Modelling of Partial Discharge Inception and Extinction Voltages Using Adaptive Neuro-Fuzzy Inference System (ANFIS)

In this paper is presented an adaptive neuro- fuzzy inference system (ANFIS) that is used for modeling of partial discharge inception and extinction voltages. The ANFIS structure is automatically generated and tuned in order to fit the available measurement data. As inputs of the adaptive neural network are used dielectric thickness, void depth and void diameter. The voids in the solid insulating materials are artificially created. Finally, estimation of the model error is given.