Construction of prototype on-line expert system for insulation degradation diagnosis of polymeric insulation material suffering partial discharge (PD)

The authors have constructed a prototype of an online PD pulse measurement and expert system (PDM/ES-I) using a personal computer for insulation degradation diagnosis of materials suffering partial discharge (PD). For 0.1-mm-thick LDPE (low-density polyethylene) film, using the CM-II electrode system and PDM/ES-I, they measured the temporal changes of Φ-q-n characteristics of PD pulses and analyzed 33 kinds of statistical PD parameters up until breakdown. They found that five new statistical PD parameters correlated with aging time. PD pulse occurrence distribution also proved to have a correlation with aging time. Diagnosis was made under the same conditions with either a diagnosis system using statistical parameters or one using pattern recognition. Both permitted an exact determination of the degree of degradation at a given aging time, predicting the breakdown residual life of the CM-II electrode system using LDPE