基于多元线性回归模型的高压开关柜触点温度预测

Based on Multiple Linear Regression Model to Predict the Contact Temperature of High Voltage Switchgear

本文针对开关柜的触点温度预测这一关键问题,以采集到的110k V高压开关柜触点温度数据为基础,运用Mapreduce模型对一个监测点数据进行分析处理,运用多元线性回归方法,建立关于温度的回归模型,以F回归关系的显著性的判定标准,对预测结果进行分析,结果表明多元线性回归中的纵向回归,能够很好的拟合高压开关柜触点的温度,预测误差小,适用于温度的长期预测。

The aim of this paper was to use the Map-reduce model, analyzing and processing a monitoring point data and use multiple linear regression method to establish model about temperature regression, then the prediction results were analyzed with significant regression relationship between the F criteria. The result shows that the multiple linear regression longitudinal regression can well fit High Voltage Switchgear contact temperature with small prediction error, and it is suitable for long-term forecast of temperature.