Insulation condition assessment of high‐voltage rotating machines using hybrid techniques

Abstract

The enduring life span of the machine insulation will be decided based on degradation level in motor and generator stator windings. The non-destructive diagnostic tools like dielectric loss and capacitance test and partial discharge (PD) analysis, recognized to access the deterioration in the insulation system of rotating machines. The experiments reveal various characteristic parameters such as leakage current, dielectric dissipation factor, the capacitance value, and PD magnitude. The integrity of the rotating machine can be find out by analyzing these parameters. This research study shows the hybrid method for prediction of insulation condition in the stator winding by utilizing the artificial neural network (ANN) with gravitational search algorithm in comparison with ANN and ANN–genetic algorithm. The advent of expert systems ensures the quality assurance and service life assessment of the high-voltage assets. It offers a predictive maintenance solution to personnel dealt with power utilities thereby increasing the uptime, reliability, and productivity, which in turn reducing the operating costs, downtime and unplanned outages. For testing and predicting the insulation status, several 11 kV machines are considered. The predicted results using hybrid techniques extend a close agreement with reference to the data obtained from the experiments performed. The proposed method indicate the competent and trustworthy, by the presented test results.