Optimization of Short-term Load Forecasting with Extrapolation Algorithm Based on Fractal Theory
首发时间:2014-12-23
Abstract:Power load forecasting is an important part in the planning of power transmission construction. Considering the importance of the peak load to the dispatching and management of the system, the error of peak load is proposed in this paper as criteria to evaluate the effect of the forecasting model. The accuracy of short term load forecasting is directly related to the operation of power generators and grid scheduling. Firstly, the historical load data is preprocessed with vertical and horizontal pretreatment in the paper; Secondly, it takes advantage of fractal and time serial characteristic of load data to design a fractal dimension calculate method for disperse sampling data; Thirdly, the forecasting data image is made by fractal interpolation, the vertical proportion parameter which be used in the interpolation is determined by the similar historical load data, the image can review change condition between load spot. In the view of the nonlinear and complexity in the change of the short-term load, according to current load forecasting technology application and project needs in practice, combined with fractal theory (F&T), this paper built a short-term load forecasting model, and obtain good results. And to prove the effectiveness of the model, SVM algorithm and ANN network was used to compare with the result of F&T. The results show that the new model is effective and highly accurate in the forecasting of short-term power load than the other models. The root-mean-square relative error (RMSRE) of new model is only 2.55%, the SVM model and BP network is 3.26%, 3.60% separately
keywords: Fractal theory Data pretreatment Extrapolation algorithm Load forecasting
点击查看论文中文信息
基于分形理论外推算法优化的电力负荷预测研究
摘要:电力负荷预测是电力系统进行计划和营销的重要部分。考虑到在实际的电力调度和管理中最大负荷的重要性,本文将最大负荷的预测误差分析作为衡量负荷预测模型精度的重要指标。电力负荷预测的准确性直接影响着电源和电网系统的正常运转。首先,本文将历史的负荷数据利用立体式预处理技术进行了归类;第二,利用分形理论和时间序列特征建立了分形维数计算模型来对样本数据进行有效聚类;第三,利用分形插值函数分析出预测数据映像值,其中,选择垂直比例参数受到了历史相似负荷数据的影响,所以其映像值在负荷变化过程中也在不断的调整。由于负荷在变动中的非线性和复杂性,利用分形理论将分形插值函数引入到电力负荷理论中,得到了较好的预测结果。通过测试将所建立的模型所取得的结果与支持向量机和人工神经网络预测模型进行对比均具有较好的精度,基于分形插值外推算法得到的均方根误差仅为2.55%,而利用SVM和BP神经网络预测的模型得到均方根误差分别是3.26%和3.60%。
基金:
论文图表:
引用
No.4623752102477514****
同行评议
勘误表
基于分形理论外推算法优化的电力负荷预测研究
评论
全部评论0/1000