基于EEDM-Arima算法的某地区电力系统负荷预测及算例分析

Load forecasting and analysis of power system based on EEDM-Arima algorithm

采用Arima算法来预测各个分量,结合实际负荷数据来分析具体算例,通过测试发现以EEDM分解处理得到的预测值达到了比分解前的预测值更高的精度。先为预测值构建相应的时间序列,之后实施差分平稳化处理,利用自相关系数以及偏自相关系数得到模型的阶数,接着估计各个参数并对其进行验证,最后利用含有合适参数的模型来完成预测过程。实验算例负荷预测在Matlab软件上完成,采用EEDM自适应方法来分解负荷,对于不平稳分量要先对其实施差分平稳化处理,置信度95%。各模型预测得到的相似度与准确度数据等于0.999 2,高于Arima算法,接近负荷实际变化趋势。

the Arima algorithm was used to predict each component, and the actual load data was used to analyze the specific calculation examples. Through testing, it was found that the predicted value obtained by EEDM decomposition reached a higher precision than the predicted value before decomposition.The corresponding time series were constructed for the predicted values, and then differential smoothing was carried out. The order of the model was obtained by using autocorrelation coefficient and partial autocorrelation coefficient, and then each parameter was estimated and verified. Finally, the model with appropriate parameters was used to complete the prediction process.The load prediction of the experimental example was completed on Matlab software, and EEDM adaptive method was adopted to decompose the load. For the unstable components, differential smoothing treatment should be implemented first, with a confidence of 95%.The predicted similarity and accuracy of each model is equal to 0.999 2, which is higher than the Arima algorithm and close to the actual change trend of load.