PD recognition by means of statistical and fractal parameters and a neuralnetwork

A novel partial discharge (pn) defect identification method is described. S tarting with PD data on different families of specimens, a suitable set of parameters are determined and then used as input variables to a neural netw ork for the purpose of identifying the defects within the insulation. In th is procedure the statistical Weibull analysis is performed on PD pulse ampl itude histograms to obtain the scale parameter cr and the shape parameter b eta. Thereafter, the two statistical operators (skewness and kurtosis) and two fractal parameters (fractal dimension and lacunarity) are evaluated. fr om the FD phase on the discharge epoch histogram and from the 3-dimensional (pulse amplitude/phase/discharge rate) histogram, respectively, Following the exposition of the basic mathematical concepts regarding the above param eters, experimental results are reported on the recognition capability of t he method in defining the defect category in a number of different specimen s.