Fast Assessment Method of Transmission Lines Switching Overvoltage Based on BP Artificial Neural Network

In the response of grid emergency and the design of electric transmission and transformation equipment, fast assessment method is involved to check the no-load switching overvoltage of transmission lines. This paper presents fast assessment method of no-load switching overvoltage of risk transmission lines based on error back-propagation (hereinafter referred to as BP) artificial neural algorithm. According to the theoretical equations of no-load switching overvoltage of transmission lines, the system intensity, the line parameters and the line length are selected as the input variables. While the switching overvoltage risk levels, which are converted from simulation results to reduce the statistical property of the transient overvoltage, are selected as the output variables. Then, the BP neural network algorithm could construct the mapping relationship be-tween the input and output variables. Finally, several 500kV transmission lines in the actual grid are gathered to verify the trained BP neural network. In this case, the error caused by the discreteness of circuit breaker switching time, the statistic properties of transient overvoltage and the uncertainty of artificial intelligence algorithm can be reduced by multiple estimations to obtain the most probable value.