Research on Power Safety Monitoring using the Laplace Score Method and A BP Neural Network
Yang Xue, Runan Song, Penghe Zhang, Yining Yang, Zhongqiang Wu · 2024
The safety of power users is very important to ensure the productivity of enterprises and social stability. In-depth mining of acquired information and monitoring data on users' power use is required to evaluate the safety of users' power use in advance to prevent safety risks and improve the safety of power use. In this study, the safety risk posed to power users is assessed using the Laplace score method, feature extraction, fault determination based on the Markov distance, and a safety risk assessment using a BP neural network. Practical examples are used to identify five fault types: overvoltage, overcurrent, overload, loss of current/voltage, and abnormal neutral current. The identification accuracy of the five fault types in test data samples is more than 96%, which validates the model proposed in this study.