基于时间序列突变误差校正的超短期风速联合预测模型

An Ultra-short-term Wind Speed Joint Prediction Model Based on Time Series Catastrophe Error Correction

为了提高风速预测的准确性,提出了一种超短期风速联合预测模型.该模型首先利用经验模态分解与局部均值分解分别将风速数据分解为一系列相对平稳的分量,然后采用灰狼算法进行参数寻优的支持向量机(GWO-SVM)对分量进行预测,最后整合所有分量的预测结果得到风速预测结果.此外,为了减小预测过程中存在的误差,对误差类型进行了分类和分析,提出了一种基于时间序列突变的误差校正方法,采用时间序列预测模型直接对误差值进行校正,有效地减小了风速预测的误差.最后,通过仿真实例,证明了该风速联合预测模型和误差校正方法可以显著地提高风速预测的准确性.

In order to improve the accuracy of wind speed prediction,this paper proposes an ultra-short-term wind speed joint prediction model.The model first uses empirical mode decomposition and local mean decomposition to decompose the wind speed data into a series of relatively stable components,and then uses the gray wolf algorithm to perform parameter optimization support vector machine(GWO-SVM)to predict the components,and finally integrates the prediction results of all components get wind speed prediction results.In addition,in order to reduce the errors in the prediction process,the types of errors are classified and analyzed,and an error correction method based on time series mutation is proposed.The time series prediction model is used to directly correct the error value and effectively reduce the error of the wind speed prediction.Finally,simulation examples show that the wind speed prediction model and error correction method can dramatically improve the accuracy of wind speed prediction.