Intelligent molded case circuit breaker based on power carrier technology and LSTM

Abstract In response to the problems of interference and low fault diagnosis accuracy in traditional circuit breaker communication, a hybrid model of high-speed power line carrier communication and convolutional neural network-long short-term memory network is studied to construct an intelligent molded case circuit breaker system. In terms of communication, the use of orthogonal frequency division multiplexing modulation and parallel cascaded convolutional code encoding technology significantly improves anti-interference ability and transmission stability. In terms of fault diagnosis, a dual input convolutional neural network-long short-term memory network model is constructed, which integrates current and voltage waveform image features with multi-dimensional temporal state data to achieve accurate identification and early warning of multiple faults. The experimental results show that the high-speed power line carrier communication module can effectively improve communication performance, achieving a communication success rate of over 99% and a transmission delay of less than 100 ms at a rate of 1 Mbps. In addition, the proposed model achieves an average accuracy of 98.2%, which is 3.7%∼12.4% higher than advanced models such as Transformer and XGBoost. Especially in the case of contact overheating faults, the F1 Score reaches 96.0%. On the STM32H743 microcontroller, the model’s single inference time is only 45.2 ms, with an average warning advance of 280.5 ms, fully meeting the real-time requirements of ‘warning trip’. This system significantly improves the accuracy of fault diagnosis and the timeliness of early warning, providing an efficient communication computing integrated solution for the intelligent upgrade and predictive maintenance of circuit breakers in low-voltage distribution networks.