Research on mechanical life prediction algorithm of circuit breaker based on LSTM

Aiming at the mechanical life prediction of high-voltage circuit breakers, and combining relevant theories of deep learning, this paper proposes a mechanical life state prediction algorithm for high-voltage circuit breakers based on the Long Short-Term Memory (LSTM). This method uses the contact action travel data of the circuit breaker as a characterization quantity for mechanical life state, constructs a prediction algorithm model based on a time-series sample dataset, and studies the influence of sample quantity on algorithm performance. Experimental results show that the method proposed in this paper can predict the travel curve of future operating cycles based on historical data; the deviation between the algorithm’s prediction results and actual data is very small, and the algorithm’s performance is excellent. The method proposed in this paper can provide maintenance suggestions for the operation and maintenance of high-voltage circuit breakers and can help transform the current substation operation and maintenance mode from regular maintenance to predictive maintenance.