基于时间序列的风电功率日前预测模型及其应用

Research of Day-ahead Forecasting Model for Wind Power Based on Time Series and Its Application

风电场发电功率随风速的无序变化是电网无法大规模接纳风电的关键因素,准确地预测风电场输出功率对电力系统大量接入风电有重要意义。针对风电功率无序变化的特征,基于时序分析法分别建立了指数加权移动平均和一阶差分自回归滑动平均的风电功率日前预测模型,进而运用穷举搜索法确定了指数加权移动平均模型的最佳加权因子为0.7,并得到此模型的风电功率预测值。同时,通过样本自相关函数定阶和最小二乘估计的方法,求得一阶差分自回归滑动平均模型的风电功率预测值。结果表明,一阶差分自回归滑动平均模型的风电场功率预测值的均方根误差比指数加权移动平均模型低0.88%,相应的准确率和合格率较高,可见一阶差分自回归滑动平均模型更能提高风电功率的预测精度。

For the key issue of the randomness characteristics of wind power changing with wind speed,the large amounts of wind power can not be applied in power grid.Therefore,it is important to predict the wind power accurately.Based on the time-series analysis,the exponentially weighted moving average（EWMA）model and the auto-regressive and moving average（ARMA）model were established to predict the wind power of the next 24 hours.Using the exhaustive search method,the EWMA model＇s optimal weighting factor was determined to be 0.7.And then,the EWMA model＇s wind power prediction value was calculated.Combined with the autocorrelation function and the least squares estimation,the ARMA model＇s wind power prediction value was obtained.The results show that the RMSE of ARMA model is lower 0.88%than the RMSE of EWMA;its accuracy rate and pass rate of ARMA are higher.Therefore,the ARMA model can improve the prediction accuracy of wind power.