基于LSTM的高压输电电缆接头温度异常预警系统设计

Design of high voltage transmission cable joint temperature abnormal warning system based on LSTM

电缆接头温度容易受到噪声干扰,导致温度预测精度较低,为此结合LSTM设计高压输电电缆接头温度异常预警系统。分析高压输电电缆接头温度场,通过参数电阻率控制连接点接触电阻。使用DS18B20温度传感器传输电缆接头温度,由预警终端根据传感器传输的温度进行预警。使用集成经验模态分解法分解处理接头初始温度,获取多个子序列,用来抵消加入白噪声所带来幅值影响。构建LSTM预警模型,用一维矩阵表示内外温差。以此为依据,将其与设定的温度上限进行对比,实现高压输电电缆接头温度预警。实验结果表明,该系统与实际温度数据只存在0.2℃的最大误差,能够有效监测电缆接头温度变化情况,具有良好的预警效果。

The cable joint temperature is easy to be disturbed by noise,resulting in low temperature prediction accuracy.Therefore,combined with LSTM,an abnormal temperature early warning system for high voltage transmission cable joints is designed.The temperature field of high voltage transmission cable joint is analyzed,and the contact resistance of the connection point is controlled by the parameter resistivity.DS18B20 temperature sensor is used to transmit the cable joint temperature,and the early warning terminal will give an early warning according to the temperature transmitted by the sensor.The integrated empirical mode decomposition method is used to decompose and process the initial temperature of the joint,and multiple subsequences are obtained to offset the amplitude impact caused by the addition of white noise.Build LSTM early warning model,and use one⁃dimensional matrix to express the internal and external temperature difference.On this basis,it is compared with the set upper temperature limit to realize the temperature early warning of high voltage transmission cable joints.The experimental results show that there is only a maximum error of 0.2℃between the system and the actual temperature data,which can effectively monitor the temperature changes of cable joints and has a good early warning effect.