............................................................... 23 Model of Structural Damage to Carbon Fibre Composites Due to Thermo-electric Effects D1.2 24 of Lightning Strikes ..................................................................................................................................... D1.3 Switching Ferroresonant Transient Study using Finite Element Transformer Model ................................. 25 D1.4 Transient Modelling of Offshore Wind Farm Connections ....................................................................... 26 D1.5 Surface Gradient Calculation for Overhead Lines ...................................................................................... 27 D1.6 Modelling of Electroluminescence in Polymers Using a Bipolar Charge Transport Model ....................... 28 Posters: High Voltage Plant ................................................................................................................................ 29 Power Transformer End-of-life Modelling: Incorporating Thermal Lifetime Analysis A2.1 30 with Ordinary Statistical Analysis ............................................................................................................... Location of Partial Discharges within a Transformer Winding Using Principal A2.2 31 Component Analysis ................................................................................................................................... Frequency Response Analysis of Transformer Winding Deformation Based on MultiA2.3 32 conductor Transmission Line Model ........................................................................................................... A2.4 Effect of Climatic Condition on Polymeric Insulators ................................................................................ 33 A2.5 Acoustic Noise Evaluation for Overhead Lines .......................................................................................... 34 A2.6 Transient Fault Location in Low Voltage Distribution Networks ............................................................... 35 A2.7 A Survey on the Potential of CF3I Gas as an Alternative for SF6 .............................................................. 36 A New Technique to Enhance the Earthing System by Increasing the Horizontal Earth A2.8 37 Electrode Effective Length ......................................................................................................................... i UHVnet 2011 A2.9 A Novel Portable Testing Device for Surge Protective Systems ....................................................................... 38 A2.10 A Solar Powered Wireless Data Acquisition System for High Voltage Substations ......................................... 39 A2.11 The Performance of Nanocoating on High Voltage insulators ........................................................................... 40 A2.12 Performance of Tower Footings Resistance under High Impulse Current ......................................................... 41 A2.13 High Frequency Performance of a Vertical Rod Electrode ................................................................................ 42 Posters: Condition Monitoring ............................................................................................................................ 43 FDTD Modelling of Partial Discharge Detection in Power Distribution Cables using B2.1 44 HFCTs ................................................................................................................................................................ B2.2 Use of Hidden Markov Model for Partial Discharge-led Failure Development Modelling ............................... 45 Dynamically Weighted Ensemble of Neural Networks for Classifying Partial Discharge B2.3 46 Patterns ............................................................................................................................................................... B2.4 A Successful On-site PD Testing Experience of 11kV EPR Cable Insulation Systems .................................... 47 B2.5 Radiometric Arc Fault Detection ........................................................................................................................ 48 B2.6 Voltage Transducer for Transient Measurements on High Voltage Overhead Lines ........................................ 49 B2.7 Fault Location using FPGAs and Power Line Communication ......................................................................... 50 A New Method to Improve the Sensitivity of Leak Detection in Self-Contained FluidB2.8 51 filled Cables ....................................................................................................................................................... Energy Harvesting from Electric Fields in Substations for Powering Autonomous B2.9 52 Sensors................................................................................................................................................................ Ageing and Temperature Influence on Polarization/Depolarization Current Behaviour of B2.10 53 Paper Immersed in Natural Ester ........................................................................................................................ B2.11 An On-line Lightning Monitoring System for Transmission Lines ................................................................... 54 Energy Harvesting in Substations for Wireless Sensors and a New Arc Capacitor B2.12 55 Structure Posters: Materials ................................................................................................................................................ 56 On the use of Raman and FTIR Spectroscopy for the Analysis of Silica-based C2.1 57 Nanofillers .......................................................................................................................................................... C2.2 Dielectric Breakdown Strength of Polyethylene Nanocomposites ..................................................................... 58 Influence of Temperature and Moisture Absorbed on Electrical Degradation and C2.3 59 Breakdown in Epoxy Resins ............................................................................................................................. C2.4 Space Charge Behaviour in Oil-Paper Insulation with Different Aging Condition .......................................... 60 Modelling the Non-equilibrium Electric Double Layer at Oil-pressboard Interface of C2.5 61 High Voltage Transformers ................................................................................................................................ Investigation of Impulsive Corona Discharges for Energisation of Electrostatic C2.6 62 Precipitation Systems ......................................................................................................................................... A Comparison of Polymeric Cable Insulation Properties Following Lightning Impulse C2.7 63 Ageing ................................................................................................................................................................ C2.8 Properties and Analysis of Thermally Aged Poly(ethylene oxide) .................................................................... 64 C2.9 Smart Materials as Intelligent Insulation ........................................................................................................... 65 C2.10 AC Breakdown Characteristics of LDPE in the Presence of Crosslinking By-products.................................... 66 C2.11 DC Impulse Discharge Degradation of Mica 67 Posters: Theories, Methods and Models ............................................................................................................. 68 D2.1 Modelling of Partial Discharge Activity in Cavity within a Dielectric Insulation Material ............................... 69 Full Wave Modelling of Partial Discharge Phenomena in Power Transformers using D2.2 70 FDTD Methods ................................................................................................................................................... D2.3 Evaluation of an Iterative Method used for Partial Discharge RF Location Techniques ................................... 71 Numerical Modelling of Needle-Grid Electrodes Negative Surface Corona Charge D2.4 72 System ................................................................................................................................................................ Mathematical Modelling of End-of-Life of Power Transformers in Perspective of System D2.5 73 Reliability ........................................................................................................................................................... D2.6 A Comparison between Electroluminescence Models and Experimental Results ............................................. 74 An Improved Pulsed Electroacoustic System for Space Charge Measurement under AC D2.7 75 Conditions........................................................................................................................................................... Authors Index ...................................................................................................................................................... 76 ii UHVnet 2011 Timetable Tuesday 18th January 2011 1700 – 2000 Registration, Reception and Poster Session Wednesday 19th January 2011 0800 – 0830 Registration 0830 – 0845 Welcome by Meeting Chair 0845 – 1015 Session 1: High Voltage Plant 1015 – 1045 Coffee break 1045 – 1215 Session 2: Condition Monitoring 1215 – 1330 Lunch and Poster Session 1330 – 1500 Session 3: Materials 1500 – 1515 Coffee break 1515 – 1645 Session 4: Theory, Methods and Models 1645 – 1700 Closing Remarks iii UHVnet 2011 UHVnet Members Glasgow Caledonian University Prof. Brian Stewart (Chairman) School of Engineering, Science and Design 70 Cowcaddens Road Glasgow, G4 0BA Email: [email protected] University of Southampton Prof. Paul Lewin The Tony Davies High Voltage Laboratory University of Southampton Highfield Southampton, SO17 1BJ Email: [email protected] Cardiff University Prof. Manu Haddad High Voltage Research Group School of Engineering Cardiff University PO Box 925 Cardiff, CF24 0YF Email: [email protected] University of Surrey Prof. Gary Stevens University of Surrey Guildford Surrey, GU2 7XH Email: [email protected] Areva T&D Technology Centre Dr. Fabrice Perrot HV & Electrical Materials Consultancy AREVA T&D Technology Centre St. Leanards Avenue Stafford, ST17 4LX Email: [email protected] University of Strathclyde Dr. Martin Judd Dept. Electronic & Electrical Eng. University of Strathclyde Royal College 204 George Street Glasgow, G1 1XW Email: [email protected] PPA Energy Mr. Cliff Walton 1 Frederick Sanger Road Surrey Research Park Guildford Surrey, GU2 7YD Email: [email protected] University of Leicester Prof. J. Forthergill Department of Engineering University of Leicester University Road Leicester, LE1 7RH Email: [email protected] Narec Mr. Alex Neumann National Renewable Energy Centre Eddie Ferguson House Ridley Street Blyth Northumberland, NE24 3AG Email: [email protected] University of Liverpool Dr. Joe W. Spencer Centre for Intelligent Monitoring Systems University of Liverpool Liverpool, L69 3BX Email: [email protected] University of Manchester Dr. Ian Cotton School of Electronic & Electrical Eng. Room C3, Ferranti Building University of Manchester Manchester, M60 1QD Email: [email protected] National Grid Dr. Jenny Cooper National Grid Warwick Technology Park Gallows Hill Warwick CV34 6DA E-mail: [email protected] iv UHVnet 2011 2011 Organising Committee Colloqium Chair: Treasurer: Technical Chair: Local Arrangements Chair: IOP Dielectrics Group Liaison: Publicity / Media Chair: Registration Chair: Prof. Paul Lewin Prof. Alun Vaughan James Pilgrim Jack Hunter Dr. David Swaffield David Mills Nicky Freebody Email: Email: Email: Email: Email: Email: Email: [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] Session Chairs High Voltage Plant Condition Monitoring Materials Theory, Methods and Models Qi Li Peter Baker David Smith Minan Zhu Martin Reading Abdelghaffar A Abdelmalik Liwei Hao Fabian Moore v University of Manchester University of Strathclyde Glasgow Caledonian University University of Strathclyde University of Southampton University of Leicester University of Southampton University of Cardiff UHVnet 2011 Transport South West Trains service to Weymouth / Poole Departs Winchester 1705 1725 1733 1800 1825 1830 1900 Arrives Southampton Airport 1714 1733 1749 1816 1844 1846 1918 South West Trains service to London Waterloo Departs Winchester 1717 1724 1747 1817 1824 1847 1917 Arrives London Waterloo 1823 1834 1847 1920 1934 1951 2020 Winchester Taxis Wessex Cars Wintax Winchester Taxi Co City Taxis Ltd +44 (0)1962 877749 +44 (0)1962 878727 +44 (0)1962 841212 vi UHVnet 2011 A1 Oral Session 1: High Voltage Plant 0845 – 1015 A1.1 Characterisation of Earth Electrodes including Experimental Tests on Large-Scale Invited 2 Systems .......................................................................................................................................................... Lecture The Variation in Tangential Electric Fields for Different Bushing Shed Profiles in a A1.2 3 Polluted Environment .................................................................................................................................... A1.3 Environmental Friendly Compact High Voltage Substations ........................................................................ 4 A1.4 A Cascaded Flying Capacitor Multilevel Converter for HVDC and FACTS ................................................ 5 A1.5 Thermal Performance of High Voltage Power Cables ................................................................................... 6 Detection and Location of Underground Power Cable using Magnetic Field A1.6 7 Technologies .................................................................................................................................................. 1 UHVnet 2011 A1.1 Characterisation of Earth Electrodes including Experimental Tests on Large-Scale Systems H. Griffiths Cardiff University, UK E-mail: [email protected] Understanding the behaviour of earthing systems under impulse currents is fundamental for the design of new systems and for improving the performance of existing systems under transient conditions. Several analytical approaches and powerful computation tools for the analysis and design of earthing systems have been developed; Analytical calculation methods deal with uniform or layered soil structures only, and there is little experimental evidence to validate the accuracy of these methods. As can be found in the literature, there is a significant set of test results on laboratory scaled-models of earth electrodes as well as some limited data on earthing systems of operational electrical substations and transmission lines. The difficulty with laboratory scaled-models is related to the boundary conditions, which makes them less suitable for the validation of calculation/computation methods. On the other hand, operational electrical installations are located in areas characterised by non-uniform soil structures with both lateral and vertical variations and there is limited scope for high current testing due to system operation and safety constraints. This paper describes experimental investigations carried out on earth electrodes (i) in test cells under laboratory conditions, (ii) at operational electricity substations and , (iii) at two outdoor earthing test facilities; the first at Llanrumney Fields where several different types of earth electrode have been installed, and, the second at a pumped-storage power station in Dinorwig, North Wales, where a test rig has been built on a large water reservoir in order to achieve close to uniform resistivity conditions. The experimental results (DC, AC and impulse) are compared with computed values obtained from different numerical models. 2 UHVnet 2011 A1.2 The Variation in Tangential Electric Fields for Different Bushing Shed Profiles in a Polluted Environment D. J. Smith*1, S. G. McMeekin1, B. G. Stewart1 and P. A. Wallace1 1 Glasgow Caledonian University, UK *E-mail: [email protected] Today, the majority of power transformer high voltage bushings are of the oil impregnated paper (OIP) condenser type. Typically, these bushings have low partial discharge activity, low dielectric losses, and are cost effective [1]. However, over one quarter of transformer failures are as a result of a bushing fault, often an electrical discharge. External surface flashover can occur from excessive surface pollutant on the bushing sheds, and can deteriorate the component, reducing asset reliability. The distribution of the electric field is dependent on the shed profile, and OIP bushings have different profiles specific to the operational environment and manufacturer. During operation, the accumulation of pollutant on the shed surfaces results in uneven distribution of the electric field, and increases the possibility of surface flashover. Using numerical methods, it is possible to model and compare complex bushing geometries under different pollutant conditions. In the literature, numerical models of the potential and electric field distributions for a bushing with varying surface conditions are under investigation [2]} . However, current studies do not consider different shed profiles for similar bushing voltage rating, or the variation in maximum tangential electric field due to the geometry. A finite element method (FEM) numerical model is built for an OIP condenser high voltage power transformer bushing. COMSOL Multiphysics software is utilised using the electric currents application mode, at 50Hz power frequency. The dielectric properties of the material within the model are based on values published in the literature. Using the model, the tangential electric field is evaluated over the bushing creepage path for four shed profiles. For porcelain bushings, surface flashover may occur for a tangential electric field of 5kV/cm [ 3], and this is used as a comparison value. The maximum tangential electric field and its location are determined for each different shed profile for a medium industrial polluted environment. It is expected this work will help develop an improved understanding of the sensitivity and criticality of shed profile selection in polluted environments. Figure 1: Equipotential and tangential electric field distribution over the bushing creepage path. [1] [2] [3] J. S. Graham, "High voltage bushings," in 15th IET International School on High Voltage Engineering and Testing, Newcastle Upon Tyne, UK, 2008, pp. 375–398. P. Cardano et al., "Application of Composite Housing to High Voltage Bushings," CIGRE, Paris, A3-307, 2008. M. Vitelli, "Numerical Performance Analysis of Semiconductor Coatings for Corona Suppression," IEEE Transactions on Dielectrics and Electrical Insulation, vol. 6, no. 6, pp. 774-780, December 1999. 3 UHVnet 2011 A1.3 Environmental Friendly Compact High Voltage Substations M. Albano * 1 , A. Haddad 1 , H. Griffiths 1 , and P. Coventry 2 1 2 Cardiff University, UK National Grid, U.K *E-mail: [email protected] In electric power systems, the preferred gaseous insulating medium is air, and for specific applications such as circuit breaker and very compact substations, sulphur hexafluoride (SF6) is successfully adopted. However, many studies show that the high environmental impact of SF6 due to its green-house effect [1]. In addition, the increase in electricity demand and difficulties for obtaining land for new developments introduce the challenge of building new compact substations with higher voltage ratings on existing sites. Research at Cardiff investigates the possible reduction in size of transmission substations using new airinsulated designs, avoiding to use of large quantity of insulating gas and proposing the substitution of small quantities of SF6 as insulating media in HV apparatus, e.g. in circuit breakers, with the more environmental friendly CF3I gas. The investigations of the adoption of CO2 - CF3I mixture for circuit breaker applications show promising properties. The mixture exhibits high specific heat and high thermal conductivity, and these two properties are the preliminary characteristics for a gas to be adopted in circuit breakers. As a high voltage insulating medium replacement for SF6, the gas mixture 70%-30% (CO2 – CF3I) gives the best performance with optimal gas phase stability to replace SF6 [2]. The application of proposed design and combinations of compact solutions to a standard 400kV switch bay indicates that significant reduction in ground area can be achieved [3]. The introduction of new technologies offer significant additional benefits in the measurement performance and avoidance of risk of explosion compared with equipment filled with oil. The adoption of small quantities of CF3I as insulating medium in HV equipment permits to reduce significantly the potential green-house effects associated with current insulating media. Therefore, the overall footprint of future high voltage substations can be significantly reduced. Figure 1: 400kV substation switch bay footprint for a conventional (on the left) and a compact substation (on the right). [1] Working Group I Contribution to the Fourth Assessment Report of the IPCC - Intergovernmental Panel on Climate Change, Addendum-Errata of Climate Change 2007 - The Physical Science Basis IPCC WG1 AR4 Report, 31 July 2008, web source: http://ipcc-wg1.ucar.edu/wg1/Report/AR4WG1_Errata_2008-08-05.pdf, accessed on 23.11.2009. [2] M. S. Kamarudin, M. Albano, N. Harid, A. Haddad, P. Coventry, "Study on the potential of CF3I gas as alternative for SF6 in high voltage applications", Proceedings of International International Universities‘ Power Engineering Conference –UPEC2010, Cardiff, Wales, UK, 2010. [3] M. Albano, A. Haddad, H. Griffiths, and P. Coventry, "Electric and magnetic fields in 400kV compact Substations", Proceedings of International Symposium on High Voltage Engineering – ISH 2009, Cape Town, South Africa, 24-28 August 2009. 4 UHVnet 2011 A1.4 A Cascaded Flying Capacitor Multilevel Converter for HVDC and FACTS I. B. Efika*1 and L. Zhang2 1 University of Leeds, UK *E-mail: [email protected] Over the last three decades, semiconductor devices like IGBT‘s (insulated gate bipolar transistors) and SIT‘s (static induction transistors) have been developed to reach high switching frequencies at considerable power levels. This opens up completely new research and development areas and fields of applications especially for power transmission and distribution systems such as high voltage direct current (HVDC) and Flexible AC Transmission systems (FACTS). The limits concerning voltage and power can be pushed even further by placing these converter circuits in series. This paper presents a Cascade Flying Capacitor Multilevel Converter (CFCMC). The basic unit in this system is a 3-level full bridge flying capacitor converter. Each block switches between 0 to +Vdc and 0 to Vdc at any point. One leg of a three phase CFCMC system can be formed as shown in Figure 1 by cascading these units. This circuit has some advantages when compared to the cascaded H-Bridge converter; the waveform performance is better due to the additional voltage level control, the stress upon switches is also lower. The presence of capacitors in each unit may seem to have cost implications, however each unit can handle higher power ratings therefore fewer units are required. The size of the capacitors required is also small when compared to those required for a standard flying capacitor multilevel converter. The particular merit of CFCMC also lies on its ability of supplying reactive power due to additional capacitors in each unit. This will improve system stability margin especially when many large renewable sourced generators with low reactive power output are connected to the utility network. Future research will be to employ the estimation and reduction of switching losses. Phase Shifted Sinetriangle PWM scheme [1], is used to generate the switching signals for each individual block and the entire cascaded system. The phase shift angle of the carrier signal, calculated as 2 can be adjusted to l control the stress impact on the switches in flying capacitor block. This feature is important when the effect of switching losses is considered. Find below the simulation results. (fS =540Hz), THD = 3.27%. Figure 1: Switch mode configuration; (Waveforms Top: 1 – Line to ground Voltages, 2 – Line to Neutral Voltages, 3 – Line to Line voltages, 4 – Current Waveforms 4 – FFt Analysis) [1] L. Yiqiao and C.O. Nwankpa, ―A Power-Line Conditioner Based on Flying-Capacitor Multilevel VoltageSource Converter with Phase-Shift SPWM‖, IEEE Transactions on Industry Applications, Vol. 36, No. 4, July/August 2000. 5 UHVnet 2011 A1.5 Thermal Performance of High Voltage Power Cables J. A. Pilgrim1, D. J. Swaffield1, P. L. Lewin1 and D. Payne2 1 2 University of Southampton, UK National Grid, UK *E-mail: [email protected] The UK high voltage electricity transmission network continues to face annual rises in demand, with ever greater volumes of power supplied to load centres throughout the country. To operate this network effectively, it is vital to accurately calculate the maximum allowable electric current which can be safely carried by each component in the power system. In high voltage power cables, this limit is defined by the maximum operating temperature of the cable insulation. Specify this current rating to be too low and the cable asset will never be used to its full potential; conversely setting the rating to be too high risks damage to the asset as the excessive heating can cause premature failure. Thus the rating calculation must be optimised to maintain security of supply by minimising the risk of cable failure, while also maximising the returns from capital investment on the power network. This project has employed a variety of mathematical techniques to improve the methods by which current ratings are calculated. Modern computational techniques such as finite element analysis (e.g Figure 1) and computational fluid dynamics are used to create more advanced circuit rating techniques. These have been compared and refined with input gained from field data. By eliminating simplifications from existing methods, it has been possible to identify ways of increasing the utilisation of the existing network. In addition the new techniques allow examination of the potential benefits of future developments in cable technology. Benefits are being derived from this work on both a day to day and strategic planning levels. For instance, by re-evaluating the current rating method for cables installed in tunnels, it has proved possible to consider the benefits from co-locating more cables in one tunnel to best use these expensive assets. The application of this method has allowed the quantification of the benefits which might be available from next generation cable technologies, enabling the prioritisation of future research effort in cable materials. Upon completion, the knowledge gained from this work is to be used to revise the international standard on calculating current ratings in cable tunnels. Techniques such as these underpin the concept of smart grids with improved operational flexibility and capability. Simultaneously the requirement to build expensive new components into the network is limited, whilst still meeting the need to supply ever increasing volumes of power across the country. Figure 1: Contours of air velocity within a cable tunnel 6 UHVnet 2011 A1.6 Detection and Location of Underground Power Cable using Magnetic Field Technologies P. Wang1, K. F. Goddard1, P. L. Lewin1 and S. G. Swingler1 1 University of Southampton, UK E-mail: [email protected] The location of buried underground electricity cables is becoming a major engineering and social issue worldwide. Records of utility locations are relatively scant, and even when records are available, they almost always refer to positions relative to ground-level physical features that may no longer exist or that may have been moved or altered. The lack of accurate positioning records of existing services can cause engineering and construction delays and safety hazards when new construction, repairs, or upgrades are necessary. Hitting unknown underground obstructions has the potential to cause property damage, injuries and, even deaths. Thus, before commencing excavation or other work where power or other cables may be buried, it is important to determine the location of the cables to ensure that they are not damaged during the work. This paper describes the use of an array of passive magnetic sensors (induction coils) together with signal processing techniques to detect and locate underground power cables. The array consists of seven identical coils mounted on a support frame; one of these coils was previously tested under laboratory conditions, and relevant results have been published in [1]. A measurement system was constructed that uses a battery powered data acquisition system with two NI 9239 modules connected to the coil array, and controlled by a laptop. The system is designed to measure the magnetic field of an underground power cable at a number of points above the ground. A 3 by 3 m test area was chosen in one of our campus car parks. This area was chosen because the university‘s utility map shows an isolated power cable there. Measurements were taken with the array in 16 different test positions, and compared with the values predicted for a long straight horizontal cable at various positions. Finally, error maps were plotted for different Z-coordinate values, showing the minimum fitting error for each position in this plane. One such map is shown in Figure 1; the low error values of 4-5% give a high degree of confidence that most of the measured signal is due to a cable near to these positions. This view is supported by the fact that the university‘s utility map shows the cable at X = 1.4 m, and by amplitude measurements taken with a hand-held magnetic field meter. 0.14 -0.2 0.12 -0.4 Depth (m) -0.6 0.1 -0.8 0.08 -1 -1.2 0.06 -1.4 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 0.04 Horizontal distance, X (m) Figure 1: a typical result for car park tests at Z = 2m. (possible cable position is X=1.4 m and Depth = 0.6m. 96% accuracy) P. Wang, P. Lewin, K. Goddard, and S. Swingler, ―Design and testing of an induction coil for measuring the magnetic fields of underground power cable‖, IEEE International Symposium on Electrical Insulation, San Diego, California, USA 2010 7 UHVnet 2011 B1 Oral Session 2: Condition Monitoring 1045 - 1215 B1.1 Invited Condition Monitoring for High Voltage Equipment ...................................................................................... 9 Lecture Automated Phase-agnostic Time Domain Analysis of RF Partial Discharge Pulse Data B1.2 10 for Low-power Wireless Sensing Applications................................................................................................... B1.3 Partial Discharge Analysis of Defective Three-phase Cable .............................................................................. 11 B1.4 Optimum Coil Design for Inductive Energy Harvesting in Substations ............................................................. 12 Study of Signal Processing Techniques used for Denoising Partial Discharge Signals in B1.5 13 Cables .................................................................................................................................................................. B1.6 Instrumentation and Condition Monitoring of Composite Cross Arms .............................................................. 14 8 UHVnet 2011 B1.1 Condition Monitoring for High Voltage Equipment Martin D Judd University of Strathclyde , UK E-mail: [email protected] Advances in electronic technology have yielded impressive processing power, increased bandwidth and low power consumption, opening up a myriad of possibilities for new, specialised condition monitoring systems. Many businesses have sprung up to exploit these opportunities by bringing new condition monitoring equipment to market. After initial enthusiasm, some techniques have not proved successful, or their effectiveness has turned out to be variable, depending on the expertise of the system supplier or the end user. This has led to the evolution of a more cautious approach to new condition monitoring techniques, which must be properly scrutinised by those who would adopt them. This presentation will begin by addressing some of the issues that researchers should take into account when developing new technologies for deployment on high voltage equipment in the context of power transmission and distribution. Some accounts of experience with the development of commercial monitoring technologies will be given to illustrate the key role that sustained fundamental scientific research can play in ensuring that, where possible, research activities lead on to real industrial and commercial benefits. Examples will be drawn from the author‘s past experience in the field of partial discharge monitoring and from current work on autonomous wireless sensor nodes for use in substations. 9 UHVnet 2011 B1.2 Automated Phase-agnostic Time Domain Analysis of RF Partial Discharge Pulse Data for Low-power Wireless Sensing Applications P. C. Baker*1, A. J. Mair1, M. D. Judd1 and S. D. J. McArthur1 1 University of Strathclyde, UK *E-mail: [email protected] The use of partial discharge (PD) diagnostics on substation equipment is limited by operational and financial constraints involved in deploying sensing systems at the plant level. To circumvent these constraints, modern wireless sensor network technology can provide a robust and flexible architecture upon which condition monitoring applications can be built, without the need for extensive wiring and at a much lower cost than conventional wired systems. A low-power approach to PD diagnostics using this technology has previously been demonstrated in [1] . One key constraint on the deployment of wireless PD sensors is the availability of an absolute phase reference, which is a prerequisite for conventional phase resolved PD pattern analysis. Phase resolution requires a VT connection to the electrical phase under observation, which for wireless sensing applications is impractical, as cabling must be kept to a minimum. Therefore, an alternative approach is ideally required which does not rely on an absolute phase reference. To address this, a novel PD diagnostic technique is proposed that is capable of ―phase-agnostic‖ timedomain analysis of RF PD pulse data, specifically suited to operation within the resource constraints of a wireless sensor node. Derived from an automated diagnostic technique for identification of faults in shipboard power systems [2] this method analyses relative pulse magnitude, pulse density and inter-pulse distance for a travelling window of PD pulse measurements. Using laboratory data representing typical PD patterns for defects in GIS, machine learning algorithms were trained on random data samples with their accuracy tested on previously unseen data. The split between training and test data was 50/50 and the time window considered was well below 1 second. The machine learning algorithms tested were: a C4.5 decision tree, a Naïve Bayesian network, a Support Vector Machine, a Nearest neighbour classifier, and Radial Basis Function algorithm. It was determined that out of these algorithms the C4.5 produced the most accurate classifier for this data type. Testing different window sizes established that training on a window size of 4 electrical cycles generated the most accurate classifier. An initial study has demonstrated that the diagnostic technique, upon identifying the presence of PD, gives promising results across seven defect types. Future work will use a wider training set to train the classifier, ideally leading to a more robust and generalised classifier. The classifier will then be deployed on a wireless sensor node connected to a low-power PD detector and validated in laboratory tests. [1] P. C. Baker, ―Enhancing substation condition monitoring through integrated diagnostics, wireless sensor networks and multi-agent systems‖, PhD Thesis, University of Strathclyde, September 2010 [2] Mair, A.J.; Davidson, E.M.; McArthur, S.D.J.; Srivastava, S.K.; Schoder, K.; Cartes, D.A.; ―Machine learning techniques for diagnosing and locating faults through the automated monitoring of power electronic components in shipboard power systems,‖ Electric Ship Technologies Symposium, 2009. ESTS 2009, pp.469-476, 20-22 April 2009 10 UHVnet 2011 B1.3 Partial Discharge Analysis of Defective Three-phase Cable J. A. Hunter*1, L. Hao1, D. J. Swaffield1, P. L. Lewin1, N. Cornish2, C. Walton2 and M. Michel3 1 2 University of Southampton, UK PPA Energy, UK 3 UK Power Networks *E-mail: [email protected] Power distribution cable networks represent a dynamic and complex challenge with regard to the issues of maintenance and providing a reliable, high quality supply of electrical power. Utilities historically used regular off-line testing to investigate the health of their assets. This method of testing is reasonably effective for this purpose but does have certain drawbacks associated with it; customer supply can be interrupted during the testing process and the cables are generally not tested under normal operating conditions. Meaning that the test data is not representative of the Partial discharge (PD) activity that is apparent under on-line conditions and the testing activity itself could trigger previously dormant PD sources. The modern approach for understanding the health of medium voltage (MV) cable distribution networks is to continuously monitor the assets whilst on-line. Analysis if the field data is then used to inform decisions regarding asset replacement and maintenance strategies. PD activity is widely recognised as a symptom linked to the degradation of the dielectric properties of high voltage plant. UK Power Networks sponsored research is being undertaken to investigate the evolution of PD activity within three-phase paper insulated lead covered (PILC) cables containing introduced defects. An experiment has been designed to stress cable lengths in a manner that is representative of the conditions met by on-line circuits [1]. A cable section containing a defect that is known to lead to the premature failure of in-service cables has been PD tested over a range of operating temperatures. The experiment utilizes three-phase energization at rated voltage as well as thermal cycling of the cable to replicate the daily load pattern experienced by circuits in the field. The extension to this work involves PD testing cable samples containing a range of defects to produce a data set consisting of PD pulses produced by varied sources. Analysis of this data should lead to a better understanding of the signals produced by the premature ageing of these types of cable. Figure 1: 3PARD plot showing produced by cable sample at 37 ºC. a: PD activity generated by the defect. [1] J. A Hunter, L. Hao, D. J. Swaffield, P. L. Lewin, N. Cornish, C. Walton and M. Michel, ―Partial discharge in medium voltage three-phase cables‖, Conference record of the 2010 IEEE International Symposium on Electrical Insulation, 2010 11 UHVnet 2011 B1.4 Optimum Coil Design for Inductive Energy Harvesting in Substations N. M. Roscoe*1 and M. D. Judd1 1 University of Strathclyde , UK *E-mail: [email protected] The reliability and life expectancy of electrical supply equipment can be improved through the use of condition monitoring. However, expansion of existing condition monitoring through the addition of new sensors is challenging since power to condition monitoring sensors has traditionally been supplied by mains power, which is not available in many locations where monitoring would be useful, or by batteries, which require their own maintenance regime. Energy harvesting in substations has the potential to power a new class of ―fit-and-forget‖ wireless sensors, thus enabling more affordable expansion of condition monitoring. There are many potential sources of energy in substations (solar, wind, thermal, etc.), each of which may have a role in a particular range of sensor applications. This paper is concerned with inductive energy harvesting from the ambient magnetic fields, focussing on optimal coil design. By this means, the intention is to develop a class of ―free-standing‖ inductive energy harvesting devices, which can be placed at a safe distance from high voltage conductors, as defined in Figure 1. In this paper, all aspects of coil design are considered, including core material, core geometry and number of turns. A coil is then designed for a representative application, taking practical limitations into account, and its output power is characterised while harvesting from a uniform 50 Hz magnetic field generated within a set of Maxwell coils. Experimental results are presented and discussed. Figure 1: (a) ―Threaded‖ inductive energy harvesting. (b) ―Free-standing‖ inductive energy harvesting. 12 UHVnet 2011 B1.5 Study of Signal Processing Techniques used for Denoising Partial Discharge Signals in Cables F.P.Mohamed*1, W.H.Siew1, J.J. Soraghan1, S.S.Strachan1 1 University of Strathclyde, UK *E-mail: [email protected] Ageing of cable insulation is an increasing problem which necessitates the development of on-line condition monitoring for cables. On-site Partial discharge (PD) measurements have been widely considered to be the most effective diagnostic tool for insulation assessment. As a result of PD, fast varying current pulses flow in the cable conductors and these pulses can be detected using high frequency current transformers which are working in the VHF band. However on-site measurements are often hampered by various noise sources which makes the process of extracting useful information from the raw data more difficult. Presented here is a review of the various signal processing techniques used for extracting PD information from the corrupted data with special focus on the advantages and disadvantages of each. Signal averaging is the most common method for removing random noise from the raw data. This can be done by alighning the various data sets and then added together. Moving average filter, ensemble average filters are widely used to remove random noises in the signal. Over smoothing of the signals, by this method, results in loss of sharp edges in the PD signal which is not recommended for PD mapping. Wavelet transform is the most widely accepted tool for denoising on-site data which involves selecting a wavelet basis function followed by decomposition, threshold selection and reconstruction [1]. Basis function can be scaled and translated to match the desired signal information to be extracted from the raw data. Selection of proper basis function makes the denoising more efficient which is often more difficult since the PD signal undergoes dispersion in the cable. Debauchees‘ family wavelets are most widely used by researchers to extract PD from the raw data. Both discrete wavelet transform (DWT) and stationary wavelet transform (SWT) are used. Due to the dyadic decomposition in DWT, shift invariance property is lost which will introduce uncertainties in PD mapping. SWT is most suitable method for locating partial discharges due to shift invariance. However processing time required by SWT is high. Recently second generation wavelet transform (SGWT) which is used for the construction of wavelets has been applied to denoise the on-ste PD data. SGWT is a fast way of implementing wavelet transform which makes the processing time low. Also using the prior information about the PD signature, filter banks can be modified which makes denoising more efficient than classical wavelet based techniques [2]. Wavelets techniques discussed so far require prerequisites like choice of wavelet basis function and decomposition level to denoise the data. Hence novel techniques of signal processing which does not require any prerequisites are being investigated to compare its performance with respect to wavelet based denoising. Various signal processing tools discussed above were applied to on-site PD data and the results were compared [1] Hao Zhang, T.R. Blackburn, B.T. Phung and D. SenA Novel Wavelet Transform Technique for On-line Partial Discharge Measurements Part 1: WT De-noising Algorithm, IEEE Transactions on Dielectrics and Electrical Insulation Vol. 14, No. 1; February 2007 [2] Xiaodi Song, Chengke Zhou, Donald M. Hepburn, Guobin Zhang ,Second Generation Wavelet Transform for Data Denoising in PD Measurement. IEEE Transactions on Dielectrics and Electrical Insulation Vol. 14, No. 6; December 2007 13 UHVnet 2011 B1.6 Instrumentation and Condition Monitoring of Composite Cross Arms C. A. Veerappan*1, C. Zachariades1, S. M. Rowland1, I. Cotton1, P. R. Green1, F. Allison1 1 University of Manchester, UK E-mail: [email protected] * Providing an alternative solution to glass and porcelain insulators, composite insulators exhibit enhanced performance under polluted conditions. Tangentially, while the idea of composite cross-arms is not new, they have only seen commercialisation on lower voltage towers, where structural requirements permit standard engineering solutions. Over the last few years a team from The National Grid High Voltage Research Centre at The University of Manchester and an industrial partner, EPL Composite Solutions Ltd have developed a composite cross-arm platform aiming to increase power transfer capacities of existing transmission infrastructure. This paper presents an instrumented field-trial of the composite cross-arm. The trial consisted of installing composite cross-arms on four adjacent towers on a line being decommissioned in the Scottish Highlands. This unenergised trial will examine mechanical performance, particularly with respect to ice, snow and wind. The remote location and the area's designation as a Site of Special Scientific Interest limited certain aspects of the system's design; having a big impact on power sourcing and data retrieval, where simple but reliable wire-line methods could not be employed. Power is generated by wind turbines and employs a battery storage system. The use of conventional communication networks for data recovery was not viable; bespoke systems for data storage and periodic physical collection were required. The premises of a local a ski centre were used for data collection and wireless networking equipment was used to bridge the gaps to and between the towers. Mechanical performance is to be measured through the use of embedded strain gauges and a combination of a load cell, accelerometer and inclinometer at the cross-arm nose. An industrial data capture and control platform is used to capture sensor outputs and store them until retrieval. Network cameras with local storage capabilities are used to capture normal and time-lapse videos of the installed cross-arms. Ultimately this trial has become more about instrumentation and data transfer in hostile environments, rather than cross-arm performance. After the first winter a second generation system is expected to be implemented. Accumulated data will be used to optimise the cross-arm design. Figure 1, shows the installation site. Figure 1: Picture of Installation Site, depicting, relay station, one non-instrumented tower and the first instrumented tower 14 UHVnet 2011 C1 Oral Session 3: Materials 1330 - 1500 C1.1 Invited Towards Recyclable Insulation Materials for High Voltage Cables ....................................................... 16 Lecture C1.2 Five-Electrode Inclined-plane Tests of Textured Silicone Rubber Samples ................................................. 17 C1.3 A Raman Microprobe Study of Corona Ageing in a Controlled Atmosphere ............................................... 18 FTIR Spectrum of Layered PET Insulation with Artificial Voids Subjected to Electrical C1.4 19 Stressing ......................................................................................................................................................... C1.5 Dielectric Behaviour of Alkyl Esters of Seed-based Oil................................................................................ 20 C1.6 The Influence of Spherical Cavity Surface Charge Distribution on PD Events ............................................ 21 15 UHVnet 2011 C1.1 Towards Recyclable Insulation Materials for High Voltage Cables I. L. Hosier*, A. S. Vaughan and S. G. Swingler University of Southampto n, UK *E-mail: [email protected] The preferred material for modern extruded high voltage transmission cables is cross-linked polyethylene (XLPE). This material has excellent thermo-mechanical and dielectric properties, however it is not easily recycled at end of use, raising questions as to its long term sustainability [1]. Therefore research work at Southampton has sought to identify suitable recyclable alternatives to XLPE. Such candidate materials need to have low temperature flexibility and high temperature mechanical stability combined with a sufficiently high electrical breakdown strength. Initially ethylene based systems [2] were considered, however, low density polyethylene (LDPE) has poor mechanical stability at temperatures exceeding 80 oC whereas high density polyethylene (HDPE) is too brittle at low temperatures. To overcome these difficulties, a series of blends combining either an ethylene vinyl acetate (EVA) co-polymer or a low density polyethylene (LDPE) with a high density polyethylene (HDPE) were considered. A blend of 20 % HDPE in LDPE crystallised relatively rapidly (Figure 1a), was found to offer a good balance between high temperature mechanical stability and flexibility at low temperatures (Figure 1b) combined with excellent dielectric strength. In the remaining EVA based blends, increasing the vinyl acetate content resulted in a more rubbery composite but with a reduced high temperature stability and breakdown strength. Propylene based systems were then considered, these included traditional syndiotactic (sPP) and isotactic polypropylene (iPP) and a range of propylene-ethylene co-polymers. Such systems offered enhanced high temperature stability and with sufficient ethylene content, low temperature flexibility [3], combined with good dielectric breakdown strength provided that the crystallisation was rapid enough to avoid the formation of large spherulites. In further efforts to optimise the properties, two blend systems composed of iPP mixed with either a propylene ethylene co-polymer (with 40 % ethylene content; ―PE40‖) or sPP were considered. Provided that the crystallisation was relatively rapid, both blends provided excellent dielectric performance and high temperature stability. A blend of 20 % iPP in PE40 (Figure 1c) offers the best level of mechanical flexibility at low temperatures and would therefore be suitable for the manufacture of enhanced, recyclable high voltage cables. Figure 1: (a) Morphology of the optimised LDPE/HDPE blend system (b) mechanical properties of selected ethylene based systems (c) mechanical properties of iPP, PE40 and its blends [1] C. P. Martin, ―The impact of mechanical stress on the integrity of XLPE cables‖, PhD Thesis, University of Southampton, 2004. [2] I. L. Hosier, A. S. Vaughan and S. G. Swingler, ―An investigation of the potential of ethylene vinyl acetate/polyethylene blends for use in recyclable high voltage cable insulation systems‖, , J. Mat. Sci., vol. 45, no. 10, pp. 2747-2459, 2010. [3] I. L. Hosier, L. Cozzarini, A. S. Vaughan and S. G. Swingler, ―Propylene based systems for high voltage cable insulation applications‖, J. Phys.: Conf. ser., vol. 183, 012015, 2009. 16 UHVnet 2011 C1.2 Five-Electrode Inclined-plane Tests of Textured Silicone Rubber Samples P. Charalampidis*1, A. Haddad1, R. T. Waters1, H. Griffiths1, N. Harid1 and P. Sarkar2 1 2 Cardiff University, UK National Grid, UK *E-mail: [email protected] The shank region of polluted polymeric insulators is susceptible to thermal damage where high current density and high electric field magnitude occurs. The recently proposed [1] textured design consisting of hemispherical protuberances, aims to reduce the current density and the electric field gradient and increase the longitudinal creepage length without increasing the overall length of the insulator. Moreover, the formation of parallel current paths can lead to less harmful discharges. For the purpose of having a more detailed insight of the formation of these parallel current paths, the IEC-60587 standard [2] that deals with the inclined plane test was modified by replacing the single ground electrode by five smaller separate electrodes. The test results of this modified inclined plane test of textured and conventional untextured rectangular samples are presented along with an analysis of the current distribution and the discharge activity. The improved performance of textured samples, in terms of resistance and tracking, has been excellent [3]. Figure 1: Evidence of thermal damage after testing on untextured and textured samples. [1] R. T. Waters and A. Haddad, ―Insulating Structures‖, UK Patent GB2406225B, Dec 2006. [2] IEC 60587: 1984:, ―Methods for evaluating resistance to tracking and erosion of electrical insulating materials used under severe ambient conditions‖, IEC standard. [3] P.Sarkar et al., ―Inclined Plane Tests of Textured Silicone Rubber Samples‖, 2010 International Conference on High Voltage Engineering and Application, New Orleans, U.S.A., 11-14 October 2010 17 UHVnet 2011 C1.3 A Raman Microprobe Study of Corona Ageing in a Controlled Atmosphere N. A. Freebody 1, A. S. Vaughan 1 1 University of Southampton, UK *E-mail: [email protected] Raman microprobe spectroscopy is widely used in the analysis of polymers due to its high spatial resolution and its ability to characterise the exact chemical composition of a material and, for this reason, it can be applied to study electrical ageing in solid dielectrics. For example, it enables us to probe the chemical processes involved in electrical treeing, whereby solid polymer is converted into decomposition products through a number of electrical processes [1]. This study takes a novel approach to this problem through ex-situ experiments that seek to reproduce the chemistry of electrical treeing in bulk. Plaque specimens of a range of polymers, including polyethylene, polystyrene, PEEK and silicone rubber, were subjected to surface ageing via corona discharge, and the residual products on both the sample surface and the high voltage electrode (as seen in figure 1) were characterised by Raman microprobe spectroscopy. These experiments were performed as a function of applied voltage, electrode geometry etc both in air and within a closed cell that enabled the atmosphere to be controlled and adjusted. The resulting Raman fingerprints were compared with those previously identified within electrical trees [2,3]. After corona discharge was applied to the samples, despite a large change in surface texture, no residues were seen on the sample and few chemical changes were detected via Raman spectroscopy, thus implying that erosion of the material occurs by fragmentation of the polymer. Analysis of the electrodes aged in air and nitrogen, revealed varying evidence of sp2 hybridized carbon, and fluorescence, both of which are products previously associated with the processes involved in electrical treeing. The similarity in these results and previous published works [2,3] suggest that there are common processes involved between corona surface ageing and electrical treeing especially when an inert atmosphere is used. Figure 1: Optical image of deposit found on electrode of aged Si rubber (scale bar = 10µm). [1] A.S. Vaughan, S.J. Dodd, and S.J. Sutton, ―A Raman microprobe study of electrical treeing in polyethylene‖. J. Matter. Sci. 39(1): p. 181-191, 2004. [2] X.S. Liu, A.S. Vaughan, and G. Chen, ―A Raman spectroscopic study of bulk and surface ageing phenomena in polyethylene‖. Annual Report Conference on Electrical Insulation and Dielectric Phenomena: p. 145-148, 2003. [3] A.S. Vaughan, I.L. Hosier, S.J. Dodd, S.J. Sutton, ―On the structure and chemistry of electrical trees in polyethylene‖. J. Phys. D-App. Phys. vol 39(5): pp. 962-978, 2006. 18 UHVnet 2011 C1.4 FTIR Spectrum of Layered PET Insulation with Artificial Voids Subjected to Electrical Stressing D. Adhikari*1, D. M. Hepburn1 and B. G. Stewart1 1 Glasgow Caledonian University, UK *E-mail: [email protected] Partial Discharge (PD) results from faults, such as voids, in power insulation systems and exacerbates failure. It is difficult to completely eliminate voids in polymeric materials. These may be formed during the manufacturing process by air leakage into the mould or by evaporation of volatile decomposition products. Sometimes insufficient pressure on a liquid epoxy during curing causes a cavity to develop. Voids in polymer insulation are gas filled cavities which have a much lower permittivity and lower breakdown strength than the solid insulation material. The lower permittivity gives rise to a higher electric field in the cavity, as a result of which the gas in the cavity will generally break down as the applied voltage is raised. The samples used in this work are created from layers of poly-ethylene-terephthalate (PET) which are 50μm thick. Multiple layers of polymer are set between a plane electrode connected to ground and a second plane electrode connected to a high voltage AC source. Artificial cylindrical voids are created by removing circular sections from one layer of polymer. Samples with single void and multiple voids (placed both horizontally and vertically) were created. These samples are electrically stressed and PD activity is monitored and recorded [1]. To determine the changes to the surface of the voids as a result of the chemical processes produced by PD, the FTIR spectrum of the polymer specimens subjected to PD were examined. The FTIR spectra from samples of unstressed material, and from surfaces of single void and multiple voids were compared and analysed. The measurement of FTIR spectrum was conducted by Attenuated Total Reflection (ATR) method, which is commonly used for the chemical analysis of the surface layer of specimens. As ATR is surface specific, depth of investigation being ~5 microns, this allows investigation of initial processes in material degradation. The changes in the spectra shown below indicate changes to the carbonhydrogen bonds (~3000cm-1) and production of carbon-oxygen (~1700cm-1) bonds. The paper will discuss changes to surface chemistry in stressed samples and correlate change with stress applied. Figure 1: Comparison of the FTIR Spectra of Unstressed PET and Stressed PET with artificial voids [1] D. Adhikari, D. M. Hepburn and B. G. Stewart, ―Analysis of Partial Discharge Characteristics in Artificially Created Voids‖, 45th International Universities’ Power Engineering Conference, Cardiff, Wales, UK 2010 [2] L. A. Dissado and J. C. Fothergill, ―Electrical Degradation and Breakdown in Polymers‖, Peter Peregrinus Ltd., pp.292-310, 1992 [3] P. Hyvonen, ―Prediction of Insulation Degradation of Distribution Power Cables Based on Chemical Analysis and Electrical Measurements‖, Doctoral Dissertation – Helsinki University of Technology, 2008 19 UHVnet 2011 C1.5 Dielectric Behaviour of Alkyl Esters of Seed-based Oil A. A. Abdelmalik*1, J. C. Fothergill1, S. Dodd1 1 University of Leicester, UK * E-mail: [email protected] The dielectric response of palm kernel oil alkyl esters of straight and branched carbon chain were studied over the frequency range of 10-3 to 104 Hz and temperature range of 20 ˚C to 80 ˚C. The straight chain alkyl ester (PKOME) was synthesized through esterification reaction of laboratory purified palm kernel oil with methanol in the presence of catalyst. It was then followed by epoxidation reaction involving the alkyl ester product and insitu peracid in the presence of catalyst to synthesize the corresponding epoxy alkyl ester. The epoxy alkyl ester was then reacted with propionic anhydride in the presence of catalyst under nitrogen. This reaction opens the epoxy ring of the epoxy alkyl oleate component of the epoxy alkyl ester to attach side chains for the synthesis of a side-branched alkyl ester (PropPKOAE). The dielectric response of these materials shows that the relaxation processes corresponds to ionic conduction and electrode polarization phenomena. The real part is constant and the imaginary part is inversely proportional to frequency up till about 10-1 Hz. This is symptomatic of a constant capacitance in parallel with conductance. Below 10-1 Hz, the imaginary part of the relative permittivity maintains a slope of -1, whilst the real part of PKOME and PropPKOAE acquired an average slope of -1.2 and -1.5 within the temperature range studied. This suggests that grafting of side chain alters the dynamics of the adsorbed ions at the electrode-liquid interface. The complex impedance plot which separates the interfacial effect from the bulk has a discontinuity at frequencies that correspond to tan δ peak. This peak corresponds to the relaxation frequency of electrode polarization, fEP. The estimated fEP of PKOME and PropPKOAE at 20 ˚C are 7.1 × 10-3 Hz and 3.7 × 10-2 Hz respectively. This implies that the relaxation time for electrode polarization in PropPKOAE is lower than that of PKOME, suggesting a faster dynamics of ions involved in electrode polarization in branched alkyl ester of vegetable oil. The change in the dynamics of electrical double layer formation may be responsible for increase in conductivity of PropPKOAE. Temperature increase lead to a shift in fEP towards high frequency with a corresponding increase in AC conductivity, and thinner diffuse electrical double layer. 5 10 PKOME Real Part PKOME Imag Part PropPKOAE Real Part PropPKOAE Imag Part 4 10 3 Relative Permittivity 10 2 10 1 10 0 10 -1 10 -2 10 -3 10 -3 10 -2 10 -1 10 0 1 10 10 Frequency (Hz) 2 10 3 10 4 10 Figure 1: Relative Permittivity of PKO Alkyl esters at 20 ˚C [1] A.K. Jonscher, 1996, Universal Relaxation Law, Chelsea Dielectrics Press, London. [2] A.K. Jonscher, 1983, Dielectric relaxation in solids, Chelsea Dielectric Press, London. [3] R.J. Sengwa, S. Choudhary, S. Sankhla, 2008, Low frequency dielectric relaxation processes and ionic conductivity of montmorillonite clay nanoparticles colloidal, suspension in poly(vinyl pyrrolidone)-ethylene glycol blends, eXPRESS Polymer Letters, Vol. 2, No. 11, Pp. 800–809. 20 UHVnet 2011 C1.6 The Influence of Spherical Cavity Surface Charge Distribution on PD Events H. A. Illias*1, G. Chen1 and P. L. Lewin1 1 University of Southampton, UK *E-mail: [email protected] Modelling of partial discharge (PD) events allows a better understanding of the phenomena itself. In this work, an improved model representing PD behaviour within a spherical cavity in a homogeneous dielectric material has been developed to study the influence of cavity surface charge distribution on the electric field distribution in the cavity and the material. Comparison of measurement and simulation results has been undertaken to validate the model. The model uses a two-dimensional (2D) axial symmetric Finite Element Analysis (FEA) method, which is solved for local electric potentials. Figure 1 shows the model geometry, which consists of a homogenous dielectric material (2.0 mm thickness, 5 mm radius) and a hemispherical cavity (1.4 mm diameter). The upper and lower cavity surfaces are divided into 10 boundaries each. A sinusoidal voltage is applied to the upper electrode while the lower electrode is always grounded. Discharge is assumed to occur along the symmetry axis in the cavity. Once the discharge has passed through the cavity to the opposite surface, it is assumed that charge then propagates along the cavity wall [1, 2]. The charge propagation is assumed only on the first 2 boundaries from the symmetry axis of the upper and lower cavity surfaces. Charge distribution is assumed identical on the upper and lower cavity surface. During discharge, charge density increases on the cavity surface boundaries where charge propagates, until the electric field in the cavity centre is less than the extinction field. To model the charge movement along the cavity surface through conduction at other times, the change in the charge density on each cavity surface boundary is set as dependent on the cavity surface conductivity and the electric field on each boundary. Thus, surface charge distribution will become non-uniform, influencing the electric field distribution in the cavity and the material, affecting the likelihood of the next PD event. The simulation result from the model agrees with a range of measurement results for a 50 Hz, 16 kV ac applied sinusoidal voltage (Figure 1b). Therefore, the modelling of PD events by consideration of charge distribution on the cavity surface is reasonable. (a) (b) Figure 1: (a) 2D axial symmetric model geometry and (b) measurement and simulation results for 500 applied voltage cycles [1] W. Kai, S. Yasuo and L. A. Dissado, ―The contribution of discharge area variation to partial discharge patterns in disc-voids‖, Journal of Physics D: Applied Physics, vol. 37, pp. 1815-1823, 2004 [2] W. Kai, T. Okamoto and Y. Suzuoki, "Effects of discharge area and surface conductivity on partial discharge behavior in voids under square voltages," IEEE Transactions on Dielectrics and Electrical Insulation, vol. 14, pp. 461-470, 2007 21 UHVnet 2011 C1.6 Oral Session 4: Theories, Methods and Models 1515 - 1645 D1.1 Invited Stochastic and Deterministic models for Electrical Tree Growth ................................................................ 23 Lecture Model of Structural Damage to Carbon Fibre Composites Due to Thermo-electric D1.2 24 Effects of Lightning Strikes ................................................................................................................................ D1.3 Switching Ferroresonant Transient Study using Finite Element Transformer Model ......................................... 25 D1.4 Transient Modelling of Offshore Wind Farm Connections ................................................................................ 26 D1.5 Surface Gradient Calculation for Overhead Lines .............................................................................................. 27 Modelling of Electroluminescence in Polymers Using a Bipolar Charge Transport D1.6 28 Model .................................................................................................................................................................. 22 UHVnet 2011 D1.1 Stochastic and Deterministic models for Electrical Tree Growth S. Dodd 1 University of Leicester, UK E-mail: [email protected] Electrical treeing is a long term electrical degradation process in polymeric insulating materials which can lead to early failure of HV electrical equipment. A number of approaches have been taken in the literature for the computer simulation of electrical tree growth. These are usually based on either stochastic or deterministic principles, see figure 1, for tree growth extension (or combinations of these two approaches). In this presentation, the physics on which these two methodologies are based are introduced and discussed in terms of their appropriateness in representing the electrical tree growth mechanisms driven by partial discharge activity. The presentation will conclude with a discussion on what can be learnt from the different approaches which contribute to our understanding the electrical tree growth mechanisms. Figure 1. (a) Stochastic and (b) deterministic approach for the computer simulation of electrical tree growth. 23 UHVnet 2011 D1.2 Model of Structural Damage to Carbon Fibre Composites Due to Thermo-electric Effects of Lightning Strikes R. D. Chippendale*1, I. O. Golosnoy1, P. L. Lewin1, G.S. Murugan1, J Lambert1 1 University of Southampton, UK *E-mail: [email protected] The impact of a lightning strike causes a short high electrical current burst through Carbon Fibre Composites (CFC). Due to the electrical properties of CFC the large current leads to a rapid heating of the surrounding impact area which degrades and damages the CFC. It is therefore necessary to study in detail the thermal response and possible degradation processes caused to CFC. The degradation takes place in two ways, firstly via direct mechanical fracture due to the thermal expansion of the CFC and secondly via thermo-chemical processes (phase change and pyrolysis) at high temperatures. The main objective of this work is to construct a numerical model of the major physical processes involved, and to understand the correlation between the damage mechanisms and the damage witnessed in modern CFC. For this work we are only considering the thermo-chemical degradation of CFC. Bespoke numerical models have been constructed to predict the extent of the damage caused by the two thermo-chemical processes separately (e.g. a model for phase change and a model for pyrolysis). The numerical model predictions have then been verified experimental by decoupling of the damage mechanisms, e.g. the real Joule heating from a lightning strike is replaced by a high power laser beam acting on composite surface. This was done to simplify the physical processes which occur when a sample is damaged. The experimentally damaged samples were then investigated using X-ray tomography to determine the physical extent of the damage. The experimental results are then compared with the numerical predictions by considering the physical extent of the polymer removal. The extent of polymer removal predicted by the numerical model, solving for pyrolysis, gave a reasonable agreement with the damage seen in the experimental sample. Furthermore the numerical model predicts that the damage caused by polymer phase change has a minimal contribution to the overall extent of the damage. 24 UHVnet 2011 D1.3 Switching Ferroresonant Transient Study using Finite Element Transformer Model R. Zhang1, H.Y. Li1 and Z.D. Wang*1 1 University of Manchester , UK *E-mail: [email protected] In the UK a typical distribution network (Figure 1) is configured as a grid transformer in the downstream substation to be operated by the circuit breaker in the upstream substation via a fair length of cable or overhead line. De-energising a transformer with a long cable/overhead line connected to it can induce the occurrence of switching ferroresonant transients. Following the normal de-energising switching sequence which is to open downstream circuit breaker first to shed the load and after a few more minutes to open the upstream circuit breaker in order to de-energise the no-loaded transformer, the no-load transformer would give a loud noise. This noise is caused by the core saturation and vibration when the cable/overhead line discharges via the non-loaded transformer. Core saturation and strayed flux can have some impacts and the worst scenario could be eddy current heating up insulation locally leading to excessive gassing and eventually localized core melting and failure. This paper aims to introduce two methods for switching transient ferroresonance study: Firstly using EMTP to build a distribution network circuit model in the aim to understand how each component influences the transient voltage and current results. Examples will be 1) circuit breaker grading capacitor and opening time which would affect initial condition of ferroresonance, 2) ground capacitor of cable and transformer core characteristics which would affect the magnitude of overvoltage and overcurrent or/and over-fluxing; and 3) the total resistance in the system which would affect the resonance damping time; Secondly using Finite Element to build a detailed transformer model in the aim to understand the details inside the transformer such as how flux density and power loss distribute and their influence on insulation ageing when ferroresonance and inrush transient phenomena occur. Examples will be the comparison of main and leakage flux distribution under normal operating and core saturation conditions; the influences made by design factors of transformer such as joint angles, steel materials and structures. This study intends to present the methodologies on how to merge traditional transient study with the insulation system studies to predict potential damages of system events on insulation materials. Figure 1 – Typical UK distribution network diagram 25 UHVnet 2011 D1.4 Transient Modelling of Offshore Wind Farm Connections F. Moore *1, A Haddad1 and H Griffiths1 1 Cardiff University, UK *E-mail: [email protected] There is currently over 30GW of offshore wind generation projects at various stages of development, resulting from government targets and incentives for renewable energy [1]. As a result, several large offshore wind generation projects are being connected to the GB Transmission system. The location of these wind farms out at sea means that long lengths of submarine cable are required to connect the wind farms to the onshore transmission network. It is thought that these long lengths of cable will influence the level of electromagnetic transient overvoltages seen on the transmission network which is principally constructed using overhead lines. Understanding the transient overvoltage levels allows plant to be correctly specified, and protected from transient overvoltages. Overvoltages can potentially be mitigated by using particular switching sequences, installing surge arrestors, or installing harmonic filters. Offshore transmission networks are governed by various industry codes and practices. These include the National Electricity Transmission System Security and Quality of Supply Standard (NETS SQSS) [2], which influences network architecture. By interpreting these standards, and the guidance provided by National Grid in their Offshore Development Information Statement [3], a representative offshore network is used to investigate the transient response of such networks. In this paper, a simple offshore network model (figure 1) for transient simulation in EMTP-ATP, is used to explore switching surges at various locations on the onshore and offshore networks. Figure 1: Simple Network Model in ATPDraw (EMTP-ATP) [1] http://www.thecrownestate.co.uk/rounds-one-two and http://www.thecrownestate.co.uk/round3 (Accessed November 2010). [2] http://www.nationalgrid.com/uk/Electricity/Codes/gbsqsscode/DocLibrary/ NETS SQSS (Accessed November 2010) [3] http://www.nationalgrid.com/uk/Electricity/ODIS Offshore Development Information Statement 2010, National Grid (Accessed November 2010) 26 UHVnet 2011 D1.5 Surface Gradient Calculation for Overhead Lines Q. Li*1, S. M. Rowland1 and R. Shuttleworth1 1 University of Manchester, UK E-mail: [email protected] * The most important factor that influences the generation of corona is the electric field distribution in the vicinity of the conductor surface, so calculation of the electric field strength on the surface of HV conductors becomes critical when studying corona phenomenon. The calculation of surface gradients on overhead conductor dates back to the 1950s when the Maxwell's Potential Matrix was first employed as an analytical tool [1] Over the past 60 years, a number of numerical methods [2, 3] have been applied on this subject due to the increasing power of computers. All these calculations are based on a simplified model of transmission line conductors—‗a series of cylinders in parallel to smooth ground‘. In the first part of this paper, five major methods are reviewed in detail and programmes using these methods have been written using MATLAB. One of National Grid‘s transmission line configurations—L2 RUBUS— has been selected as an example to compare the results for different methods. Following the theoretical study of existing methods, the second part represents the characteristics of different methods, and analyzes the possibilities for improving the calculation accuracy. ‗Finite Element Analysis‘ has the advantage of being able to analyze geometries with irregular shapes (Fig 1) coupled with different fields. However it is limited by the scale of geometries it can simulate. The effect of stranding has been considered in a 2-D model built in commercial software COMSOL. By making best use of the advantages and bypassing the disadvantages, a comprehensive method which employs both analytical methods and the ‗Finite Element Method‘ was developed. The transmission line is then converted to relatively small scale geometry to analyze the surface stress. Not only stranding shapes but also protrusions, dust or water droplets were analysed. Figure 1: FEA simulation results for GAP-Type Conductor [1] M. Temoshok, ―Relative Surface Voltage Gradients of Grouped Conductors‖, Trans. AIEE, Vol. 67, 1948, pp. 1583-1591. [2] Maruvada P. Sarma and W. Janischewskyj ―Electrostatic Field of a System of Parallel Cylindrical Conductors‖, IEEE Trans. on Power Apparatus and Systems, vol. PAS-88, pp. 1069 1969. [3] H. Singer, H. Steinbigler, P. Weiss, ―A Charge Simulation Method for the Calculation of High Voltage Fields‖, IEEE Trans. PAS, Vol. 93, pp. 1660-1668 Sept. 1974. 27 UHVnet 2011 D1.6 Modelling of Electroluminescence in Polymers Using a Bipolar Charge Transport Model J. Zhao*, D. H. Mills, G. Chen and P. L. Lewin University of Southampton, UK *E-mail: [email protected] c.uk Electroluminescence (EL) in polymeric materials is thought to occur due to the energy dissipation process from the recombination of opposite polarity charge carriers. It is considered as an indication of storage and transport of charge carriers in cable insulation subject to electrical stresses and may indicate the change in charge movement due to aging or degradation processes. Under ac electric fields, the interaction of opposite polarity charge carriers at the interface of polymer/conductor is enhanced compared with dc conditions, and seems to contribute a lot to the electroluminescence rather than the charge behaviours in the bulk of polymers. The dynamics of charge carriers both at the interface of polymer/conductor and in the bulk of polymers is investigated through a simulation work using a bipolar charge transport model. Figure 1 compares experimental electroluminescence results with simulated data from the recombination of injected charge carriers. The paper will give more details on EL model and comparison under various waveforms and frequencies. 6 Simulation 6kV Experiment 6kV Normalized EL intensity 5 4 3 2 1 0 0 45 90 135 180 225 angle () 270 315 360 Figure 1: Comparison of electroluminescence simulation and experimental measurement [1] P. L. Lewin, S. J. Dodd and A. M. Ariffin, ―Simulation of Electroluminescence using a Bipolar Recombination Model‖, IEEE International Conference on Solid Dielectrics, 2007, pp. 15-18. [2] J. Zhao, Z. Xu, G. Chen and P. L. Lewin, ― Numerical Modeling of Space Charge in Polyethylene under AC Fields‖, 2010 IEEE International Conference on Solid Dielectrics, 2010, Potsdam, Germany. pp. 565-568. 28 UHVnet 2011 A2 Posters: High Voltage Plant A2.1 A2.2 A2.3 A2.4 A2.5 A2.6 A2.7 A2.8 A2.9 A2.10 A2.11 A2.12 A2.13 Power Transformer End-of-life Modelling: Incorporating Thermal Lifetime Analysis with Ordinary Statistical Analysis ................................................................................................................... 30 Location of Partial Discharges within a Transformer Winding Using Principal Component Analysis ........................................................................................................................................ 31 Frequency Response Analysis of Transformer Winding Deformation Based on Multiconductor Transmission Line Model ............................................................................................................... 32 Effect of Climatic Condition on Polymeric Insulators .................................................................................... 33 Acoustic Noise Evaluation for Overhead Lines............................................................................................... 34 Transient Fault Location in Low Voltage Distribution Networks ................................................................... 35 A Survey on the Potential of CF3I Gas as an Alternative for SF6 .................................................................. 36 A New Technique to Enhance the Earthing System by Increasing the Horizontal Earth Electrode Effective Length ............................................................................................................................. 37 A Novel Portable Testing Device for Surge Protective Systems .................................................................... 38 A Solar Powered Wireless Data Acquisition System for High Voltage Substations ....................................... 39 The Performance of Nanocoating on High Voltage insulators ........................................................................ 40 Performance of Tower Footings Resistance under High Impulse Current ...................................................... 41 High Frequency Performance of a Vertical Rod Electrode ............................................................................. 42 29 UHVnet 2011 A2.1 Power Transformer End-of-life Modelling: Incorporating Thermal Lifetime Analysis with Ordinary Statistical Analysis D.Y. Feng1, Z.D. Wang1* and P. Jarman2 1 2 University of Manchester, UK National Grid, UK * E-mail: [email protected] In many developed countries such as the UK, majority of power transformers operating in the power system networks have passed their designed lifetime. It is of great importance to establish a model to accurately predict transformers‘ lifetime so that this model can help asset managers plan the replacement of wear-out transformers, because firstly designed lifetime was not based on operation of old transformers but rather on educated guess; secondly the re-investment in transmission and distribution system infrastructure requires excessive capitals and must be done in a planned manner. In this paper, an intensive statistical analysis on UK‘s transmission transformer population during 1952 to 2004 has been carried out and confirms, with high confidence, a low and stable failure rate for transformers exist till 40 years of age. However, due to the very limited data of older transformers, statistical tools are unable to predict the failure rate of aged transformers and have therefore lost their values to asset managers. In order to compensate statistical tools‘ incapacity in dealing with incomplete dataset, the conservative view of ‗transformer lifetime equals to insulating paper‘s life‘ is taken to assist the statistical approaches. Thermal lifetimes of 77 retired transformers are derived using the lowest DP measured from the paper insulation samples. A correlation of thermal failure rate and transformer ages could thereby be derived and combined with the statistical results to form a complete trend of the transformer failure rate as age progresses (Fig.1). Since the 77 retired transformers‘ representativeness to the entire transformer population in terms of thermal lifetime span is uncertain, to enlarge this sample size, the IEC thermal model is used to model a transformer population‘s thermal lifetimes. The second part of the paper focuses on discussing the suitability and applications of IEC thermal model: firstly inputs to the model are defined with a statistical range to reflect the design and operating conditions of a specific transformer population, secondly introducing various practical thermal aging mechanisms (e.g. oxidation- or hydrolysis-dominated) to improve the IEC thermal model. Ultimately these improvements on thermal modelling approach would assist the statistical analysis in transformers‘ end-of-life modelling. Figure 1 : Full life cycle of a transformer‘s failure rate (0.24% derived from historical data using statistical approaches incorporated with the thermal failure rate derived from the IEC thermal model) 30 UHVnet 2011 A2.2 Location of Partial Discharges within a Transformer Winding Using Principal Component Analysis M. S. Abd Rahman*1, P. L. Lewin1 and L. Hao1 1 University of Southampton, UK *E-mail: [email protected] Partial discharge (PD) may occur in a transformer winding due to ageing processes or defects introduced during manufacture. A partial discharge is defined as a localised electric discharge that only partially bridges the dielectric insulator between conductors when the electric field exceeds a critical value. The presence of PD does not necessarily indicate imminent failure of the transformer but it is a serious degradation and ageing mechanism which can be considered as a precursor of transformer failure. PD might occur anywhere along the transformer winding and the discharge signal can propagate along the winding to the bushing and neutral to earth connections. As far as maintenance and replacement processes are concerned, it is important to identify the location of PD activity so any repair or replace decision is assured to be cost effective. Therefore, identification of a PD source as well as its location along the transformer winding is of great interest to both manufacturers and system operators. The wavelet transform is a mathematical function that can be used to decompose a PD signal into detail levels and an approximation. Wavelet filtering is often used to improve signal to noise ratio (SNR) of measured signals, but in this case it is used to identify the distribution of signal energies in both the time and frequency domains. This method produces a feature vector for each captured discharge signal. The use of principle component analysis (PCA) can compress this data into three dimensions, to aid visualisation. Data captured by sensors over hundreds of cycles of applied voltage can be analysed using this approach. An experiment (Figure 1) has been developed that can be used to create PD data in order to investigate the feasibility of using PCA analysis to identify PD source location. Figure 1: Experimental diagram for measuring partial discharge within transformer winding [1] L. Hao, P. L. Lewin and S. G. Swingler, ―Identification of Multiple Partial Discharge Sources‖, IEEE International Conference on Condition Monitoring and Diagnosis, 2008. [2] E. M. Lalitha and L. Satish ―Wavelet Analysis for Classification of Multi-source PD Patterns‖, IEEE Transactions on Dielectrics and Electrical Insulation, vol. 7, no. 1, pp.40-47, February 2000 [3] K. X. Lai, B. T. Phung and T. R. Blackburn, ―Partial Discharge Analysis using PCA and SOM‖, IEEE Power Tech, 2007, pp.2133-2138. 31 UHVnet 2011 A2.3 Frequency Response Analysis of Transformer Winding Deformation Based on Multi-conductor Transmission Line Model T. Y. Ji1, W. H. Tang1*, C. H. Wei1 and Q. H. Wu1 1 University of Liverpool, UK *E-mail: [email protected] This paper presents a multi-conductor transmission line (MTL) model of a transformer winding to perform frequency response analysis (FRA) for the detection of winding deformation. The MTL model is built based on the travelling wave theory [1-3]. It views every disc of a transformer winding as a transmission line and these lines are parallel with each other and with the ground. The winding conductor is composed of infinite segments of length dx and each unit is described by the parameters of inductance, resistance, capacitance and conductance. These parameters need to be identified in order to model a transformer for FRA analysis. The construction of the MTL model is explained in detail in the paper. The geometry parameters of a test transformer and its measured FRA data in normal condition are provided in this paper. The winding parameters of the MTL model of the transformer are firstly obtained by theoretical calculation, then refined by optimisation towards the minimum error between the frequency response of the model and the measured frequency response. By comparing the measured frequency response of the test transformer and the frequency response of its MTL model, it has shown that the MTL model is accurate enough for FRA, as long as the winding parameters are properly given. The winding parameters of the MTL model with deformed winding are estimated using finite element method (FEM), and the frequency response is calculated correspondingly. Two deformation scenarios are involved in the simulation study, and the FRA results have been presented to show that the deformation can be detected and the degree of deformation can be reflected as shown in Figure 1. Figure 1: The frequency responses of the normal and the deformed windings. [1] E. P. Dick and C. C. Erven. Transformer diagnostic testing by frequency responce analysis. IEEE Transactions on Power Apparatus and Systems, 1(6):2144–2153, 1978. [2] E. Rahimpour, J. Christian, and K. Feser. Transfer function method to diagnose axial displacement and radial deformation of transformer windings. IEEE Transactions on Power Delivery, 18(2):493–505, 2003. [3] R. Rudenberg. Electrical shock waves in power systems: traveling waves in lumped and distributed circuit elements. Harvard University Press, Cambridge, Massachusetts, 1968. 32 UHVnet 2011 A2.4 Effect of Climatic Condition on Polymeric Insulators A.S. Nekeb*1, N. Harid1, A. Haddad1 1 Cardiff University, UK *E-mail: [email protected] The problem of aging or degradation of outdoor polymeric insulators has been a major concern for electrical utilities and researchers. Their surface hydrophobicity can be significantly reduced with time as they are subjected to variable climatic conditions such as temperature, humidity, and ultraviolet radiation. These can accelerate the aging process depending on their severity and the geographic location of the insulators, and may lead to loss of their insulating properties and subsequent flashover. Fig. 1 summarises the effect of various climatic conditions on polymeric insulators. This paper reviews the effect of climatic conditions on polymeric insulators, with emphasis on the effect of ultraviolet radiation. Previous research work on this subject is reported and discussed. The standards covering UV irradiation methods on insulator samples are reviewed. Initial tests consisting of subjecting insulator samples to various cycles of irradiation are described. The test facility, which also provides automatic temperature and humidity control for a given test cycle, is described. Figure 1: Effect of Climatic conditions on polymeric insulators 33 UHVnet 2011 A2.5 Acoustic Noise Evaluation for Overhead Lines Qi. Li*1, G. Zhang1, S. M. Rowland1 and R. Shuttleworth1 1 University of Manchester, UK E-mail: [email protected] * Followed by the rapid increase of voltage level in modern power systems, audible noise is now becoming one of the critical design and environmental considerations for overhead lines. Power utilities have carried on plenty of experimental measurements to evaluate the noise level emitted from high voltage conductors. Not only outdoor measurements which are carried out close to a whole span of overhead line [1] but also indoor measurements [2, 3] which employ cage configurations to simulate sections of overhead line have been reviewed in the first section of this paper. Processes for predicting audible noise of overhead lines have been summarized. Cage experiments have been well proven to be an effective way to study the environmental impact of transmission lines. It is employed to control the surface gradient in acoustic noise testings for different conductors. A cage set-up (National Grid High Voltage Lab in Manchester) for audible noise examination is proposed in the second section of this paper. An appropriate cage radius was first selected by considering the various conductor configurations. In order to mitigate the excessive corona activities in each end of conductor, a corona guard is installed at each end. The radius of this guard, so as to eliminate corona, is then determined by FEA simulation results. The nonuniformly distributed electric field within the cage is finally examined. The length of middle section which provides a uniformly distributed electric field is then determined, enabling effective experimental design. Besides the electrical design, the tensioning system design and acoustic measurement techniques for the cage experiment are introduced. An anechoic chamber was introduced to cover the whole cage so as to shield the whole cage from background noise. Figure 1: A typical cage set-up [2] [1] R.A. Popeck and R.F. Knapp "Measurement and Analysis of Audible Noise from Operating 765 kV Transmission Lines", IEEE Transactions on Power Apparatus and Systems, vol. PAS-100, pp. 2138 1981. [2] M. J. Lekganyane, N. M. Ijumba, and A. C .Britten, ―Corona Audible Noise Measurements in a Small Indoor Corona Cage under HVDC Voltages‖, 2006 International Conference on Power System Technology [3] Minhua Ma, Yuming Zhao, Zhicheng Guan, Liming Wang, ―The influence of contaminations on HVDC conductor corona characteristics‖, Annual Report Conference on Electrical Insulation and Dielectric Phenomena, 2007. 34 UHVnet 2011 A2.6 Transient Fault Location in Low Voltage Distribution Networks Yuxian Tao*1, W.H.Siew1 and J.J. Soraghan1 1 University of Strathclyde, UK *E-mail: [email protected] Underground cables are widely used in the UK for electricity distribution. Additionally, many of the cables are approaching the end of their design-life. Distribution Network Operators (DNOs) will normally keep these aged cables in service to extend their service lifetime. However, the aged cables are prone to develop faults, which result in loss of power supply. This leads to customer minutes lost -- a parameter that is monitored by the Electricity Regulator (ER). Hence, it becomes desirable to know where in their power network, a fault might be developing. The early stage of a fault could be classified as transient faults. Time domain reflectometry (TDR) is mainly used to pre-locate faults in low voltage underground cable network. However the success of this technique to address transient faults relies both on the simplicity of the cable circuit being diagnosed and the point of diagnosis. This is because T-joints in a cable network and parallel circuits at the point of diagnosis could result in a complicated waveform being acquired and therefore making unreliable pre-location. Hence current technologies for identifying and locating transient faults either require access to residential homes or use of a heavy blocking inductor to ensure that only a particular circuit is being investigated. Furthermore, the network operators are not able to pre-locate after the event because transient faults are not predictable and may not recur during the investigation. Transient faults that occur in underground cable networks are mostly electric arcing caused by insulation failure. Electric arcing in low voltage cables is often self-extinguishing [1] and does not present an immediate hazard. This is the reason for their unpredictability. This presentation will therefore review the existing technologies and to present a strategy for overcoming the restrictions or limitations posed by the existing strategies for locating transient faults. [1] W. Charytoniuk, Wei-Jen Lee, Mo-Shing Chen, & J. Cultrera. ―Arcing fault detection in underground distribution networks - feasibility study‖, IEEE Transactions on Industry Applications, Volume 36, Issue 6, pp. 1756-1761, 2000 35 UHVnet 2011 A2.7 A Survey on the Potential of CF3I Gas as an Alternative for SF6 M. S. Kamarudin*1, M. Albano1, P. Coventry2, N. Harid1 and A. Haddad1 1 2 Cardiff University, UK National Grid UK *E-mail: [email protected] Sulphur hexafluoride (SF6) has been widely used as an insulator in gas-insulated switchgear (GIS) applications. But due to the fact that it is a greenhouse gas, many researchers have been trying to find alternative solutions for it. Furthermore, SF6 produces highly toxic and corrosive substances when it is subjected to electrical discharges. Trifluoroiodomethane (CF3I) has recently been regarded as a candidate for replacing SF6. CF3I has been used as a fire suppressor and now many investigations have been carried out throughout the world to assess its capability in high voltage applications. This paper surveys this previous work and identifies some of the properties which are relevant to high voltage applications. With a global warming potential 23,900 times greater than carbon dioxide (CO2), and atmospheric lifetime 3,200 years, SF6 is the most potent greenhouse gas in existence. Its production is now restricted under Kyoto Protocol. Table 1 shows the general properties comparison between CF3I with SF6. Table 1: General properties of CF3I and SF6 Material Molecular mass CF3I 195.91 Colourless Non-flammable Less than 5 0.0001 0.005 – 22.5°C Characteristic Global Warming Potential (GWP) Ozone Depleting Potential (ODP) Lifetime in atmosphere (year) Boiling point (0.1 MPa) SF6 146.05 Colourless Non-flammable 23,900 0 3,200 – 63.9°C At 0.5 MPa, the boiling point of CF3I is around 25°C, compared to –30°C for SF6 [1]. For this reason, it can be difficult to compress CF3I in HV switchgear at temperature common in winter. The adoption of other gases such as nitrogen (N2) or CO2 helps in reducing the boiling point, and it is required for outdoor application. Using Dalton‘s law, the partial pressure in a CF3I- N2 gas mixture can be expressed as Pgas mixture = PCF3I + PN2 (1) where Pgas mixture PCF3I PN2 total pressure of the gas mixture partial pressure of CF3I gas partial pressure of N2 A study by Toyota et. al in 2006 [2] revealed that for a same gap length of electrodes, CF 3I gas has a higher dielectric strength of SF6, which is about 1.2 times higher. Another study by the same researchers revealed that a mixture of 60% CF3I with 40% N2 has a dielectric strength as equal to that of SF6. With a dielectric strength of 1.2 times better than SF6, CF3I has been identified as a very good candidate to replace SF6 as a gas insulator. More research works should be carried out on the capabilities of CF 3I, particularly with regards of its mixtures, its performance under uniform and non-uniform field and also to control the by-products produced after a successful discharge. Work is in progress to develop a test facility to explore the properties of CF3I and its mixtures for insulation and switchgear applications. [1] M. Taki, D. Maekawa, H. Odaka, H. Mizoguchi and S. Yanabu, ―Interruption Capability of CF3I Gas as a Substitution Candidate for SF6 Gas‖, IEEE Transactions on Dielectrics and Electrical Insulation, vol. 14, no. 2, 2007. [2] H. Toyota, S. Matsuoka and K. Hidaka, ―Measurement of Sparkover Voltage and Time Lag Characteristics in CF3IN2 and CF3I-Air Gas Mixtures by using Steep-Front Square Voltage‖, Electrical Engineering in Japan, vol. 157, no. 2, 2006. 36 UHVnet 2011 A2.8 A New Technique to Enhance the Earthing System by Increasing the Horizontal Earth Electrode Effective Length A Elmghairbi*1, N Harid1, H Griffiths1, A Haddad1 1 Cardiff University, UK *E-mail: [email protected] Horizontal earth electrodes are commonly used to enhance earthing systems located in areas with high soil resistivity, in order to reduce overall earth impedance. Such electrodes are commonly used, for example, to interconnect adjacent earthing grids or the individual earthing systems of turbines on windfarms. The ability of a horizontal earth electrode in reducing earth potential rise is limited because, no further reduction is obtained by increasing its length beyond a certain length known as the effective length. It is also well known that the behaviour of earthing systems subjected to transients is different from that under power frequency faults. This results in a shorter effective length under transient conditions than at power frequency. In this paper, field experiments and computer simulations of a sectionalised 88m length horizontal earth electrode under different energisations (dc, variable frequency ac and transients of different shapes) are reported. From measurements of the voltage and current at the injection point, and by incrementally increasing the length of the test electrode, the effective length of the horizontal earth electrode was determined under the different energisations for the soil conditions pertaining to this particular installation. Moreover, current and voltage distribution along the length of the electrode was recorded. The experimental and simulation results of voltage and current show reasonably close agreement and also that quite good prediction of the effective length using simplified formulae is possible. Using the same experimental facility and test electrode, a new proposed method to increase the effective length of the horizontal earth electrode was investigated by installing an additional above ground insulated parallel conductor which is bonded to the bare underground horizontal electrode at points along its length. The results show that the current and voltage distributions are changed such that a greater length of buried conductor is utilised and that this contributes to an additional reduction in the earth impedance, and hence the developed earth potential rise, at the point of current injection. Measured CDEGS formula Conductor length (m) 50 45 40 35 30 25 20 1 1.5 2 2.5 Rise time (ms) 3 3.5 4 4.5 5 Figure 1: Effective length of a horizontal earth electrode [1] M. Muhr, S. Pack and S. Jaufer, ―Usage and Benefit of an Overhead Line Monitoring System‖, International Conference on High Voltage Engineering and Application, November 2008 [2] M. M. Werneck, and A. C. S. Abrantes, ―Fiber-Optic-Based Current and Voltage Measuring System for HighVoltage Distribution Lines‖, IEEE Transactions on Power Delivery, vol. 19, no. 3, July 2004 [3] C. A. Spellman, A. Haddad D.M. German and R. T. Waters, ―Improved Three-Phase Voltage Measurement Using Capacitive Probe‖, Proceeding of the UniversitiesPower engineering Conference, vol. 1, pp.352-355, 1999 37 UHVnet 2011 A2.9 A Novel Portable Testing Device for Surge Protective Systems C. Long*1, W. Zhou2, C. Zhou1, J. Yu2 1 2 Glasgow Caledonian University, UK Wuhan University, China *E-mail: [email protected] With significant increase in penetration of the Surge Protective Devices (SPDs) into areas such as telecommunications, electricity, meteorology, railway, petrochemical industry, field testing of SPDs has become a great concern. At the present, however, impulse wave generators, generally utilized as laboratory equipments, are too heavy to be utilised for the purpose. In this paper, a novel, portable testing device providing with a combination of maximum amplitude of 6kV/3kA and DC voltage supply of maximum amplitude of 2kV is presented. It is designed to test on site the condition of SPDs which have been utilized for many years. The device is equipped with a PWM and a high frequency circuit in order to minimize DC voltage supply. The device uses a DSP as the core component to measure and determine the peak impulse voltage and impulse current automatically. Meanwhile, simulation of EMTP-ATP has been carried out to evaluate the design of the combination wave generator circuit which needs to generate IEEE standard combination waveforms. The calculation, simulation and field testing data demonstrate that the portable test device with weight of mere 13kg is capable of characterising and diagnosing the condition of SPDs. [1] C. Long, et. al., ―Study on Portable SPD Site Test System, Shaanxi Electric Power‖, 2010, 20(6):55-58 [2] ―IEC 61643-1: 1998 Low-voltage surge protective devices – Part 1: Surge protective devices connected to lowvoltage power distribution systems – Requirements and tests‖, 1998 [3] L. Liu, G. Zhang, et. al., ―Design of Combined Wave Generator by Using Universal Variable Method‖, High Voltage Engineering, 2007, 33(1): 123-130 38 UHVnet 2011 A2.10 A Solar Powered Wireless Data Acquisition System for High Voltage Substations A.C. Bogias*1, N. Harid1, and M. Haddad1 1 Cardiff University, UK *E-mail: [email protected] Wireless communication systems can offer significant advantages when applied to power system monitoring applications and may become attractive to power utilities [1]. In this work a wireless solar battery powered data acquisition system for monitoring high voltage substation equipment was developed. An example-application to a surge arrester was used to illustrate the capabilities of the developed system. Various diagnostic methods for monitoring surge arresters exist. However, those that measure the operating voltage and leakage current and carry out signal processing can provide more accurate diagnostic information [2, 3]. The proposed WLAN system, shown in Figure 1, consists of a solar panel, four lithium-ion batteries, surge protection and signal conditioning IC‘s, power electronic IC‘s, a microcontroller and Wireless Local Area Network (WLAN – IEEE 802.11b/g) module, making up the WLAN Sensor. The WLAN Sensor acquires and transmits the current and voltage signals from the test surge arrester. The transmitted data is received by a WLAN access point connected to a remote Personal Computer (PC), data processing using LabVIEW. A prototype wireless data acquisition system has been built and successfully used to measure the leakage current and applied voltage of a surge arrester tested in the Cardiff High Voltage laboratory. The results are in close agreement with those recorded directly through a Data Acquisition (DAQ) card and transmitted via coaxial cable. . Figure 1: The WLAN Sensor with the solar panel. [1] F. Cleveland, ―Use of Wireless Data Communications in Power System Operations‖, Power Systems Conference and Exposition, 2006. PSCE '06. 2006 IEEE PES. 2006. [2] M. Haddad, D. Warne. (2004). ZnO surge arresters . In: Advances in High Voltage Engineering . London: Institution of Engineering and Technology. 191-244. [3] J. Lundquist, L Stenstrom, A. Schei, and B. Hansen, ―New method for measurement of the resistive leakage currents of metal-oxide surge arresters in service‖', IEEE Transactions on Power Delivery,Vol. 5(4), pp. 18111822, 1990. 39 UHVnet 2011 A2.11 The Performance of Nanocoating on High Voltage insulators S. Braini *1, A. Haddad* 2 Cardiff University, UK *E-mail: [email protected] The most important factor for dimensioning outdoor insulation is pollution performance. Depending on the pollution severity of the site, outdoor insulators need to have sufficient surface leakage length to ensure that dry band formed and surface arcing is minimised. Conventional insulators, made of porcelain and glass, have shown good performance over decades of inservice performance. However, they suffer from a hydrophilic surface property, thus allowing high surface leakage current to flow on the wetted surface. Such currents cause dry bands at areas of high current density and lower wetting rates, which eventually causes surface arcing and frequently complete flashover of the insulator. A number of remedies were used in the past to improve the surface properties, including greasing and RTV coatings. These mitigation techniques, although effective, are very labour intensive and expensive solutions. A recently proposed solution consists of the application of a nano-coating on the surface of the insulator. The nano-coating used in this work is a Voltshield coating, which is a chemically cross-linked polymeric resin with extremely good 'non-stick' properties. This coating bonds to the surface of the insulator and forms a very thin layer which gives highly hydrophobic properties to the surface of the insulator. The coated surface permits the rapid dispersal of water and does not allow adherence of solid pollutant due to its cross-linked molecular bonding [1]. In this work, the performance of this nano-coating is investigated under artificial pollution conditions using the dry layer method of IEC60507. It was found that the nano-coated porcelain insulator performs much better than the non-coated porcelain insulator both in suppressing the leakage current activities and retaining its surface hydrophobicity. Figure 1 shows the test results obtained using a dry pollution layer and clean fog as recommended in IEC 60507. a) nano-coated insultors b) standard insulator Figure 1: Pollution test results on standard and nano-coated porcelain insulators [1] J. Blacket, ―VOLTSHIELD- ANTI-POLLUTANT TREATMENT FOR GLASS AND GLAZED PORCELAIN INSULATORS― 20th International Conference on Electricity Distribtion, Praque,8-11 June 2009 40 UHVnet 2011 A2.12 Performance of Tower Footings Resistance under High Impulse Current M. Ahmeda, N. Harid, H. Griffiths and A. Haddad Cardiff University, UK *E-mail: [email protected] The rise of earth potential associated with the flow of fault current through transmission towers is important for assessing the lightning performance of lines, and more importantly, the risk of hazardous step and touch voltages at tower bases. Studies on full-scale tower base grounding systems available in the literature are limited [1,2]. Measurements on such electrodes offer a valuable means of understanding their lightning behaviour and for validating theoretical models. In this work, the tests on the tower footings were intended to determine impulse resistance and demonstrate its non-linear variation with current magnitude. In this paper, a high-current test set-up consisting of a 20kA impulse generator and a 30m-long overhead line suspended on wood poles of 1.6m height connecting the current (A) impulse generator to the test tower base. A ring electrode connected to 8 peripheral rods was used as a return current electrode. The tower base potential rise with reference to the current return point of the generator was measured. Figure 1 shows the impulse resistance decreasing as a function of peak current magnitude for all tower footings. This fall is attributed to soil ionisation around the test electrode. Figure 1: Variation of impulse resistance of tower footings with current magnitude. The highest reduction in resistance occurs with the footing which has the highest DC resistance (Leg 1). When the current increases from 600A to 6kA, the impulse resistances decreased by 47% for Leg 1, 40% for Leg 3, 22% for Leg 2 and 14% for Leg 4. For the complete tower base, with all four legs connected in parallel, the percentage reduction in the impulse resistance is relatively small due to the current division between the legs which leads to limited ionisation progression around the individual legs. [1] M Takeuchi, Y. Yasuda, H. Fukuzono, K. Kawabata, T. Hara and S. Sekioka, ―Impulse Characteristics of a 500kV Transmission Tower Footing Base with Various Grounding Electrodes‖, International Conference on Lightning Protection, ICLP, 1998 [2] E. A. Cherney, K. G. Ringler, N. Kolcio and G. K. Bell P. L. Lewin and S. J. Dodd, ―Step and Touch Potentials at Faulted Transmission Towers‖, IEEE Transactions on Power Apparatus and Systems, vol. PAS-100, no. 7, pp.3312-3321, 1981 41 UHVnet 2011 A2.13 High Frequency Performance of a Vertical Rod Electrode S. Mousa, N. Harid, H. Griffiths and A Haddad Cardiff University, UK E-mail: [email protected] Earth electrodes are known to have variable frequency dependence. This was verified by several published papers on the subject [1, 2]. Such behaviour is thought to be affected by inductive and capacitive stray components of the earth electrodes. Computations have shown that for given geometries and soil conditions, a sharp departure of the earth electrode impedance from the DC/low frequency values can occur. To date, most experimental and measurement techniques use low magnitude DC, low frequency ac an impulse currents. No established technique is available to investigate experimentally the performance of these electrodes under variable high frequency up to hundreds of kilohertz. In this work, we explore the performance of vertical rods using a high frequency source and a test electrode of 1.2m length in non-uniform soil. Figure 1 shows the measured impedance over the frequency range of DC to 120kHz. As can be seen, there is a sharp decrease of impedance at about 60kHz. At this value the downturn frequency happens and the capacitive behaviour appears also the ground resistance decreases from 100.95 to around 8 . Impedance () 1000 100 10 1 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 Frequency (Hz) Figure 1: Measured frequency response of 1.2m long vertical earth electrode. [1] Davies A M, H.Griffiths and T.E.Charlton, ‗High Frequency Performance of a Vertical Earth Rod‘, Proceedings of the 24th International Conference on Lightening Protection (ICLP), Birmingham UK, 1998, pp.536-540. [2] Griffiths H, Haddad A and Harid N, ‗Characterisation of earthing systems under high frequency and transient conditions‘, Proc. of the 28th Universities Power Engineering Conference (UPEC 2004), Vol. 1, pp.188-192, Bristol, England, 6-8 September 2004. 42 UHVnet 2011 B2 Posters: Condition Monitoring B2.1 B2.2 B2.3 B2.4 B2.5 B2.6 B2.7 B2.8 B2.9 B2.10 B2.11 B2.12 FDTD Modelling of Partial Discharge Detection in Power Distribution Cables using HFCTs.............................................................................................................................................................. 44 Use of Hidden Markov Model for Partial Discharge-led Failure Development Modelling ............................ 45 Dynamically Weighted Ensemble of Neural Networks for Classifying Partial Discharge Patterns ............................................................................................................................................................ 46 A Successful On-site PD Testing Experience of 11kV EPR Cable Insulation Systems .................................. 47 Radiometric Arc Fault Detection ..................................................................................................................... 48 Voltage Transducer for Transient Measurements on High Voltage Overhead Lines ..................................... 49 Fault Location using FPGAs and Power Line Communication ....................................................................... 50 A New Method to Improve the Sensitivity of Leak Detection in Self-Contained Fluidfilled Cables .................................................................................................................................................... 51 Energy Harvesting from Electric Fields in Substations for Powering Autonomous Sensors ............................................................................................................................................................. 52 Ageing and Temperature Influence on Polarization/Depolarization Current Behaviour of Paper Immersed in Natural Ester ..................................................................................................................... 53 An On-line Lightning Monitoring System for Transmission Lines ................................................................. 54 Energy Harvesting in Substations for Wireless Sensors and a New Arc Capacitor Structure 55 43 UHVnet 2011 B2.1 FDTD Modelling of Partial Discharge Detection in Power Distribution Cables using HFCTs X Hu*1, A J Reid1, M D Judd1 and W H Siew1 1 University of Strathclyde, UK * E-mail: [email protected] Partial discharge (PD) is an electrical discharge that bridges only part of the insulation material in an insulation system such as those in power cables. If these discharges are left to develop, they will gradually erode the insulating material and may lead to complete failure of the cable or accessory. PD measurement has been widely used in power industries as a practical tool to evaluate insulation condition of cables because the pulses can often be detected at the cable ends using high frequency current transformers (HFCTs). PD propagation in cables experiences significant attenuation [1-3] and thus PD pulse magnitude and shape may change a lot by the time it finally reaches a detection point, which might be up to a few kilometres away. If propagation effects can be properly characterised, diagnostic post-processing tasks like classification and identification might be enhanced. In order to bridge the gap between PD measurement and interpretation of the results, finite difference time domain (FDTD) methods are being evaluated as a means of modelling the entire PD process from current pulse at source to detection by a remote HFCT. Various dimensions and types of cables have been modelled using commercial software (XFdtd [4]) and the modelling has been oriented towards a comparison with the PD measurement using HFCTs. The source current pulse can be derived from experimental PD data while the PD current pulse at the location where the HFCT would be attached must be calculated numerically from the simulated magnetic field data by applying Ampere‘s Law. This paper describes how a PD current pulse is modelled using FDTD and examples of the propagating electric field will be given. Figure 1 shows a slice of electric fields travelling along the cable after a PD has been excited on the inner conductor. The principle of simulating the HFCT will be introduced and results will be presented to demonstrate the accuracy of the Ampere‘s Law current sensor. Output will then be suitable for combination with the HFCT‘s transfer function to obtain HFCT‘s response. The findings of this work may helpfully improve the understanding of existing results from PD measurement in cables. Figure 1. FDTD model of a 1m cable sample [1] C. Xu, L. Zhou, J. Y. Zhou, and S. Boggs. High frequency properties of shielded power cable part 1: Overview of mechanisms. IEEE Electrical Insulation Magazine, 21(6):24–28, Nov/Dec 2005. [2] S. Boggs, A. Pathak, and P. Walker.Partial discharge XXII: High frequency attenuation in shielded solid dielectric power cable and implications thereof for PD location. IEEE Electrical Insulation Magazine, 12(1):9–16, Jan/Feb 1996. [3] N. Oussalah, Y. Zebboudj, S. Boggs. Partial Discharge Pulse Propagation in Shielded Power Cable and Implications for Detection Sensitivity. IEEE Electrical Insulation Magazine, 23(6):5-10, Nov/Dec 2007. [4] http://www.remcom.com/ 44 UHVnet 2011 B2.2 Use of Hidden Markov Model for Partial Discharge-led Failure Development Modelling D. Zhou1,2, C.R.Li2 and Z. D. Wang1* 1 2 University of Manchester, UK North China Electric Power University, China * E-mail: [email protected] Partial discharge (PD) detection and monitoring is of vital importance to confirm the insulation integrity of the insulation of high voltage equipment, especially for equipment with a complex insulation system, such as power transformers. In power transformers, PD can lead to surface tracking on solid insulating materials, which causes irreversible damage and may eventually lead to a breakdown of the concerned operating component. PD can also decompose the insulating oil and pollute the oil system in such a way that the insulation properties of the oil can no longer be guaranteed. Continuous on-line monitoring of partial discharges becomes possible with the development of advanced sensors and data-acquisition systems. However, there is still a lack of effective data analysis and modelling methods for the correct interpretation of the obtained PD monitoring data to prevent PD-led failures. The severity and evolution stage of PD led insulation degradation process must be recognized before the full utilisation of continuous on-line PD monitoring. In this paper a hidden markov model (HMM) with left-right topology for PD development modelling is proposed. HMM, as a double stochastic process, comprises an underlying stochastic process that is not directly observable but can be visualized through another set of stochastic processes that produce a sequence of observations. This is in accordance with the nature of partial discharge developing process as we know. In the proposed HMM model, initiation stage, developing stage I, developing stage II and pre-breakdown stage are defined as the four underlying PD evolution stages. Pulse repetition rate, maximum pulse amplitude and average pulse amplitude are chosen as the observations. Two sets of experimental data are utilized for the training of the model. A third set of data is used for verification. Promising results show that a highly successful recognition rate of PD evolution stage can be achieved using HMM through the decoding process. Training Stage Recognition Stage Data Acquisition Data Acquisition Data Preprocessing Data Preprocessing Feature Extraction Feature Extraction HMM training HMM State Decoding Figure 1: the flowchart of partial discharge development modelling process 45 UHVnet 2011 B2.3 Dynamically Weighted Ensemble of Neural Networks for Classifying Partial Discharge Patterns A. Abubakar Mas‘ud*1, B. G. Stewart1, S. G. McMeekin1 and A.Nesbitt1 1 Glasgow Caledonian University, UK *E-mail: [email protected] Partial discharges (PD) measurements are used to monitor the degradation of insulation subjected to High Voltage (HV) electrical stress. Recently, the pursued goal has been identifying a robust method for recognition of different categories of PD data using expert systems. This can be achieved through the ensemble of neural networks (ENN), where by different NN models are trained and their predictions combined as shown in Fig.1. The training data for the ENN considered here comprises 9 statistical parameters obtained on different PD measurements of corona from a point-plane arrangement. Dynamically weighted averaging of neural networks is adopted where the ensemble output gives the weighted average of the output of each NN. Using this strategy it is shown that the ENN output has the best accuracy of 98.12%, while single NNs such as the Multi-Layered Perception Network (MLPN), Elman Recurrent Network (ERNN) and Radial Basis Function Network (RBFN) have independent accuracies of 97.5%, 96.97% and 97.4% respectively. The results demonstrate that the ENN model has potential for further application to other PD scenarios. Figure 1: An ensemble of Neural Networks. [1] Z. Zhao, Y. Zhang and H. Liao, ―Design of ensemble neural network using the Akaike information criterion‖, Engineering Applications of Artificial Intelligence, 2008. [2] D. Jimenez, ― Dynamically weighted Ensemble Neural Networks for Classification‖, IEEE World Conference on Computational Intelligence, 1998 [3] T. Hong, M.T.C Fang, ―Detection and Classification of Partial Discharge Using a Feature Decomposition-Based Modular Neural Network, vol.50, no.5, October 2001. 46 UHVnet 2011 B2.4 A Successful On-site PD Testing Experience of 11kV EPR Cable Insulation Systems Xiaosheng Peng * 1 , Chengke Zhou 1 , Donald M. Hepburn 1 , Xiaodi Song 1 Glasgow Caledonian University, UK * Email: [email protected] A power generating station in the United Kingdom has reported a number of in service failures in its 11kV single core Ethylene-Propylene Rubber (EPR) insulated cables. Several industrial companies have carried out on-site condition assessment to determine whether other insulation defects were present but no conclusive results have been found due to presence of strong background electrical noise. The present authors were invited to carry out on-site testing to demonstrate the effectiveness of their denoising techniques. This paper presents the processes of the on-site cable partial discharge signal detection experience and the signal processing of the raw data. Following a brief introduction to the tests, equipments and connections, the paper analyses sources of different types of interference signals. These are found to originate mainly from UPS Inverter Supplies or 11kV Motor drive circuits. Thereafter, second generation wavelet transform (SGWT) data denoising algorithm is introduced. SGWT is proved to be an effective denoising technique for the detected data. Also presented in the paper are PD pattern identification and PD source localization methods which are used to identify the source of the PD signal. Finally the diagnosis results, with indication of potential insulation defect and cable joint problems, are provided. [1] X. Zhou, C. Zhou, I. J. Kemp, ―An improved Methodology for Application of Wavelet Transform to Partial Discharge Measurement Denoising‖, IEEE Trans. Dielectrics and Electrical Insulation, Vol. 20, No. 2, March, 2005. [2] C. Zhou, X. Zhou, B. Stewart, A. Nesbitt, D. Hepburn and D. Guo: ―Comparisons of Digital Filter, Matched Filter and Wavelet Transform in PD Detection‖, Recommended by CIGRE UK panel to CIGRE 2006. [3] X. Song, C. Zhou, D.M. Hepburn and M. Michel, ―Second Generation Wavelet Transform in PD Measurement Denoising‖, IEEE Trans. Dielectrics and Electrical Insulation, Vol. 14 (6), 2007 47 UHVnet 2011 B2.5 Radiometric Arc Fault Detection R. M. Harris*1, M. D. Judd1 and P. J. Moore2 1 2 University of Strathclyde, UK Elimpus Ltd., UK *E-mail: [email protected] When arcs are formed they are accompanied by the radiation of electromagnetic transients which can be captured as a signal which propagates in the radio spectrum. Such captures are commonly utilised in the study and location of lightning arcs [1], where the emissions are referred to as sferics. Our aim in recent research has been to capture radiometric emissions emitted by arcs resulting from faults on power distribution networks. The advantage of being able to do this is that the detection equipment does not require a direct connection to the power system, and there may be potential for wide-area coverage. To achieve this, four monitoring stations have been installed, each equipped with a receiving antenna and sampling equipment, as shown in Figure 1. Monitoring has been carried out in the High Frequency (HF) range of the radio spectrum, since previous studies have suggested that this is the range most likely to yield successful capture of this class of signal [2]. This choice of frequency range has a sound theoretical footing stemming from the fact that the length of the radiating element determines the wavelength of the emissions. Arcs resulting from faults and the radiating structures connected to the arc current path in the power systems equipment itself are of a size that is more compatible with emissions in the HF range. This is why most lightning detection systems operate in the Very Low Frequency (VLF) range. The long-channel, high-current arcs induce VLF and LF emissions readily detectable over great distances. Despite this fact, the equipment has proved able to capture and locate the origin of lightning sferics by utilising their smaller, but numerous components in the HF range [3], as well as successfully capturing emissions emanating from power system arcs. This paper presents a selection of results to date and discusses the challenges that this type of measurement presents. Figure 1: Radiometric Arc Detection System. [1] H.D. Betz, U. Schumann and P. Laroche (Eds.), Lightning: Principles, Instruments and Applications, Springer Science and Business Media, 2009 [2] E. J. Bartlett, M. Vaughan and P. J. Moore, ―Investigations into Electromagnetic Emissions from Power System Arcs‖, IEE Electromagnetic Compatibility Conference , 1999 [3] R. M. Harris, M. D. Judd and P. J. Moore, ―A Novel Approach to Lightning Location: Potential for Reducing Electrical Supply Disruption‖, Proc. International Conference on Gas Discharges and Their Applications, Greifswald, Sept. 2010, pp.534-537 48 UHVnet 2011 B2.6 Voltage Transducer for Transient Measurements on High Voltage Overhead Lines M. F. Hussin*1, A. Haddad1 and N. Harid1 1 Cardiff University, UK *E-mail: [email protected] Overhead lines provide the best economic and practical solution for energy transport but suffer from faults. Overvoltages resulting from these faults can cause degradation or failure of high voltage plant. The main origins for the faults are lightning and switching surges in addition to weather related causes such as heavy wind and ice [1]. Therefore, it is vital that measurement and monitoring of the voltage on HV lines are assessed accurately to maintain safe and economic operation of HV substation and overhead line equipment. Voltage transformers are widely used for measurement and monitoring purposes but they are bulky and costly [2]. Moreover, their frequency response is not adequate for the measurement of fast transient. In this work, a capacitive-type transducer is developed for voltage measurements on HV lines. The transducer uses similar principles as the contactless capacitive probe developed in previous work [3]. However, in this new design, the probe is connected to the HV line instead of the ground. Laboratory experiments are carried out to calibrate the transducer under AC, fast and slow front impulses. An impulse generator is used for the impulse tests while a variac was used for the AC tests. The constructed probe was tested using a number of low-voltage arm capacitance values (2.2nF to 10nF) with the probe installed around a HV conductor above ground. A number of conductor height levels (0.5m to 2.5m) and low-voltage arm capacitance were used to investigate the effect on the output voltage. Figure 1 shows a comparison between the probe output voltage and a standard HV capacitive divider obtained at a height of 2.5m. As can be seen on the figure, good linearity is obtained for all voltage types. Increasing the value of the low-voltage-arm capacitance cause the output to decrease significantly in a non-linear way. This can be particularly useful for the measurement of higher surge voltage. Future tests are planned at the University test overhead line. Figure 1: Output Voltage produced by the transducer [1] M. Muhr, S. Pack and S. Jaufer, ―Usage and Benefit of an Overhead Line Monitoring System‖, International Conference on High Voltage Engineering and Application, November 2008 [2] M. M. Werneck, and A. C. S. Abrantes, ―Fiber-Optic-Based Current and Voltage Measuring System for HighVoltage Distribution Lines‖, IEEE Transactions on Power Delivery, vol. 19, no. 3, July 2004 [3] C. A. Spellman, A. Haddad D.M. German and R. T. Waters, ―Improved Three-Phase Voltage Measurement Using Capacitive Probe‖, Proceeding of the UniversitiesPower engineering Conference, vol. 1, pp.352-355, 1999 49 UHVnet 2011 B2.7 Fault Location using FPGAs and Power Line Communication S. Robson*1, A. Haddad1 and H. Griffiths1 1 Cardiff University, UK *E-mail: [email protected] This work demonstrates a fault location scheme based on the time domain measurement of the fault-induced transient. The combination of high speed analog to digital conversion, GPS time-stamping and customisable triggering control logic is shown to be a possible way of cost effectively resolving transient arrival times to within 30 ns. The simultaneous gathering of the initial transient induced by a fault event can be carried out by several devices scattered across a power network, facilitating the possibility of determining the position of the fault by triangulation, even in branckhed networks. Within the same FPGA chip as the fault detection control logic, an Orthogonal Frequency Division Multiplexing (OFDM) modulator can also be implemented, allowing a means of sending the measured time-stamp information back to an entry point to a wired communication infrastructure, e.g. a SCADA interface at the primary substation. Synchronisation at the receiver is achieved via FPGA hardware implementation of the Schmidl and Cox algorithm [1]. Demonstration and verification of the system has been carried out in a simulation environment. For the OFDM modulator, the various parameters (e.g. number of subcarriers, sampling frequency and cyclic prefix length) have been determined by performance evaluation on a typical rural 11 kV network modelled within the ATP/EMTP environment and MATLAB. Under assumptions of noise levels and a used bandwidth no greater than 300 kHz, it is shown that with a simple time-multiplexed multi-node implementation with a carefully considered choice of OFDM parameters, satisfactory performance can be achieved. The high level block diagram of the system is shown in fig. 1. It is recognised that extensive field trials are required to fully test the proposed concept. Potential difficulties are the non-linear shape of the fault transient (i.e. due to arcing) and ensuring an accurate timestamp. Field trials using coupling capacitors for the modulator coupling to the line and a rogowski coil current transducer are planned. Figure 1: Block diagram of the FPGA based fault locator [1] T.M. Schmidl and D.C. Cox, ―Robust frequency and timing synchronization for OFDM,‖ IEEE Trans. Commun., vol. 45, no.12, pp. 1613-1621, 1997. 50 UHVnet 2011 B2.8 A New Method to Improve the Sensitivity of Leak Detection in SelfContained Fluid-filled Cables L. Hao*1, P. L. Lewin1, S. G. Swingler1 and C. Bradley2 1 2 University of Southampton, UK National Grid, UK *E-mail: [email protected] A method of real-time detection of leaks for self-contained fluid-filled cables without taking them out of service has been assessed and a novel machine learning technique, i.e. support vector regression (SVR) analysis has been investigated to improve the detection sensitivity of the self-contained fluid-filled (FF) cable leaks. The condition of a 400 kV underground FF cable route within the National Grid transmission network has been monitored by Drallim pressure, temperature and load current measurement system. These three measured variables are used as parameters to describe the condition of the cable system. In the regression analysis the temperature and load current of the cable circuit are used as independent variables and the pressure within cables is the dependent variable to be predicted. As a supervised learning algorithm, the SVR requires data with known attributes as training samples in the learning process and can be used to identify unknown data or predict future trends. The load current is an independent variable to the fluid-filled system itself. The temperature, namely the tank temperature is determined by both the load current and the weather condition i.e. ambient temperature. The pressure is directly relevant to the temperature and therefore also correlated to the load current. The Gaussian-RBF kernel: has been used in this investigation as it has a good performance in general application. The SVR algorithm was trained using 4 days data, as shown in Figure 1, and the optimized SVR is used to predict the pressure using the given load current and temperature information. Figure 1: Predicted pressure on training data and measured pressure for cable 1_1_R_A_NC. When compared the predicted pressure and the measured pressure in 6 cables for 3 days, the average error rate ( and the maximum error rate ( , pp-predicted pressure, pm-measured pressure) is between 0.11% and 0.19% ) is between 0.24% and 0.43%. A comparison between the existing system alarm strategy and the alarm sensitivity achieved using the SVR method has also been assessed using equation , where pfa is the falling alarm pressure. The adjusted average error rate is between 0.7% and 1.09% and the maximum error rate is between 1.62% and 2.38%. The results represent that the new machine learning based technique has a 50 times better sensitivity than the existing alarm system. [1] L. Hao, P. L. Lewin, S. G. Swingler and C. Bradley, ―Leak Detection for Self-Contained Fluid-Filled Cables using Regression Analysis‖, IEEE International Symposium on Electrical Insulation, 2010 [2] G. F. Moore, Electrical Cables Handbook, 3rd edition. Blackwell Science, 1997 [3] B. Scholkopf and A. J. Smola, Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond, The MIT Press, 2002 51 UHVnet 2011 B2.9 Energy Harvesting from Electric Fields in Substations for Powering Autonomous Sensors M. Zhu*1 and M. D. Judd1 1 University of Strathclyde, UK *E-mail: min[email protected] The benefits of enhanced condition monitoring in the asset management of the electricity transmission infrastructure can only be fully realised if the sensors used to measure operating parameters are reliable and economical to operate. While adding more sensors can help to track the plant health more accurately, the installation and operating costs of any additional sensors might outweigh the benefits they bring due to the requirement for new cabling or battery maintenance. Advances in wireless communication technology are reducing the power consumption, especially when the average data transfer rate is low [1]. In a high voltage substation, condition monitoring sensors are surrounded by relatively high strength electric and/or magnetic fields. Harvesting energy from the ambient environment therefore becomes a viable alternative for powering the sensor nodes. This would remove the barriers preventing the uptake of wireless sensor networks. Previous studies, which focussed on health and safety issues for utility personnel, recorded typical maximum electric field strengths of about 10 kV/m at 1 m above ground within 400 kV substations [2]. Subsequently, field measurements targeted for potential wireless sensor node installations on the surfaces of substation structures have recorded substantially higher readings. The technical challenge for using a capacitive energy converter to harvest energy from the electric field is similar to the difficulty in developing a piezoelectric energy harvester, where a relatively large voltage is generated with an extremely high source impedance [3]. This study investigates the techniques for improving the harvesting efficiency from 50 Hz electric fields and effectively processing the energy harnessed. The harvested energy is intended for use in powering a new generation of low power consumption wireless sensors. A demonstration device which illustrates the feasibility of harvesting energy from the electric field to power a useful measurement function will be described and possible improvements and challenges will be discussed. [1] E.M. Yeatman ―Advances in Power Sources for Wireless Sensor Nodes‖, Proc. International Workshop on Wearable and Implantable Body Sensor, 2004. [Online] Available at. http://www.doc.ic.ac.uk/vip/bsn\textunderscore2004/program/papers/Eric%20Yeatman.pdf [2] J. Latva-Teikari et al. ―Measuring Occupational Exposure to Electric and Magnetic Fields at 400 kV Substations‖, Proc. Transmission and Distribution Conference and Exposition, pp. 1-4, April 2008 [3] N. Shenck and J. Paradiso, ―Energy Scavenging with Shoe-Mounted Piezoelectrics‖, IEEE Micro, pp. 30-42, MayJune 2001 52 UHVnet 2011 B2.10 Ageing and Temperature Influence on Polarization/Depolarization Current Behaviour of Paper Immersed in Natural Ester Jian Hao*1,2, Ruijin Liao2, G. Chen1 1 2 University of Southampton, UK University of Chongqing, China *E-mail: [email protected] Transformers play an important role in providing a reliable and efficient electricity supply and are one of the most critical equipments in electric power transmission and distribution systems. The most commonly used liquid in power transformers is mineral oil due to its low price and good properties. However the performance of mineral oil starts to be limited due to environmental consideration [1]. Natural ester insulating fluid offers fire safety, environment and insulation aging advantages over mineral oil and are found to be suitable for the use in transformer insulation system [1]. However, transformer owners require to assess the status of the cellulose insulation in transformer non-destructively. Polarization/depolarization Current (PDC) measurement [2] is one of the non-destructive techniques which have been used to achieve this aim. At the present, there are few publications about the PDC behaviour of natural ester-paper insulation, though the natural ester becomes more widely used in transformers. In this paper, the influence of ageing and temperature on the PDC behaviour of the paper immersed in natural ester and mineral oil were compared. Results show PDC technique can be used to assess the aging condition of the natural-ester paper insulation. The ageing and temperature have similar influence on the PDC behaviour of the paper immersed in natural ester and in mineral oil. The depolarization current of paper immersed in natural ester is lower than that immersed in mineral oil at the same test temperature. The depolarization current of the paper immersed in natural ester and mineral oil increase with the aging time increased. Therefore, the depolarization current can be used to indicate the aging status of natural ester-paper insulation. Idep/A 10 10 -8 0 day 31 days 60 days 97 days 123 days -9 Idep/A 10 -10 10 -11 10 -12 10 0 10 1 10 2 10 3 10 4 10 -7 10 -8 10 -9 10 -10 10 -11 10 -12 10 Time/s 0 day 31 days 60 days 97 days 123days 0 10 1 10 2 10 3 10 4 Time/s (a) paper immersed in mineral oil (b) paper immersed in natural ester Figure 1: The depolarization current of paper immersed in natural ester and mineral oil with different aging condition [1] D. Martin, I. Khan, J. Dai, Z. D. Wang, ―An overview of the suitability of vegetable oil dielectrics for use in large power transformers‖, Proceedings of the TJH2b Euro Tech Con, Chester, pp.1-20, 2006 [2] W. S. Zaengl, ―Application of dielectric spectroscopy in time and frequency domain for HV power equipment,‖ IEEE Electr. Insul. Mag., vol. 19, no. 6, pp. 9-22, 2003 53 UHVnet 2011 B2.11 An On-line Lightning Monitoring System for Transmission Lines Bojie Sheng*1, Wenjun Zhou2 , Chengke Zhou1 and Jianhui Yui2 1 2 Glasgow Caledonian University, UK Wuhan University, China *E-mail: [email protected] In order to identify the exact location of lightning strikes on a transmission line, an on-line monitoring system has been developed by the authors comprising front-end lightning detection systems installed on transmission towers, and a remote server running in the control room. The on-line monitoring system can not only obtain the current waveform, the tower number, the exact line and stroke polarity, but also determine the stroke pattern as a result of either lightning shielding failure or back flashover. In order to overcome the difficulties associated with multiple lightning strikes which often occur in quick succession along lightning channel after the main discharge, a novel method is developed to acquire multiple lightning strikes according to their characteristics. The system is deigned to communicate with the remote server automatically by a GPRS wireless module. Finally, the examination results of the effectiveness of the on-line monitoring system, carried out by experiments in a lab. [1] F. Fuchs, E. Ulrich Landers, R. Schmid,et al. ―Lightning Current and Magnetic Field Parameters Caused by Lightning Strikes to Tall Structures Relating to Interference of Electronic Systems‖. IEEE Trans. On EMC,1998 [2] Zhang Qi1in, Qie Xiushu, Kong Xiangzhen. ―Comparative Analysis on Return Stroke Current of Triggered and Natural Lightning Flashes‖. Proceedings of the CSEE, 2007 [3] Wang Hui, Chen Shuiming, He jinliang, Sun weimin, Luo Yi. ―Influence of Lightning Monitoring Device Position on Measured Lightning Current Parameters‖. High Voltage Apparatus. 2009 54 UHVnet 2011 B2.12 Energy Harvesting in Substations for Wireless Sensors and a New Arc Capacitor Structure Jinxin Huang*1,2, Qingmin Li2, Martin D. Judd1 and W. H. Siew1 1 2 University of Strathclyde, UK Shandong University, China *E-mail: [email protected] The past several years have seen an increasing interest in the development of wireless sensor networks. Wireless sensor nodes are usually ‗smart‘, not taking too much space, and their data can be transmitted wirelessly. These advantages make them particularly suitable for certain locations, such as dangerous environments, small spaces or locations where installing cables poses difficulties. However, the problem of powering a large number of nodes in a dense network becomes critical when one considers the prohibitive cost of wiring power to them or replacing batteries. One solution method is energy scavenging technology which can harvest energy from the environment, using sources such as solar, vibration, air flow, temperature difference, ambient electromagnetic fields, etc. An attractive characteristic of this approach is that the lifetime of the node would only be limited by failure of its own components. Equipment in a high voltage substation is surrounded by electric and magnetic fields. Some measurements recorded in ChongQing suggest that typical maximum AC electric field strength in a 500 kV substation is 18 kV/m, therefore the energy of electric field is very rich. And it would be preferable to use the electric field also because it is independent of load current and predictable. Additionally, advances in wireless communication technology have resulted in reduced power consumption of the electronic circuitry required, especially when the average data transfer rate is low due to the lower switching losses. So, harvesting energy from the electric field to power sensor nodes becomes viable. In this paper, the principle of electric field energy harvesting which can change electric field to voltage difference through plane-parallel capacitor is analyzed and a new arc structure of capacitor which reform the traditional plane capacitor plate to arc is proposed. Initial simulation results by Maxwell indicate that the new structure can make the charge induced on the surface of electrode increase by a factor of nearly threefold and can improve the harvesting efficiency. 55 UHVnet 2011 C2 Posters: Materials C2.1 C2.2 C2.3 C2.4 C2.5 C2.6 C2.7 C2.8 C2.9 C2.10 C2.11 On the use of Raman and FTIR Spectroscopy for the Analysis of Silica-based 57 Nanofillers ........................................................................................................................................................... Dielectric Breakdown Strength of Polyethylene Nanocomposites ...................................................................... 58 Influence of Temperature and Moisture Absorbed on Electrical Degradation and 59 Breakdown in Epoxy Resins ............................................................................................................................. Space Charge Behaviour in Oil-Paper Insulation with Different Aging Condition ........................................... 60 Modelling the Non-equilibrium Electric Double Layer at Oil-pressboard Interface of 61 High Voltage Transformers ................................................................................................................................. Investigation of Impulsive Corona Discharges for Energisation of Electrostatic 62 Precipitation Systems .......................................................................................................................................... A Comparison of Polymeric Cable Insulation Properties Following Lightning Impulse 63 Ageing ................................................................................................................................................................. Properties and Analysis of Thermally Aged Poly(ethylene oxide) ..................................................................... 64 Smart Materials as Intelligent Insulation ............................................................................................................ 65 AC Breakdown Characteristics of LDPE in the Presence of Crosslinking By-products..................................... 66 DC Impulse Discharge Degradation of Mica ...................................................................................................... 67 56 UHVnet 2011 C2.1 On the use of Raman and FTIR Spectroscopy for the Analysis of Silica-based Nanofillers C. Yeung*1, G. Gherbaz1 and A. S. Vaughan1 1 University of Southampton, UK *E-mail: [email protected] The potential of polymeric nanocomposites as solid insulation systems has been a topic of great discussion. Although this approach as a means to engineering materials with improved properties has been well established, many of the fundamentals aspects of this class of materials remains poorly understood. For example, the long term dielectric characteristics of so-called nanodielectrics is one of these topics. Whilst the nature of the interfacial region within such systems is believed to be key in determining performance, further investigation is required in order better to understand the macroscopic behaviour of nanocomposites. Such studies are vital for fundamental change, bringing an alternative to conventional polymers and filled composites and so making a massive impact on industry. This paper concerns interfaces in nanodielectrics and sets out to explore the effect of quantified changes in surface functionalisation. Specifically, we have used vibrational spectroscopy to examine and characterise the relevant effects of modifying the surface chemistry of nanosilica with commercial silane methods. Confocal Raman spectroscopy is used to provide qualitative data concerning the functionalization level, whilst Fourier Transform Infrared spectroscopy is used to provide more quantitative data. In this paper, we present the first step towards the design and quantification of nanoparticle surface chemistry - a step that we believe will ultimately allow the interphase to be optimized to meet demanding dielectric requirements. 57 UHVnet 2011 C2.2 Dielectric Breakdown Strength of Polyethylene Nanocomposites K. Y. Lau*1, 2, A. S. Vaughan1 and G. Chen1 1 2 University of Southampton, UK Universiti Teknologi Malaysia, Malaysia *E-mail: [email protected] The term ―nanometric dielectrics‖ or simply ―nanodielectrics‖ was introduced in 1994 when Lewis [1] anticipated the potential property changes that would benefit electrical insulation due to nano-sized inclusion. Such materials, containing homogenous dispersion of small amount (normally less than 10wt%) of nanoparticles (with at least one dimension in nanometre range) in host matrix, are of specific dielectric interest. Although much effort has been put forth to investigate the potential dielectric benefit of such newly emerging materials, many uncertainties remain unanswered, and much remains to be explored [2]. Current experimental work is to investigate the preparation of nanodielectrics via solution blending approach. Polyethylene blend composed of 20wt% of high density polyethylene (HDPE) in low density polyethylene (LDPE) is proposed as the base polymer, with varying content of nanosilica (between 0wt% and 10wt%) as the fillers. Although expensive, solution blending method, when compared with melt compounding method, is expected to provide better dispersion of nanoparticles in polymers, thus providing qualitative data in understanding the behaviour of nanodielectrics [3]. Upon successful preparation of polyethylene nanocomposites, breakdown strength based on ASTM Standard D149-87 is to be conducted to determine the feasibility of such dielectric materials in engineering point of view. Figure 1 illustrates the schematic diagram of the breakdown test configuration. The samples are placed between two 6.3mm diameter steel ball bearings immersed in silicone fluid. AC voltage at a preset ramp rate will be applied until the samples fail and the values of breakdown voltages will be recorded and analysed using two-parameter Weibull distribution. Based upon top-down research approach, the underlying physics and chemistry associated with dielectric property changes will then be explored. Figure 1: Dielectric breakdown test configuration. [1] T. J. Lewis, ‗‗Nanometric Dielectrics‘‘, IEEE Transactions on Dielectrics and Electrical Insulation, vol. 1, no. 15, pp. 812-825, 1994. [2] M. F. Frechette, A. Vijh, L. Utracki, M. L. Trudeau, A. Sami, C. Laurent, P. Morshuis, A. S. Vaughan, E. David, J. Castellon, D. Fabiani, S. Gubanski, J. Kindersberger, C. Reed, A. Krivda, J. Fothergill, F. Guastavino and H. Alamdari, ―Nanodielectrics: A Panacea for Solving All Electrical Insulation Problems?‖ IEEE International Conference on Solid Dielectrics, 2010. [3] T. Tanaka, G. C. Montanari and R. Mülhaupt, ―Polymer Nanocomposites as Dielectrics and Electrical Insulation – Perspectives for Processing Technologies, Material Characterization and Future Applications‖, IEEE Transactions on Dielectrics and Electrical Insulation, vol. 11, no. 5, pp. 763-784, 2004. 58 UHVnet 2011 C2.3 Influence of Temperature and Moisture Absorbed on Electrical Degradation and Breakdown in Epoxy Resins S.J. Dodd1, N.M. Chalashkanov1*, L.A. Dissado1, J.C. Fothergill1 1 University of Leicester, UK *E-mail: [email protected] In the current study, a set of experiments has been carried out to establish the influence of temperature and moisture on the electrical degradation mechanisms, in particular electrical treeing. A set of pin-plane samples were prepared from Araldite CY1311 epoxy resin and conditioned in sealed containers with different levels of relative humidity over the range 15-100%. The treeing experiments were performed over the temperature range 20-70oC. The glass transition temperature of this resin is 0oC, therefore all test were performed above Tg. It has been confirmed that temperature and moisture absorbed in the samples affect the tree growth in similar way. The growth time and the fractal dimension of the trees have been found to decrease with both increasing temperature and moisture concentration (see Fig.1). Pairs of values of temperature and moisture concentration were identified for which the degradation mechanism (electrical treeing) changes to thermal breakdown. Figure 1: Figure caption here, this is an example figure with the correct dimensions. a) b) c) d) Figure 1: Electrical degradation and breakdown at different levels of moisture absorbed, temperature 20 o C, applied voltage 13.5kV rms, pin-plane distance 2mm; a) moisture level less than 0.1%, b) 1.0%, c) 2.4%, d) 6.9% 59 UHVnet 2011 C2.4 Space Charge Behaviour in Oil-Paper Insulation with Different Aging Condition Jian Hao*1, 2, G. Chen2, Ruijin Liao1, Wei Li1 1 2 University of Chongqing, China University of Southampton, UK *E-mail: [email protected] Oil-paper insulation system is widely used in power transformers and cables. The dielectric properties of oilpaper insulation play an important role in the reliable operation of power equipment. Oil-paper insulation degrades under a combined stress of thermal (the most important factor), electrical, mechanical, and chemical stresses during routine operations, which has great effect on the dielectric properties of oil-paper insulation [1]. Space charge in oil-paper insulation has a close relation to its electrical performance [1]. In this paper, space charge behaviour of oil-paper insulation sample with three different ageing conditions (aged for 0, 35 and 77 days) was investigated using the pulsed electroacoustic (PEA) technique. The influence of aging on the space charge dynamics behaviour was analysed. Results show that aging has great effect on the space charge dynamics of oil-paper insulation. The homocharge injection takes place under all three aging conditions above. Positive charges tend to accumulate in the sample, and increase with the oil-paper insulation sample deterioration. The time to achieve the maximum injection charge density is 30s, 2min and 10min for oil-paper insulation sample aged for 0, 35 and 77 days, respectively. The maximum charge density injected in the sample aged for 77 days is more than two times larger than the initial sample. In addition, the charge decay speed becomes much slower with the aging time increase. There is an exponential relationship between the total charge amount and the decay time. The decay time constant increases with the increasing deterioration condition of the oil-paper insulation sample. The value may be used to reflect the aging status of oil-paper insulation. Total Charge Amount (10-7C) 3.0 Aging for 0 day Aging for 35 days Aging for 77 days 2.5 2.0 1.5 1.0 0.5 0.0 0 5 10 15 20 25 30 35 Decay Time (min) a Figure 1: Relationship between total charge amount in the oil-paper insulation sample with different aging condition and decay time Table 1: Fitting parameters of total charge amount (y) and decay time (x) x according to y A Be Constant sample A B 0 day 0.0367 1.1832 0.0715 35 days 0.2054 1.4754 0.1350 77 days 0.3386 2.1976 0.2044 : decay time constant; R2: fitting coefficient R2 0.9968 0.8945 0.9212 [1] Chao Tang, G. Chen, M. Fu, Rui-jin Liao, ―Space charge behavior in multi-layer oil-paper insulation under different DC voltages and temperatures‖, IEEE Transactions on Dielectrics and Electrical Insulation, vol. 17, no. 3, pp. 778-788. 2010 60 UHVnet 2011 C2.5 Modelling the Non-equilibrium Electric Double Layer at Oilpressboard Interface of High Voltage Transformers H. Zainuddin*1, P. L. Lewin1 and P. M. Mitchinson1 University of Southampton, UK *E-mail: [email protected] In large oil-filled power transformers, cellulose-based pressboard and paper are used throughout for electrical insulation. Microscopic views have shown that pressboard insulation is a fibrous and porous structure with non-homogeneous surface. It has been recognised that the pressboard structure is more porous towards the edge [1]. The pores within the pressboard allow oil absorption during impregnation process and provide paths for oil to penetrate until saturation is reached. The ratio of fibre and oil changes as the material structure changes from a medium of bulk oil-pressboard composite toward the bulk oil medium. The porosity of pressboard can also result in impurities within the oil being drawn into the pressboard. It has also been recognised that physicochemical process of a liquid in contact with solid wall leads to the formation of electric double layer (EDL) in the liquid region [2, 3]. The material properties and geometry of pressboard thus lead to a complex oil-pressboard interface. A 2-D model of oil-pressboard interface has been constructed using Comsol Multiphysics Finite Element Analysis software and this is shown in Figure 1. The mathematical model considers the dissociation of a generic impurity in the oil into positive and negative ions and considers the role of the porous and non-homogeneous wall of pressboard in the formation of the EDL. The pressboard, which is represented by different arrays of fibre, promotes preferential adsorption and desorption processes between ions in the oil and unoccupied fibre surfaces of oil impregnated pressboard. The model studies the non-equilibrium charge density profile in the EDL at the oil-pressboard interface when the oil is in the stationary condition. Results of the simulation will be presented in the Colloquium. Figure 1: 2-D model of oil-pressboard interface [1] P.M. Mitchinson, P.L. Lewin, B.D. Strawbridge, P. Jarman, ―Tracking and surface discharge at the oil-pressboard interface‖, IEEE Electrical Insulation Magazine, vol. 26, pp. 35-41, March/April 2010. [2] G.G. Touchard, T.W. Patzek and C.J. Kadke, ―A physicochemical explanation for flow electrification in lowconductivity liquids in contact with a corroding wall‖, IEEE Transactions on Industry Applications, vol. 32, pp. 1015-1057, September/October 1996. [3] A.P. Washabaugh and M. Zahn, ―A chemical reaction-based boundary condition for flow electrification‖, IEEE Transactions on Dielectrics and Electrical Insulation, vol. 4, pp. 688-709, December 1997. 61 UHVnet 2011 C2.6 Investigation of Impulsive Corona Discharges for Energisation of Electrostatic Precipitation Systems A. C. Mermigkas*1, I. V. Timoshkin1, S. J. MacGregor1, M. J. Given1, M. P. Wilson1 and T. Wang1 1 University of Strathclyde, UK *E-mail: [email protected] Various industrial and domestic processes as well as developing nano-technologies generate micron and submicron particles. This phenomenon is more prevalent in large cities where population density and industrial activities are much higher, meaning that a large percentage of the world population is being exposed to everyday inhalation of particulate matter (PM). This may result in negative health effects, many of which are not investigated fully yet [1]. The current research project is focused on the development of a small scale impulsive micro- electrostatic precipitator (IMP) for the removal of PM at homes or in public environments, being small in contrast with the industrial ones. This IMP will implement superimposed DC and sub-microsecond electric fields in order to charge and remove PM efficiently. As the impulse breakdown voltage in a gap is much greater than the DC one, the IMP will also avoid operating close to DC breakdown voltage levels. The designed IMP system composes of a plasma-generation and particle-collection electrodes. For the former, threaded rods of 3 and 6mm have been used as well as a smooth 1.5mm one, while the latter consists of a stainless-steel tube of 28mm internal diameter. The rods were placed coaxially into the tube, with the particle laden air flowing homogenously from the top to the bottom of the reactor. The transmission line based pulse generator developed is able to produce 270ns pulses with frequency of up to 100Hz. The efficiency of precipitation of micron sized particles was evaluated for different DC and impulse voltage levels by measurements of mass of collected particles. Breakdown voltage, corona initiation voltage and parameters of impulse coronas have been obtained under different energisation regimes. Precipitation experimental results showed that the positive or negative charging regimes play an important role in the system efficiency. The ultimate objective of this research project is to investigate precipitation levels of PM2.5, which constitutes a range of lower precipitation efficiency for available ESPs [2], as well as potential microbiological decontamination efficiency of impulsive non-thermal plasmas. [1] J. G. Ayres, ―Long term exposure on air pollution: effect on mortality‖, Committee On the Medical Effects of Air Pollutants, 2009, COMEAP report, pp. 1-4, [Online]. Available: http://www.dh.gov.uk/ab/COMEAP/DH_108151 [2] K. Parker, ―Electrical operation of electrostatic precipitators‖, The Institution of Electrical Engineers, 2003, pp. 16. 62 UHVnet 2011 C2.7 A Comparison of Polymeric Cable Insulation Properties Following Lightning Impulse Ageing N. L. Dao*, P. L. Lewin, I. L. Hosier and S. G. Swingler University of Southampton, UK *E-mail: [email protected] LDPE and HDPE are common materials used within high voltage insulation systems. These materials will be aged after working under high voltage for a long time. The ageing process of these materials may be affected by external factors. The application of repetitive lightning impulse over-voltages is one of these factors and will be considered in this paper. This paper includes the sample preparation process, the ageing of samples under identical conditions and finally the analysis of electrical properties after the ageing process. The obtained results are used to compare the effect of repetitive lightning impulses with these two materials. These results are also used to highlight the possible mechanisms behind the lighting impulse ageing process. [1] S. Boev ―Electric aging of polyethylene in pulsed electric field‖ 12th IEEE International Pulsed Power Conference. (Cat. No.99CH36358) 1999 , Monterey, CA, USA, Page(s): 1365-8 vol.2 [2] G.C. Stone, R.G. Van Heeswijk, R. Bartnikas ―Electrical aging and electroluminescence in epoxy under repetitive voltage surges‖ IEEE transaction on electrical insulation, volume 27, issue 2, April 1992 Page(s):233 – 244 [3] R.A Hartlein, V.S. Harper, Harry Ng ―Effects of voltage surges on extruded dielectric cable life project update‖ IEEE transactions on power delivery, volume 9, issue 2, April 1994 Page(s):611 – 619 63 UHVnet 2011 C2.8 Properties and Analysis of Thermally Aged Poly(ethylene oxide) M. Reading* and A. S. Vaughan University of Southampton, UK *E-mail: [email protected] Recent studies have been performed into the use of polyethylene oxide (PEO) as a model system for observing the fundamental effects of adding micro and nano sized fillers to create polymeric composite systems. Many factors contribute to the successful creation of such a composite system, including dispersion of the filler and treatment of the material during creation. For example, while producing thin films of the materials for testing, high temperatures were used for short periods of time in open air to press the samples into small discs. It is well known that prolonged high temperature exposure can alter the chemistry and structure of polymeric materials and that small variations in the original chemistry, such as longer molecular weights or introduction of fillers, can reduce or possibly accelerate this 'ageing' effect. From these previous investigations many property changes were observed during addition of filler or variation of molecular weight, therefore to accurately attribute these changes to a cause the thermal ageing of the material should be observed. This investigation looks at the same 3 molecular weight PEO systems as those used in the previous investigations and analyses them for their vulnerability to thermal ageing. One thermally aged sample is then taken and tested alongside an unaged sample to observe the effects that the process has on the properties. This includes rheology in solution, differential scanning calorimetry (DSC), AC electrical breakdown, dielectric spectroscopy and fourier transform infra-red (FTIR). By observing the property changes of aged samples it is possible to better understand the thermal ageing process occurring and possibly a way to reduce the effect, along with considering the effect with regard to the behaviour of the previously tested composite samples. 2.0 100k unaged 100k aged 5% aged 100k 5% unaged 100k 10% aged 100k 10% unaged 100k 1.5 Viscosity (Pa.s) Absorbance 0.1 1.0 0.5 0.01 0.0 4000 1 10 100 1000 3000 2000 Wavenumber cm-1 10000 Shear rate (s-1) Figure 1: Viscosity of plots of aged and unaged samples (left) and FTIR spectra of aged and unaged samples (right) 64 1000 UHVnet 2011 C2.9 Smart Materials as Intelligent Insulation A. F. Holt*1, R. C. D. Brown1, P. L. Lewin1, A. S. Vaughan1 and P. Lang2 1 2 University of Southampton, UK EDF Energy Networks Ltd, Crawley, UK *E-mail: [email protected] In order to provide a robust infrastructure for the transmission and distribution of electrical power, understanding and monitoring equipment ageing and failure is of paramount importance. Commonly, failure is associated with degradation of the dielectric material; therefore the introduction of a smart moiety into the material is a potentially attractive means of continual condition monitoring. It is important that any introduction of smart groups into the dielectric does not have any detrimental effect on the desirable electrical and mechanical properties of the bulk material. Initial work focussed on the introduction of fluorophores into a model dielectric system. Fluorescence is known to be a visible effect even at very low concentrations of active fluorophores and therefore was thought well suited to such an application. It was necessary both to optimise the active fluorophore itself and to determine the most appropriate manner in which to introduce the fluorophores into the insulating system. This presentation will describe the effect of introducing fluorophores into polymeric systems on the dielectric properties of the material and the findings thus far [1]. Alternative smart material systems will also be discussed along with the benefits and limitations of smart materials as electric field sensors. [1] A. F. Holt, A. C. Topley, R. C. D. Brown, P. L. Lewin, A. S. Vaughan and P. Lang, ―Towards Intelligent Insulation Technologies‖, Conference on Electrical Insulation and Dielectric 65 UHVnet 2011 C2.10 AC Breakdown Characteristics of LDPE in the Presence of Crosslinking By-products N. Hussin*1, and G. Chen1 1 University of Southampton, UK *E-mail: [email protected] LDPE films of 50µm thick were soaked into crosslinking byproducts which are acetophenone, αmethylstyrene and cumyl alcohol. The samples were used to perform the breakdown strength (Eb) of the LDPE with the byproducts chemical reside in the sample. The AC breakdown measurements were conducted at a ramp rate of 50V/s at room temperature. Weibull plot is used to analyse the ac breakdown result. Comparing the soaked and un-soaked (fresh LDPE) samples, it does show a small reduction of the eta values as the LDPE films were soaked into the sample. It suggests that the breakdown strength is reduced by adding the byproducts in the LDPE film. However, as the range of breakdown strength of all samples are to be compared, these values fall in the same region which indicate no significant difference can be seen in all samples. Weibull Breakdown Probability (%) 99.0 95.0 70.0 50.0 20.0 10.0 5.0 1.0 0.5 Clean LDPE confidence Bound 0.1 100 150 200 250 300 350 Breakdown Stress (kV/mm) Figure 1: The Eb plot of clean LDPE [1] T. Andrews, R.N. Hampton , A. Smedberg, D. Waldm, V. Waschk, and W. Weissenberg. ―The Role of Degassing in XLPE Power Cable Manufacture‖, IEEE Electrical Insulation Magazine December 2006 5-16. [2] N. Amyot, S.Y Lee, E. David, I. H. Lee. “Effect of residual crosslinking by-products on the local dielectric strength of HV extruded cables.‖ 2000 Annual Report Conference on Electrical Insulation and Dielectric Phenomena, Oct. 2000 Volume 2(No 2):743 – 746 [3] N. Hussin and G. Chen, ― Space Charge Accumulation and Conductivity of Crosslinking Byproducts Soaked LDPE‖. In: 2010 Conference on Electrical Insulation and Dielectric Phenomena, 17 - 20 October 2010, Purdue University, West Lafayette, Indiana, USA. pp. 125-128 66 UHVnet 2011 C2.11 DC Impulse Discharge Degradation of Mica J Paterson, A J Shields, D M Hepburn, C Zhou Glasgow Caledonian University , UK E-mail: [email protected] Mica is a vital component of the insulation systems of stator bars in High Voltage (HV) machines. Its inherent resistance to electrical discharge activity has resulted in its use for HV insulation for many years, however, the degradation of insulation containing mica is not yet fully understood. Difficulties associated with laboratory aging of mica, principally the lengthy time scales required in AC partial discharge (PD) stressing before damage occurs, has led to a more aggressive electrical stressing strategy. DC Impulse testing has the capability to reduce the time required to initiate onset of damage in mica. The transient nature and high energy delivered to a mica sample during an impulse discharge results in physical damage that is observable by traditional and advanced microscopy techniques. Material damage, such as localised dislocation and fracture of the mica surface and melting at the impact site, can be analysed. The energy of DC impulse stressing can be readily calculated enabling correlation between energy input and type and severity of damage. DC impulse testing, carried out in accordance with BS 923 1/50µs, produces damage to mica by a variety of mechanisms. The very high temperature associated with the plasma discharge causes thermal damage that manifests itself as localised melting. Electrical stress causes tracks to form on the mica surface [1] and to cause bond cleavage of mica [2] and in polymer/mica composites [3]. The work presented in this paper addresses the issue that micaceous insulation used in high power rotating plant operates at elevated temperatures. Use of thermally stressed samples in DC impulse testing allows assessment of variation in mica degradation resulting from prior thermal exposure. Comparing thermally stressed with unstressed mica samples will help quantify the contribution thermal stress has on the degradation of mica. Mica samples exposed to elevated temperature for extended periods were stressed using DC impulses and optical investigation of the material carried out. This work aims to better identify the structural and chemical changes occurring in mica under electrical and thermal stress. Although the degradation of mica has previously been investigated, e.g. [4, 5, 6,], the studies occurred prior to the micro- and nano-scale physical and chemical techniques of the present age revealing previously hidden detail. The changes identified and correlation with energy input will be reported. [1] Shields A.J and Kemp I.J, ―Impulse discharge erosion and breakdown of mica‖, Ph.D. thesis, 1980, p186 [2] Shields A.J and Kemp I.J, Degradation and breakdown of mica under partial discharge stressing, IEE Proc. Science, Measurement & Technology, May 2000, Vol 147, pp105 – 109, ISSN 1350-2344 [3] Jia Z.D., Hao Y.P. and Xie H.K., ‗The degradation Assessment of epoxy/mica insulation under multi stress aging‘, IEEE Trans. on Dielectrics and Electrical Insulation Vol. 13 No 1, Feb 2006, pages?. [4] Jia Z.D and Xie H.K, ‗The change of microstructure of epoxy mica insulation in multi-stress aging.‘ 973772searchabstractProceedings of 2001 International Symposium on Electrical Insulating materials, pp 693 – 696 [5] Davidson A.T. and Yoffe A.D., ―Dielectric breakdown in thin mica crystals‖, Nature, 1965, p.1247 [6] Silk E.C.H. and Barnes, R.S., ―The observation of dislocations in mica‖, Acta. Metall., 1961, Vol.9, p558-563 67 UHVnet 2011 D2 Posters: Theories, Methods and Models D2.1 D2.2 D2.3 D2.4 D2.5 D2.6 D2.7 Modelling of Partial Discharge Activity in Cavity within a Dielectric Insulation Material ............................... 69 Full Wave Modelling of Partial Discharge Phenomena in Power Transformers using FDTD Methods ................................................................................................................................................... 70 Evaluation of an Iterative Method used for Partial Discharge RF Location Techniques.................................... 71 Numerical Modelling of Needle-Grid Electrodes Negative Surface Corona Charge System ................................................................................................................................................................. 72 Mathematical Modelling of End-of-Life of Power Transformers in Perspective of System Reliability ............................................................................................................................................................ 73 A Comparison between Electroluminescence Models and Experimental Results .............................................. 74 An Improved Pulsed Electroacoustic System for Space Charge Measurement under AC Conditions ........................................................................................................................................................... 75 68 UHVnet 2011 D2.1 Modelling of Partial Discharge Activity in Cavity within a Dielectric Insulation Material Tianyu Bai*1, D. J. Swaffield1 and P. L. Lewin 1 1 University of Southampton, UK *E-mail: [email protected] The pattern of partial discharge（PD）occurrence at a defect site within a solid dielectric material is influenced by the conditions of the defect site. This is because the defect conditions such as its size and location determine the electric field distributions at the defect site which influence the patterns of PD occurrence. A model for a spherical cavity and ellipsoidal cavity within a homogeneous dielectric material has been developed by using Finite Element Analysis (FEA) software. The model is used to study the influence of different conditions of the cavity on the electric field distribution in the cavity and the PD activity. Also, experimental measurements of PD in spherical cavity and ellipsoidal cavity of different size within a dielectric material will be displayed. At present, only the model for the PD has been prepared. The cavity in the silicon rubber is now produced. The pre-cure time , post-cure time and temperature for the product of cavity are being determined by researchers now. And the bubbles in the small sample with silicon rubber are easier to be made than in the bigger sample. Once cavity in silicon rubber is successful to be made, PD measurement will proceed. Figure 1 shows the schematic diagram of the object test for the research. Figure 1: schematic diagram of the object test 69 UHVnet 2011 D2.2 Full Wave Modelling of Partial Discharge Phenomena in Power Transformers using FDTD Methods A. M. Ishak*1,2, M. D. Judd1 and W. H. Siew1 1 2 University of Strathclyde, UK National Defence University of Malaysia, Malaysia *E-mail: [email protected] The occurrence of defects in the insulation system of power transformers will degrade the insulation properties and compromise the life of the equipment. The consequences of unexpected failure can be costly and replacement units may not be readily available. Small electrical sparks which are known as partial discharge (PD) are usually being produced at the degraded location. PD is a localized electrical discharge that only partially bridges the insulation between conductors. Although the magnitude of the discharge is usually small at first, it can cause progressive deterioration with time and then lead to massive failure of the equipment. Therefore, it is important to detect and monitor the occurrence of PD by using non-destructive test equipment. PD signals can be detected in the ultra high frequency (UHF) band (300-3000 MHz) because the short current pulses radiate a wideband electromagnetic transient. The advantages of UHF approach are wide detection range, high sensitivity and good immunity to interference signals during on-site measurements. Once a PD pulse occurs, electromagnetic waves propagate in all directions from the PD source. Different materials impose different propagation properties on the radiated waves. Finite-difference time-domain (FDTD) is a method to model electromagnetic wave propagation and interactions with the structure of materials. Provided the results can be validated, modelling full wave PD phenomena by using the FDTD method will be much safer than attempting it with high voltage experiments. Furthermore, it will be much easier to change parameters in the model and study their influence on the PD detection, thereby contributing to your understanding and ability to interpret measurements. This paper describes the modelling of PD current sources, the modelling of UHF sensors, and demonstrates by way of example a method to estimate the location of a PD source by measuring the arrival time of UHF PD signals at several sensors. 70 UHVnet 2011 D2.3 Evaluation of an Iterative Method used for Partial Discharge RF Location Techniques O. El Mountassir*1, B.G. Stewart1, S. G. McMeekin1 and A. Ahmadinia1 1 Glasgow Caledonian University , UK *E-mail: [email protected] To date many experimental studies have succeeded in determining the location of partial discharges (PD) in 3 dimensions using time difference of arrival technique and iterative algorithms (e.g. Moore et al. 2005, Stewart et al. 2009). In order to locate PDs accurately, many studies have focused on analyzing the propagation of the PDs electromagnetic waves (e.g. Kawada, 2009). Despite the fact that Radio Frequency (RF) technique is being heavily investigated, few studies have attempted to study the performance of a number of iterative methods and their influence in accurately determining the locations of PDs. This paper evaluates the performance of the Hyperbola Least Square algorithm (HLS) to locate different PDs. A software platform was developed for the simulation and localisation of a range of PDs using a Y shaped arrangement of four antennas. The HLS algorithm was applied and evaluated using (0, 0, 0) as the initial values, with different error bounds on the iteration method evaluated. The performance of the algorithm was compared in terms of location accuracy and also computing efficiency (Figure1 shows the iteration efficiency for 64 PD points). The paper will also look at location accuracy. The results obtained by this algorithm were different in terms of (i) accuracy and (ii) number of iterations. Both were also found to be highly dependent on the selected error bounds. The error bounds and the maximum number of iterations needed in order to provide the best results have been quantified and compared to each other. This is an important result as it shows that the accuracy of an algorithm can in some way be evaluated for PD location despite the non-linear nature of the triangulation equations. Figure 1: Number of iterations for HLS using Y shaped antennas arrangements. [1] P. J. Moore, P. I. Portugués and I. Glover, ―Radiometric Location of Partial Discharge Sources on Energized HighVoltage Plant‖, IEEE Transactions on Power Delivery, vol. 20, no. 3, 2005 [2] B. G. Stewart, A. Nesbitt and L. Hall, ―Triangulation and 3D Location Estimation of RFI and Partial Discharge Sources Within a 400kV Substation‖, Proceedings of the IEEE 29 th Electrical Insulation Conference, pp.164-168, 2009 [3] Y. Tian, P. M. Kawada and K. Isaka, ―Locating Partial Discharge Source Occurring on Distribution Line by using FDTD and TDOA Methods‖, IEEJ Transactions on Fundamentals and Materials, 2009 71 UHVnet 2011 D2.4 Numerical Modelling of Needle-Grid Electrodes Negative Surface Corona Charge System Y. Zhuang*1, G. Chen1 and M. Rotaru1 1 University of Southampton, UK *E-mail: [email protected] Surface potential decay measurement is a simple and low cost tool to examine electrical properties of insulation materials. During corona charging stage, needle-grid electrodes system is often used to achieve uniform charge distribution on the surface of the sample. However, there is little report on the effects of geometrical parameters and voltage values of the charging system on the surface potential and its characteristics. In the present report simulations based on gas discharging physics similar to [1] have been carried out to investigate dynamic surface charge formation. The geometry of in the model includes a 174µm radius of curvature needle setting perpendicular to a 0.5mm thickness grid electrode and a 27.5mm diameter with 50µm thickness polyethylene. The bottom surface of the polyethylene is grounded and it is 3cm and 1.5cm away from the needle electrode and grid electrode respectively. The simulations were initially performed under the following conditions: the needle electrode was set as -8000V and the grid electrode -2000V. It has been found that an impulse current appeared after 0.3µs charging which represented the corona effect. The effect of adding a grid electrode can be clearly seen from the logarithmic plot of electrons. Finally, surface charge density on the sample has been obtained. [1] T. N. Tran, I. O. Golosnoy, P. L. Lewin and G. E. Georghiou, ―Two Dimensional Studies of Trichel Pulses in Air Using the Finite Element Method‖, 2009 IEEE Conference on Electrical Insulation and Dielectric Phenomena, 1821 October 2009 72 UHVnet 2011 D2.5 Mathematical Modelling of End-of-Life of Power Transformers in Perspective of System Reliability B. Patel1, Z.D. Wang1*, J.V. Milanovic1 and P. Jarman2 1 2 University of Manchester, UK National Grid, UK * E-mail: [email protected] Never before have the consequences of loss of supply been so great, due to the fact that electricity is now a vital part of our everyday lives. The importance of security of supply is also reflected by the financial penalties incurred onto transmission and distribution companies for not maintaining supply and the immeasurable damage this can cause to their public image. Transformers form a vital part of any power system and the UK transmission system contains nearly one thousand power transformers. Transformers are often the most expensive pieces of equipment in any transmission system. Because of the number of transformers, the capital cost of transformer assets and the financial consequences of failure it is necessary for transmission companies to have short term and medium/long term asset replacement plans in order to effectively maintain security of supply. Capital investment planning is a vital process for any business. Because of these factors mentioned above there has been a shift to condition based asset replacement programmes by transmission companies, in which assets are replaced based on condition as opposed to age. In order for such programmes to be successful a deep coherent understanding of the ageing, deterioration and failure of transformers is required. In this project the general methodology on how to create a mathematical model of transformer failure is discussed in order to determine how the probability of transformer failure is affected due to a series of operating scenarios and phenomena such as overloads, short circuits, lightning strikes and etc. The current technical debate focuses on two distinctively different approaches on choosing parameters based on which a mathematical transformer failure probability model is built: one is based on the design information and operating history and the other one is based on condition. After the mathematical transformer failure probability model is built, the subsequent effect on system reliability can then be calculated which takes account of transformer age, transformer condition and system operating conditions. A demonstration case will be given using a test system consisting of three transformers operating in parallel and by utilising a variety of statistical tools including Monte Carlo simulations and Markov models. 73 UHVnet 2011 D2.6 A Comparison between Electroluminescence Models and Experimental Results D. H. Mills*1, F. Baudoin2, P. L. Lewin1 and G. Chen1 1 2 University of Southampton, UK University of Toulouse, France *E-mail: [email protected] Electrical insulation ages and degrades until its eventual failure under electrical stress. One cause of this relates to the movement and accumulation of charge within the insulation. The emission of a low level of light from polymeric materials while under electrical stressing occurs before the onset of currently detectable material degradation. This light is known as electroluminescence (EL) and under an ac electric field is thought to relate to the interaction of charge in close proximity to the electrode-polymer interface. Understanding the cause of this light emission gives a very high-resolution method of monitoring charge interaction and its influence on material ageing. A possible cause of this light emission is the bipolar charge recombination theory. This theory involves the injection, trapping and recombination of charge carriers during each half cycle of the applied field [1]. This work compares two models that to simulate the EL emission according to this bipolar charge recombination theory. Model 1 assumes a fixed space charge region and all injected charge is uniformly distributed in this region with charges able to either become trapped or to recombine with opposite polarity charge carriers [2]. This recombination relates directly the excitation needed for the emission of a photon of light as measured in experiments. Model 2 develops on this by accounting for the transport and extraction of charge with an exponential distribution of trap levels rather than a uniform distribution [3]. Figure 1 shows a good correlation between the two models and experimental data. The full paper will describe the models in more detail and present results comparing the simulated and experimental results under various applied waveforms. Model 1 and model 2 both provide a good correlation with experimental data but model 2 allows a greater understanding of the space charge profile in the region close to the electrodes as well as the shape of the conduction current. Further work involves developing these models to support changes in the charge trapping profiles due to material ageing and supporting simulated results with measured conduction current. Figure 1: Phase resolved electroluminescence emission and simulation results under a 50Hz sinusoidal 6kVpk applied field [1] A. Mohd Ariffin and P. L. Lewin, ―Phase-Resolved Measurement and Modelling of Electroluminescence Phenomenon in Polyethylene Subjected to High Electrical Stress‖, International Conference on Condtion Monitoring and Diagnosis, 2008 [2] A. Mohd Ariffin, N. Mat Tajudin, S. Sulaiman, Y. H. M. Thayoob and P. L. Lewin, ―Comparing Simulation Results and Experimental Measurements of Electroluminescence Phenomenon in Dielectric Materials‖, IEEE International Symposium on Electrical Insulation, 2010 [3] F. Baudoin, D. H. Mills, P. L. Lewin, S. Le Roy, G. Teyssedre and C. Laurent, ―Contribution to the Modelling of Electroluminescence in High Voltage Polymeric Materials‖, Conference on Electrical Insulation and Dielectric Phenomena, 2010 74 UHVnet 2011 D2.7 An Improved Pulsed Electroacoustic System for Space Charge Measurement under AC Conditions Z. Xu*1, J. Zhao1 and G. Chen1 1 University of Southampton, UK * Email: [email protected] In this paper, an improved space charge measurement system based on the pulsed electro acoustic technique (PEA) is presented. The new system gives an essential way to examine the role of space charge in electrical aging process under AC conditions. The system setup for AC measurement is presented and detailed in this paper with comparison to the old system. There are two features with improved PEA system. A pulse generator with a 3 kHz repetition rate is utilized to reduce the measurement time. The Eclipse data acquisition system is used to achieve the high data acquisition rate. The results which were taken from both old and new PEA system show that hetero-charge can be formed in the region close to the lower electrode under AC electric field. Apparently the results captured from the new system have better phase resolution than the old system. The space charge decay profile measured by the new system can reflect vividly on the charge dynamically changing. The utmost space charge information was saved as the measurement time was dramatically shortened by the improved system. Figure 1: Space charge profile udner AC condition, f=1Hz, every 15° in one cycle. [1] C. Laurent, G. Teyssedre, G. C. Montanari, ―Time-resolved space charge and electroluminescence measurement in polyethylene Under ac stress,‖ IEEE Trans. Diel. Electr. Insul., vol. 11, pp554-560, 2004 [2] Wu X, Chen G, Davies A E, Tanaka Y and Sutton S J, ―Space charge measurements in polyethylene under dc and ac operating conditions using the PEA technique,‖ IEE DMMA conf. pub473, 2000, pp 57–62. [3] C. Thomas, G. Teyssedre and C. Laurent, ―A new method for signal averaging resorting to space charge measurement by the pulsed electro-acoustic method under AC stress,‖ IEEE ICSD, Winchester, UK, 2007, pp 490493 75 UHVnet 2011 Authors Index Abd Rahman, M. S. ............................................31 Abdelmalik, A. A................................................20 Abubakar Mas‘ud, A. .........................................46 Adhikari, D. ........................................................19 Ahmadinia, A......................................................71 Ahmeda, M. ........................................................41 Albano, M. ......................................................4, 36 Allison, F. ...........................................................14 Bai, T. .................................................................69 Baker, P. C. .........................................................10 Baudoin, F. .........................................................74 Bogias, A. C. ......................................................39 Bradley, C. ..........................................................51 Braini, S. .............................................................40 Brown, R. C. D. ..................................................65 Chalashkanov, N. M. ..........................................59 Charalampidis, P.................................................17 Chen, G. .............. 21, 28, 53, 58, 60, 66, 72, 74, 75 Chippendale, R. D. .............................................24 Cotton, I. .............................................................14 Coventry, P. ....................................................4, 36 Dao, N. L. ...........................................................63 Dissado, L. A. .....................................................59 Dodd, S. J. ..............................................20, 23, 59 Efika, I. B. ............................................................5 El Mountassir, O. ................................................71 Elmghairbi, A. ....................................................37 Feng, D. Y. .........................................................30 Fothergill, J. C. .............................................20, 59 Freebody, N. A. ..................................................18 Gherbaz, G. .........................................................57 Given, M. J. ........................................................62 Goddard, K. F. ......................................................7 Golosnoy, I. O. ...................................................24 Green, P. R. ........................................................14 Griffiths, H. .................. 2, 4, 17, 26, 37, 41, 42, 50 Haddad, A. ..... 4, 17, 26, 33, 36, 37, 39, 40, 41, 42, 49, 50 Hao, J. ...........................................................53, 60 Hao, L. ....................................................11, 31, 51 Harid, N. ................... 17, 33, 36, 37, 39, 41, 42, 49 Harris, R. M. .......................................................48 Hepburn, D. M. .............................................19, 47 Holt, A. F. ...........................................................65 Hosier, I. L. ...................................................16, 63 Hu, X.................................................................. 44 Huang, J. ............................................................ 55 Hunter, J. A. ....................................................... 11 Hussin, M. F. ..................................................... 49 Hussin, N. .......................................................... 66 Illias, H. A.......................................................... 21 Ishak, A. M. ....................................................... 70 Jarman, P...................................................... 30, 73 Ji, T. Y. .............................................................. 32 Judd, M. D. ................ 9, 10, 12, 44, 48, 52, 55, 70 Kamarudin, M. S................................................ 36 Lambert, J. ......................................................... 24 Lang, P. .............................................................. 65 Lau, K. Y. .......................................................... 58 Lewin, P. L.6, 7, 11, 21, 24, 28, 31, 51, 61, 63, 65, 69, 74 Li, C. R. ............................................................. 45 Li, H. Y. ............................................................. 25 Li, Qi. ........................................................... 27, 34 Li, Qingmin........................................................ 55 Li, W. ................................................................. 60 Liao, R. ........................................................ 53, 60 Long, C. ............................................................. 38 MacGregor, S. J. ................................................ 62 Mair, A. J. .......................................................... 10 McArthur, S. D. J. .............................................. 10 McMeekin, S. G. ...................................... 3, 46, 71 Mermigkas, A. C. ............................................... 62 Michel, M. ......................................................... 11 Milanovic, J. V................................................... 73 Mills, D. H. .................................................. 28, 74 Mitchinson, P. M. .............................................. 61 Mohamed, F. P. .................................................. 13 Moore, F. ........................................................... 26 Moore, P. J. ........................................................ 48 Mousa, S. ........................................................... 42 Murugan, G. S.................................................... 24 Nekeb, A. S. ....................................................... 33 Nesbitt, A. .......................................................... 46 Patel, B. .............................................................. 73 Payne, D. .............................................................. 6 Peng, X. ............................................................. 47 Pilgrim, J. A. ........................................................ 6 Reading, M. ....................................................... 64 Reid, A. J. .......................................................... 44 76 UHVnet 2011 Robson, S. .......................................................... 50 Roscoe, N. M. ..................................................... 12 Rotaru, M. .......................................................... 72 Rowland, S. M. ....................................... 14, 27, 34 Sarkar, P. ............................................................ 17 Sheng, B. ............................................................ 54 Shuttleworth, R............................................. 27, 34 Siew, W. H. ................................ 13, 35, 44, 55, 70 Smith, D. J. ........................................................... 3 Song, X. .............................................................. 47 Soraghan, J. J. ............................................... 13, 35 Stewart, B. G. ..................................... 3, 19, 46, 71 Strachan, S. S. .................................................... 13 Swaffield, D. J. ......................................... 6, 11, 69 Swingler, S. G. ................................... 7, 16, 51, 63 Tang, W. H. ........................................................ 32 Tao, Y. ................................................................ 35 Timoshkin, I. V. ................................................. 62 Vaughan, A. S. ..................... 16, 18, 57, 58, 64, 65 Veerappan, C. A. ................................................ 14 Wallace, P. A. ....................................................... 3 Walton, C. .......................................................... 11 Wang, P. ...............................................................7 Wang, T. .............................................................62 Wang, Z. D. ...................................... 25, 30, 45, 73 Waters, R. T. .......................................................17 Wei, C. H. ...........................................................32 Wilson, M. P. ......................................................62 Wu, Q. H.............................................................32 Xu, Z. ..................................................................75 Yeung, C. ............................................................57 Yu, J....................................................................38 Yui, J...................................................................54 Zachariades, C. ...................................................14 Zainuddin, H. ......................................................61 Zhang, G. ............................................................34 Zhang, L. ..............................................................5 Zhang, R. ............................................................25 Zhao, J. .........................................................28, 75 Zhou, C. ..................................................38, 47, 54 Zhou, D. ..............................................................45 Zhou, W. .......................................................38, 54 Zhu, M. ...............................................................52 Zhuang, Y. ..........................................................72 77
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
advertisement
© 2021