电力负荷非线性预测方法与研究方向探讨

进行精确的负荷预测是电力系统规划、计划、用电、调度等部门的基础工作。传统的预测方法主要是运用经典的确定性理论，而电力需求受众多因素影响，具有复杂的非线性特征．运用经典的预测方法可能会抹杀负荷变化复杂性本质。为此，通过对影响电力负荷变化因素及电力市场非线性特征的分析．介绍了几种非线性预测方法的构成原理与特点，包括基于人工神经网络（ANN）、小波分析、模糊理论、灰色理论、分形理论和支持向量机的负荷预测，并对近年来各非线性预测方法及其组合预测的主要研究成果进行了总结与评述。同时根据国内外电力系统负荷预测的现状与经验。对未来的主要研究方向作了一些有益的探讨。

Review of power load forecasting based on nonlinear method and its development

Precise power load forecasting is the foundation of power system planning and operation.The traditional load forecasting methods are mainly based on classical analytics equation.However,the actual demand for electricity is influenced by many uncertainty factors; it possesses the characteristic of complicated nonlinear.Therefore,those forecasting methods based on classical theory may fail to express the complexity of power load undulation.Some factors affecting load variety and the nonlinear characteristic of power market are analyzed in this paper.Several nonlinear forecasting methods are introduced in composing principle and peculiarity.They are the methods based on artificial neural network (ANN),wavelet theory,fuzzy theory,gray theory,fractional theory and support vector machine (SVM).Furthermore,the main research results of the above nonlinear forecasting methods and their combination in recent years are reviewed and discussed.Based on these discussions and the experience of China and other countries,some advice is given on future research direction.