A hybrid approach for investigation of electrical characteristics for rod to point electrode in gas insulators

Abstract Designing and conducting experiments to simulate real-world conditions with varying gas insulators and electrode geometries can be intricate. Ensuring repeatability and accuracy across experiments poses a challenge due to the inherent complexity of electrical phenomena in gas insulators. This manuscript proposes an investigation of electrical characteristics for rod to point within gas insulators. The proposed approach combines the Mother Optimization Algorithm (MOA) and Circular Dilated Convolutional Neural Network (CDCNN), Hence it is named as MOA-CDCNN technique. The major objective of the paper is to investigate the electrical characteristics of rods and govern breakdown voltage, electric field distribution, and discharge patterns in gas-insulated systems. The MOA is used to optimize the gas insulation system. The CDCNN method is utilized to predict the insulator parameters. The performance of the proposed method is put into practice on the MATLAB platform and contrasted with methods that are currently in use. Particle swarm optimization (PSO), heap-based optimizer (HBO), and wild horse optimizer (WHO) all yield inferior results compared to the proposed strategy. From the result, it is concluded that the proposed method shows a maximum electric field of 6.8 kV/mm compared with other existing methods.