A Dynamic Intelligent Approach Based on Gaussian Function for Prediction of the Flashover Voltage Conditions on Polluted Polymer Insulators

This research presents a dynamic modeling approach to predict flashover voltage (FOV) of polymer insulators. Pre-flashover conditions can be used to represent analytical formulations for different stages of dry band arcing activities on the surface of polluted insulator. In this study, experimental results of pollution flashover tests are used to develop a circuit model of pollution arcs based on Obenaus’ model. Estimation of circuit values is done by use of intelligent algorithm with no dependency on profile of insulators which is also more beneficial than typical analytical calculations. Then, separation and investigation of discharge and pollution voltage is done and proper classification is made. This classification is later used to predict pre-flashover conditions and pollution flashover. Gaussian model is used to make fitness of voltage of pollution layer and voltage of discharge on the surface of insulator. Gaussian representation of each stage in pollution flashover contains critical signals leading to complete discharge. Based on the obtained model, a flowchart is presented to predict dynamic stages of flashover. Comparison to various experimental data shows close correlations with the results of proposed model. An online monitoring system can be developed according to obtained results in order to prevent pollution flashover.