At present

the closing time dispersion of circuit breakers used in 800 kV filter fields is remarkable. In order to improve the effectiveness of phase control and reduce the impact of inrush current on electrical components

the prediction method of closing time based on L-M neural network is proposed based on actual operating data. This prediction method comprehensively considers the influence of temperature and interval time on the closing time of circuit breakers. After processing and classifying the on-site operation data

a database of closing time is established. The closing time prediction model is established by learning the training set. The effectiveness of the algorithm is verified through error analysis of the validation set. The predictive ability of Support Vector Machines and L-M neural networks is compared. The results show that the closing time prediction error of the L-M neural network algorithm is superior to that of the Support Vector Machine algorithm. The L-M neural network algorithm has improved the prediction accuracy of a single circuit breaker

which meets the requirements for phase-controlled accuracy.