Discussion on "Power Quality Disturbance Waveform Recognition Using Wavelet-Based Neural Classifier—Part 2: Application"

We believe that the field on the surface of the weathersheds can be reasonably well estimated by a combination of 3-D calculations, on insulators without weathersheds (but with representation of tower and conductors) to calculate fields along the axis of the insulator, and 2-D axi-symmetric calculations on insulators with weathersheds to determine the relative relationship between the field along the axis and the field along the weathershed surface. This avoids the necessity of 3-D modeling of the weathersheds, a step which would add a significant amount of time to the computations. Another way to estimate the field on the surface is to model a limited number of weathersheds at the locations which can be determined by 3D calculations of the insulator without weathersheds. We also believe that using the potential calculated on the insulator axis in a 3-D model without weathersheds as a boundary condition in the 2-D model, as indicated in Dr. Olson’s discussion, will increase the degree of accuracy to the configuration where the insulator is installed. Verification by typical cases may give us more confidence in this estimation.