Partial Discharge Type Detection utilizing Statistical Techniques (n-q) and Random Forest Method

Partial Discharge (PD) designs are critical instrument for the findings of high voltage (HV) protection frameworks. Human specialists can find conceivable protection absconds in different portrayals of the PD information. One of the most broadly utilized portrayals is Phase-Resove,d PD (PRPD) designs. So as to guarantee the dependable activity of H.V hardware, it is vit,al to rela,te the noticeable measurable attributes of P.Ds to t,he prope,rties of the imperfection and at last to decide the kind of the deformity. In present work, we have obtained and analyzed PRPD pattern (n-q) using statistical parameters such as mean, standard deviation, variance, skew-ness and kurtosis to detect type of PD & we have verified the obtained results by providing obtained statistical para-meters as an in-put for training of Artificial Neural Net-work (ANN) in Google colaboratory using Python for Random Forest Method to detect type of discharge such as either void, surface or corona.