BREAKDOWN VOLTAGE TEST OF DIFFERENT SOLID INSULATING MATERIALS USING ARTIFICIAL NEURAL NETWORK MODEL

Abstract

In this paper  we are presenting  Artificial  Neural  Network (MFNN) model which   has different  possible  inputs  affecting  the  breakdown voltage  that  are working  temperature, the  insulating  material thickness, dielectric  strength  of insulating material,  volume  resistivity of materials, dissipation factor, conductivity of  materials, and  the  materials  relative  permittivity  that also  predicts  the  breakdown voltage  as  a  function  of all these  inputs parameters.  It is important to train the Artificial Neural Network to determine the Breakdown Voltage as close as possible. For the purpose of training of neural network model,   the data which are obtained by performing the experiment are used. After  the  completion  of  training  of artificial neural  network  model,  the  breakdown voltages  as  a  function  of  the all these input  parameters  are  determined.  We can also predict the effect regarding thickness of insulating materials, moisture content, temperature and various environment conditions and parameters on solid insulating materials.  Necessity and used of MFNN model is also discuss in this paper with taking references of different research papers.  Important of back propagation technique to train MFNN is also discussing.  Different causes of breakdown and detonations of different solid insulating materials also focused in this paper. Software MATLAB 2010 is used for designed, trained and tested of the ANN.