Ablation state assessment of SF6 circuit breaker contacts based on BP neural network and mean impact value

Abstract

In order to evaluate the ablation state of the SF6 circuit breaker contacts, simulated ablation tests were carried out on the circuit breaker, and the dynamic resistance curve as a function of contact travel was measured with three levels of injected current. The results showed that the increasing of test current value can reduce the fluctuation of the dynamic resistance curve as a function of contact travel. In addition to the ablation effect of current, the loosening of the operating mechanism will affect the accuracy of opening and closing action of the contact, resulting in the continuous reduction of the main contact travel in the dynamic resistance curve as a function of contact travel. Based on the measurement results, a BP neural network evaluation model was built to evaluate the ablation state of the contact. Seven parameters were extracted from the dynamic resistance curve as a function of contact travel as model inputs, and mean impact value algorithm was used to optimize the model. After the optimization, the average relative error of the evaluation to the verification set is less than 10.4%.