考虑气象因素的季节型电力负荷预测方法仿真

Simulation of Seasonal Power Load Forecasting Method Considering Meteorological Factors

由于电力负荷易受气象因素的影响,使电力负荷时间序列变化不稳定。为了提高电力负荷预测精度,以保证电力系统的稳定运行,提出一种考虑气象因素的季节型电力负荷预测方法。使用最小二乘支持向量机方法建立电力负荷异常数据训练模型,通过该模型训练电力负荷历史数据,根据训练结果补充负荷历史数据中残缺的数据。基于数据处理结果与电力负荷变化特征,提取易受气象因素影响的电力负荷,并分解不同季节的电力负荷,修正气象因素累积效应;运用回归神经网络改进时间序列预测模型,对样本进行训练,加强时间序列预测模型的准确性,通过引入加权均方误差,调整回归神经网络的连接权,改进电力负荷预测模型,得出电力负荷预测最优结果。实验结果表明,所提方法的季节型电力负荷预测精确度较高,其中,春季电力负荷预测精度平均值较现有方法的预测精度平均值分别高出3.8%和5.7%,验证了上述方法的应用效果。

At present,the electricity load is often affected by meteorological factors,which leads to instability in the time series of electricity loads.To improve the forecasting accuracy of electricity load and ensure the stable operation of the power system,this paper proposed a forecasting method for seasonal electricity load considering meteorological factors.Firstly,a training model of anomaly data was constructed by the least squares support vector machine.Based on this model,historical data of electricity load was trained,and then the missing data in historical data was supplemented according to the training results.Based on the data processing results and the characteristics of the electricity load,the electricity load susceptible to meteorological factors was extracted,and then the cumulative effect of meteorological factors was corrected by decomposing the electricity load in different seasons.Moreover,the regression neural network was used to improve the accuracy of the time series prediction model and train the sample,thus strengthening the accuracy of the model.Furthermore,the weighted mean square error was introduced to adjust the connection weights of the regression neural network,thus improving the prediction model.Finally,the optimal result of electricity load prediction was obtained.Experimental results show that the proposed method has high accuracy in forecasting seasonal electricity load.The average prediction accuracy of spring electricity load is 3.8%higher and 5.7%higher than that of existing methods,which verifies the application effect of the above method.