电/气能源消耗与温度的相关性研究

Correlation study between energy consumption and temperature

针对能源消耗和保障需要精确的负荷预测,有必要研究温度和能源消耗之间的关系,利用经验模态分解和线性回归分析确定了天气与用电和天然气之间的关系,作为数据处理工具,经验模态分解可以将原始数据划分成若干个本质模态函数。在对温度和电/气负荷进行分解后得到的最低频率模态函数进行线性回归分析,可获得温度与负荷的相关性。实验结果表明,相对于直接运用实际负荷数据,经过经验模态分解后得到的结果可以更好地展示负荷与温度之间的相关性,电负荷、天然气与温度的相关性分别为0.959和-0.859。

In order to understand energy consumption and ensure more precise load prediction, it is essential to identify the variations of energy consumption in response to temperature change. Empirical mode decomposition and linear regression analysis were utilised to identify their relationship. By applying data mining techniques, some outliers from gas load and temperature data points were detected and excluded beforehand to ensure data accuracy. The correlation coefficients of linear regression between gas load and temperature could be identified, which were important index to quantify their relationship. According to half-hourly analysis, the correlation was generally higher at nights than that during day time. For seasonal level, it droped slightly, varying from 0.68 in summer to 0.83 in winter. The correlation coefficients were 0. 959 and 0. 859 respectively.