Evaluation of FDTD modelling as a tool for predicting the response of UHF partial discharge sensors

Ultra high frequency (UHF) partial discharge sensors are important tools for condition monitoring and fault location in high voltage equipment. There are many designs of UHF sensors which can detect electromagnetic waves that radiate from partial discharge sources. The general types of UHF PD sensors are disc, monopole, probe, spiral, and conical types with each type of sensor having different characteristics and applications. Computational modelling of UHF PD sensors using Finite-difference time-domain (FDTD) simulation can simplify the process of sensor design and optimisation, reducing the development cost for repeated testing (in order to select the best materials and designs for the sensors), and giving greater insight into how the mechanical design and mounting will influence frequency response. This paper reports on an investigation into the application of FDTD methods in modelling and calibrating UHF PD sensors. This paper focuses on the disc-type sensor where the sensor has been modelled in software and the predicted responses are compared with experimental measurements. Results indicate that the FDTD method can accurately predict the output voltages and frequency responses of disc-type sensors. FDTD simulation can reduce reliance upon costly experimental sensor prototypes, leading to quicker assessment of design concepts, improved capabilities and reduced development costs.