基于多特征量的GIS触头温度预测方法

GIS contact temperature prediction method based on multiple parameters

为防止因气体绝缘开关(gas insulated switchgear,GIS)触头温升造成的事故,有必要对GIS触头温度进行监测与预测。针对触头温度不易直接测量以及其温度易受运行工况与外界因素影响的问题,文中提出了一种基于多特征量的GIS触头温度预测方法。文中通过建立三维仿真模型,分析了在不同接触电阻值、负荷电流、环境温度、风速、SF 6压强、太阳辐射强度下GIS的温度分布规律,结合热路理论定性验证了仿真模型的可靠性。通过分析可知,GIS触头温度预测的关键因素为外壳温升、负荷电流、风速、SF 6压强、太阳辐射强度,而环境温度影响可忽略,采取反向传播(back propagation,BP)神经网络用以上多特征量预测触头温升,将得到的预测值与建模方法的计算结果进行对比,误差为-0.7~0.68℃。该预测方法综合考虑多种影响因素对GIS温度场的影响,为基于外置传感器的GIS触头温度预测提供参考。

In order to prevent the faults caused by the overheating phenomena of gas insulated switchgear(GIS)contacts,it is necessary to monitor and predict the temperature of GIS contacts.In view of the problems that the temperature of contacts is not easy to be directly measured and the temperature is easily affected by the operation conditions and external factors,a prediction method of GIS contact temperature based on multiple parameters is proposed in this work.The temperature distribution law of GIS under different influencing factors is investigated by employing a three-dimensional simulation model.The influencing factors are contact resistance,load current,ambient temperature,wind speed,SF 6 pressure and solar radiation intensity.The reliability of the model is validated by use of heat circuit model.The results indicate that the key factors to predict the contact temperature are the shell temperature rise,load current,wind speed,SF 6 pressure and solar radiation intensity.At the same time,the influence of ambient temperature can be ignored.Further,the back propagation(BP)neural network is adopted to predict the temperature rise of the contacts with the above factors.The predicted temperature rise is compared with the calculated one of the model,and the error is in the range of-0.70~0.68℃.The method comprehensively takes into account the influence of various factors on the GIS temperature field,and it helps to give a reference for the temperature prediction of GIS contacts based on external sensors.