基于RBF神经网络的短期负荷预测方法综述

摘要

This paper introduces the concepts of the power system short-term load forecasting methods based on RBF neural network, and discusses its specific implementation approach. The continuous improving process of this method is reviewed by analogy analysis, and the progress made in practice is pointed out. Then, the basic principles and technical characteristics of several maturer RBF neural network prediction models are given with comparison and evaluation. According to the actual characteristics and new situations of power system operation, the method is analyzed from three aspects of improved algorithm, the original load data selection and combination of the actual load characteristics. The development space of continuous improvement in the field is discussed, finally the further development trend of technology in this field is prospected.%介绍了基于RBF神经网络的电力系统短期负荷预测方法的相关概念,论述其具体实现途径.通过类比分析的方法对该类预测方法改进的过程进行回顾,指出其在实践中取得的进步.阐述了一些比较成熟的基于RBF神经网络预测模型的基本原理和技术特点,并对它们进行了评价.根据电力系统运行的实际特点和面临的新情况,从算法改进、原始负荷数据筛选和如何结合实际负荷特点等三方面对该方法进行分析.探讨了该领域持续改进的发展空间,指出了该领域进一步发展的技术趋势.