Design of LSTM-Based Anomaly Warning System for High-Voltage Transmission Cable Joints

High-voltage transmission cable joints abnormal warning is easily affected by harmonic currents, resulting in poor warning effect. For this reason, a high-voltage transmission cable joints abnormal warning system based on LSTM is designed. We need to detect the abnormal current harmonics of high-voltage transmission cable joints, and obtain the dielectric state of cable joints. This paper designs CC2430 microprocessor warning module, responsible for processing the current data in the warning area, builds ground wire current detection Detection module, which is responsible for evaluating cable joint anomalies. We also construct the LSTM neural network warning model, compare and analyze the realtime status with the detection point status in the model, and obtain the prediction value of the current cable joint status. By comparing and analyzing the realtime status and the detection point status in the model, we obtain the prediction value of the current cable joint status. The moving average technique is used to analyze the overall change direction of the current data and determine the relationship between the current prediction data and the threshold value. This paper determines the relationship between the current prediction data and the threshold value, and distinguishes the abnormality level according to the color of the system indicator. From the experimental results, it can be seen that the maximum current value of the system is 125 A, and the change curve is consistent with the actual curve. The value is 125 A, and the change curve is consistent with the actual curve, which indicates that the warning effect of the system is good.