Haiyan Yao, Xueheng Yao, Qiang Guo, Zhengfang Shi, Qintian Lin, Peiqin Jia
Partial discharge model, temperature rise model, switch cabinet,fault detection, GA-BP, LSTM
To improve the accuracy and real-time performance of fault detection in switch cabinets, a new type of fault detection method is studied and proposed. First, the partial discharge (PD) signal of the switch cabinet is analysed by using the long short-term memory (LSTM) network. By taking advantage of its superiority in time series data processing, the faults caused by PD are accurately identified. Second, the backpropagation neural network optimised by a genetic algorithm can effectively predict the temperature rise (TR) trend of switch cabinets and identify potential faults caused by temperature rise. The experimental results show that the average recognition accuracy rate of the PD model is as high as 98.52%, and the loss rate of the TR model is 4.21%. The hybrid fault detection platform combined with the two achieves a fault detection accuracy rate of more than 95% in practical applications, while maintaining a low false alarm rate and detection delay. The research can achieve early warning of switch-gear faults, reduce the downtime of equipment, and thereby improve the stability and operational efficiency of the power system.
Important Links:
Go Back
