Edinburgh Napier University

  Partial discharge (PD) activities are the essential contributor of insulation failure, which constitute more than 60% for damaging high voltage equipment. This article presents a critical review on the modelling of partial discharge using different techniques. Monitoring partial discharge activities is essential to enhance the comprehension on the phenomena of partial discharge through modelling process. A lot of tools have been introduced to assist PD measurement in diagnosing and monitoring the condition of high voltage insulation system. This article will evaluate the development of important PD models such as three-capacitance model, Pedersen's model, Finite Element model (FEA), and Niemeyer's model. This review attested that the research studies done on partial discharge are still deficient and there are still countless of study on partial discharge to be carried out. The parameters that have influence on partial discharge occurrence inside cavities in solid dielectric materials are statistical time lag, inception field, effective charge decay time constant, and cavity surface time constant. The initial free electron required for the existence of a PD is provided by the electron surface emission and volume ionization with the cavity while applied voltage's amplitude and frequency are the stress conditions affecting PD behavior in cavities as well as cavity's size, shape, and location. A Comparison regarding modeling time and accuracy of the model between different models of partial discharge have been done in this article. From where it can be concluded that the FEM and Niemeyer's model are suitable for PD due to their average cost and low degree of application, but FEM is commonly used because it is more accurate and easiest model. Moreover, from a comparison of advantages and limitations of different models, it is shown that FEM model has one major limitation that the large amount of data is required during meshing process. Besides, Niemeyer's model is not accurate due to the fundamental assumptions made and these assumptions are not clearly justified with high reliance over immeasurable free parameters.