Search

Electric power supply involves three main stages which are generation, transmission and distribution. Properly maintained electrical equipment at each stage can assure the supply is continuous at all time to customers and it is important for safety requirements where fire and explosion accidents can be eliminated. The electrical equipment includes generators, motors, power transformers and switchgear. Electrical discharges (small sparks) which usually occur in defective insulation system of electrical equipment can cause equipment failure or even explosion if the discharges are not detected and treated in an appropriate time. The discharges generate electromagnetic waves which can be detected using ultra-high frequency (UHF) sensors. A lot of research have been done to design effective and low cost UHF sensors for different applications in high voltage systems. UHF sensor manufacturers are required to meet certain sensitivity standards so that the sensors are able to detect the mini mum level of discharges and at certain frequency range. Manufacturers will fabricate a new sensor and then measure the sensitivity of the sensor using a calibration system. If the sensor does not achieve the minimum sensitivity standard, manufacturers will fabricate another one after changing certain parameters of the sensor that could improve the response and then the sensor will be characterised again. This repetitive and expensive process can be eliminated using numerical electromagnetic simulation to design and calibrate the sensor. Finite-difference time-domain (FDTD) method is a computational technique that can be used to design UHF sensors and predict the sensitivity of the sensors. In this study, a calibration system was modelled using FDTD technique to predict the responses of several existing UHF partial discharge sensors. The simulated calibration system was simplified to overcome the complex design of the experimental calibration system. However, the main function of the system was retained in the simulation. The performance of the simulated calibration system was validated by comparing the experimental responses of physical sensors with the simulated results. The percentage difference between the measured and simulated responses of the existing sensors is 7.65% on absolute average with standard deviation of 2.99. The factors that affect the responses of the sensors for examples, the electrical property of insulating materials and sensor sizes have been studied using this method. A new UHF partial discharge sensor was designed and modelled using FDTD simulation and then fabricated for testing. The measured and simulated responses of the sensor showed good agreement. Design variations of the same sensor were simulated to improve the sensor response. Then, the optimum simulated response and simplest design to manufacture was chosen as the final design of the sensor. The experimental and simulated results of the sensor also showed excellent agreement. The FDTD method can also study the characteristics of electromagnetic wave propagation generated by PD source in power transformer.