Ultra High Voltage Directc Current Converter Valve Saturation Reactor Heat Prediction and Early Warning

To proactively detect potential faults in saturation reactors of converter valves, a method for predicting and warning of saturation reactor temperature based on multi-factor correlation analysis is proposed. Utilizing infrared detection statistical data from the valve hall, the algorithm comprehensively analyzes the correlation between various thermal influence factors. The LSTM neural network is employed to predict the temperature of the saturation reactors, and the results indicate that the accuracy of the model reaches 95%. Finally, early warning is conducted based on the temperature prediction data, and after the warning, the system captures images of the faulty component locations for inspection personnel to troubleshoot. This approach enables early detection of equipment issues, ensuring the safe and stable operation of the DC system.