Classification of partial discharge patterns in GIS using adaptive neuro-fuzzy inference system

Abstract

Partial discharge (PD) measurement is among the most important methods of diagnosing insulation systems in high-voltage equipment. It is a convenient means of evaluating the state of the insulation and its prospective condition. PD activities may arise from various defects, and they vary according to the defects that cause them. The PD patterns that are generated by three laboratory models of defects in gas-insulated switchgears (GISs) are recorded and analyzed. This research involves PD tests that involve three sets of GIS apparatus with prefabricated defects. Five of 74 statistical PD features were selected as the inputs of adaptive neuro-fuzzy inference system (ANFIS) according to the training errors in 10000 epochs. The ANFIS was utilized to construct a fuzzy inference system (FIS). This FIS was then used to identify the source of the PDs. The results reveal that ANFIS classification has a high success rate, reaching an acceptable classification accuracy 91.5% at the lowest possible test voltage.

Exploring the Impact of Parallel Architecture on Improving Adaptable Neuro-Fuzzy Inference Systems for Gas-Insulated Switch Defect Recognition

“…Initializing the fuzzy inference system (FIS) model of ANFIS is more complex, involving the selection of inputs [33] and the selection of FIS models, the design of if-then rules, and membership functions within the FIS model. After the initial FIS model is designed, training data can be input to train the ANFIS before initial validation is performed by inputting test data.…”