Artificial neural networks modelling of breakdown voltage of solid insulating materials in the presence of void

A major field of artificial neural networks (ANN) application is function estimation because of its useful properties, such as non-linearity and adaptivity particularly when the equation describing the function is unknown. In this study, the partial discharges (PD) breakdown voltage of five insulating materials under AC conditions has been predicted as a function of four input parameters, such as the thickness of the insulating sample t, the thickness of the void t₁, diameter of the void d and relative permittivity of materials ∈_r by using two different ANN models. The requisite training data are obtained from experimental studies performed on a cylinder-plane electrode system. The voids are artificially created with different dimensions. Detailed studies have been carried out to determine the ANN parameters which give the best results. Studies have also been carried out to assess the extrapolation capabilities of the networks considered here. On completion of training, it is found that the ANN models are capable of predicting the breakdown voltage V_b= f(t, t₁, d, ∈_r) very efficiently and within a small value of mean absolute error with the multi-layer feedforward neural network (MFNN) model marginally better than the radial basis function network (RBFN) model.