Calculation of Action Times of Arrester with Clearance Based on GA-BP Neural Network Technology

The operation times of arrester counter are one of the main bases for judging the operation status of arrester, but it is not only time-consuming and laborious, but also difficult to implement the operation times of arrester read by artificial and unmanned aerial photographs. This paper presents a new method for calculating the action times of lightning arresters with gap based on the combination of lightning location system, transmission line data and GA-BP neural network technology. Firstly, the maximum lightning current in the lightning-induced area is determined. Then, the structure of GA-BP neural network is determined and trained according to the selected samples. The trained network is used to calculate whether the arrester acts after lightning strike, so as to realize the online prediction of the number of arrester actions. Examples show that the accuracy rate of this method is 90% for action recognition of lightning-induced arresters, 86.25% for non-action samples of lightning-induced arresters, and 80% for action recognition of lightning-induced arresters. The accumulative action times of the arrester calculated by the network basically coincide with the actual action times, and the overall accuracy rate is better.