基于层次分析法和径向基函数神经网络的中长期负荷预测综合模型

A Comprehensive Model for Long-and Medium-Term Load Forecasting Based on Analytic Hierarchy Process and Radial Basis Function Neural Network

中长期负荷预测是电力系统规划与运行的基础工作,提出基于3指标量,即指标总量、指标增长量和指标增长率的综合模型。首先构建层次分析(analytic hierarchy process,AHP)模型,分别对3个指标量进行分析评价,优选出每个指标量的最优预测模型,然后利用径向基函数(radial basic function,RBF)神经网络对3个最优模型的预测结果进行拟合,并将GDP因素也作为神经网络输入数据之一,输出最终的预测结果。AHP模型中综合考虑了模型预测误差和模型拟合度,并成功地加入了人工干预的因素,依据专家经验判断模型的信任度和预测结果趋势可信度。AHP模型采用与预测时刻最近的历史数据进行分析,因此具有较好的实时性。实验结果表明该综合模型具有较高的预测精度,实际应用效果较好。

Long- and medium-term load forecasting is the foundation of power system planning and operation, for this reason a comprehensive model based on three index quantities, i.e., the total index quantity, the increasing index quantity and the index growth rate, is proposed. Firstly, an analytic hierarchy process （AHP） model is constructed to analyze and estimate the three index quantities respectively, then select out optimal forecasting model for each index quantity; secondly, by use of radial basic function neural network （RBFNN） the forecasted results from the three optimal models are fitted; thirdly, taking the GDP factor as one of the input data of neural network, the final forecasting result is output. In AHP model both forecasting error and the fitting degree of the model are comprehensively considered as well as the manual intervention is successfully added, finally according to expert experience the trust degree of the model and the confidence level of the trend of forecasting results are judged. The proposed AHP model uses the historical data nearest to the time point to carry out the forecasting for the analysis, therefore, it possesses better real-time performance. Experimental results show that using the proposed model a more accurate forecasting result can be obtained, and practical application results verify the practicability of this method..