基于自适应滤波的日负荷预测新陈代谢模型

提出了对日负荷进行预测的新方法。基于自适应滤波算法进行预测,在预测过程中对原始数据进行新陈代谢处理,且根据预测日的属性对预测结果进行加权,并依据历史负荷中负荷的变动情况对结果进行校正,以求最佳预测效果。利用自适应滤波预测结果的残差建立时间序列的AR(p)模型,与自适应滤波模型形成组合模型,从而实现了短期电力负荷样本资料随时间变化而更新、样本量和计算量不增加而预测精度能得到保证的目标。与传统的预测方法相比较,该模型用于日负荷预测具有计算迅速、精度高的优点。

Metabolic Models about Daily Load Forecasting Based on Adaptive Filtering Algorithm

A novel daily load forecasting method based on adaptive filtering algorithm for short-term daily load is presented.During the forecasting,the history data is dealt with using metabolic theory and the weighted coefficient is given to the result according to the day property,then the result is adfusted on historical date variable to get better accuracy.Furthermore,the AR model for the error is set up and formed the combined model with the adaptive filtering model.The combined model realizes the target of high accuracy without improving the number of the sample data and computing.Simulation results demonstrate the effectiveness of the model and the feasibility of the proposed algorithm.