基于改进粒子群-径向基神经网络模型的短期电力负荷预测  

Short-Term Load Forecasting Based on Modified Particle Swarm Optimization and Radial Basis Function Neural Network Model

作　　者：师彪[1] 李郁侠 于新花[2] 闫旺[1] 何常胜[1] 孟欣[1] 

机构地区：[1]西安理工大学水利水电学院,陕西省西安市710048 [2]青岛科技大学高职技术学院,山东省青岛市261000

出　　处：《电网技术》2009年第17期180-184,共5页Power System Technology

基　　金：国家火炬计划创新基金(07C26213711606);陕西省自然科学基础研究计划基金(SJ08E220);山东省软科学基金(2007RKB188)

摘　　要：为了准确、快速、高效地预测电网短期负荷,提出了改进的粒子群–径向基神经网络算法。用改进的粒子群算法训练径向基神经网络,实现了径向基函数神经网络的参数优化。建立了短期电力负荷预测模型,综合考虑气象、天气、日期类型等影响负荷的因素进行短期负荷预测。算例结果表明,该算法优于径向基神经网络法和粒子群–径向基网络算法,克服了径向基网络和粒子群优化方法的缺点,改善了径向基神经网络的泛化能力,输出稳定,预测精度高,收敛速度快,平均百分比误差可控制在1.2%以内。To forecast short-term power load fast, accurately and efficiently, the features and defects of particle swarm optimization （PSO） algorithm are analyzed and an modified PSO- radial basis function neural network （RBFNN） algorithm is proposed, in which the RBFNN is trained by improved PSO to implement the optimization of RBFNN parameters and a short-term load forecasting model is built. In load forecasting such factors impacting loads as meteorology, weather and date types are comprehensively considered. Calculation example results show that the forecasting results by the proposed mehtod are better than those by RBFNN method and PSO- radial basis function network （RBFN） method, the defects of the latters are remedied and the generalization ability of RBFN is improved. The proposed method possesses following advantages： stable output, high precision of forecasting, fast convergence and its average percentage error is within the range of 1.2%, thus the proposed forecasting method is available to short-term load forecasting.

关 键 词：负荷预测 改进粒子群-径向基神经网络模型 泛化能力 预测精度 

分 类 号：TM715[电气工程—电力系统及自动化]