短期负荷预测中的负荷求导法及天气因素的使用

Load Derivation in Short Term Forecasting Using Weather Factor

为了将超短期负荷预测中具有高准确率和高可靠性的负荷求导法移植到短期负荷预测中,从采用差分运算、降低直接预测的数值,以及充分使用负荷“近大远小”、“相似日”性质等角度对其进行了分析。结果表明可以将负荷求导法用于短期负荷预测。将其与标准日负荷曲线取平均作为预测结果,示例表明日平均误差小于2%。若将最近15天左右预测值与实际负荷值之间的差值,再通过各敏感点负荷与平均气温等作回归分析,可以实现考虑天气的负荷求导法。

To apply the load derivation method to short term load forecasting, which is usually used in ultra short term load forecasting for its high precision and reliability, the strategies adopted in the method, such as difference operation, reducing forecasting value and using ＂near large and far small＂ and ＂similar day＂ characteristics, are analyzed in this paper. The analysis shows that it is feasible for short term load forecasting. Combining it with the standard daily load curve can make the mean daily forecasting error with in 2%. After calculating the residual loads with the latest 15-day forecasting results and actual values, the weather factor can be taken account to improve the forecasting precision by regression analysis between the sensitive residual loads and temperature.