基于多元经验模态分解的电网短期负荷预测及薄弱线路辨识

Power grid short⁃term load prediction and weak line identification based on multivariate empirical mode decomposition

为了更加准确地进行短期电力负荷预测,提高预测精度,利用不平稳电力负荷时间序列具有周期性和非线性的特征,提出一种新的电力系统短期负荷预测组合模型。该模型将多元经验模态分解法与支持向量回归算法相结合,将电力负荷时间序列及重要影响因素序列同时分解成个数相同的子序列,然后利用上述模型分别预测,再一一对应组合得到最终预测结果。以N市电网2017年1月—12月的日历史负荷为研究对象,并根据气象资料记录的日平均温度、湿度、气压、负荷率等影响因素,对建立的模型进行仿真分析,预测结果与2017年12月24日—31日实际的数据做对比。结果表明该模型预测精度较高,方法可行。将预测得到的结果,用于对N市电网的薄弱线路辨识中,给出了12月24日—31日每天的薄弱线路排序表,以便于工作人员对薄弱线路进行重点监测与保护。

In order to carry out short⁃term power load forecasting more accurately and improve the forecasting accuracy,a new combined model of short⁃term load forecasting is proposed on the basis of the periodic and nonlinear features of unstable power load time sequence.In the model,the multivariate empirical mode decomposition(MEMD)are combined with support vector regression(SVR)algorithm,the power load time sequence and important influencing factor sequence are decomposed into the same number of sub⁃sequences at the same time,and then the model above⁃mentioned is used to make prediction respectively,and the final prediction results are obtained by corresponding combination.The daily historical load of the power grid in the N city from January to December 2017 is taken as the research object.The established model was simulated and analyzed according to the daily average temperature,humidity,air pressure,load rate and other influencing factors recorded in meteorological data.The predicted results were compared with the actual data from December 24 to December 31,2017.The comparison results show that the established model has high prediction accuracy and the method is feasible.And then,the predicted results are used to identify the weak lines of the power grid in the N city.A ranking list of weak lines for each day from December 24 to December 31 is given,so that the staff can focus on the monitoring and protection of the weak lines.