基于混沌时间序列GA-VNN模型的超短期风功率多步预测  

Super-Short-Term Multi-Step Prediction of Wind Power Based on GA-VNN Model of Chaotic Time Series

作　　者：江岳春[1] 张丙江 邢方方[1] 张雨[1] 王志刚[2] 

机构地区：[1]湖南大学电气与信息工程学院,湖南省长沙市410082 [2]国网河南省电力公司驻马店供电公司,河南省驻马店市463000

出　　处：《电网技术》2015年第8期2160-2166,共7页Power System Technology

基　　金：国家自然科学基金项目(51277057);科技部技术创新项目(12C26214305038)~~

摘　　要：随着风电在电力系统中的渗透水平不断提高,能准确、可靠地进行风功率预测至关重要。为提高风功率超短期预测精度,利用风功率时间序列的混沌特性,推导分析了Volterra泛函模型和3层前馈(back propagation,BP)神经网络在结构上的一致性,提出混沌时间序列遗传算法-Volterra神经网络(genetic algorithm-Volterra neural network,GA-VNN)模型,对超短期风功率进行多步预测。该模型将实用的Volterra泛函模型和BP神经网络结合起来,解决了求解Volterra泛函模型高阶核函数的问题。同时设计了一种混沌时间序列GA-VNN模型的学习算法,在算法中利用GA全局寻优能力来优化BP神经网络,获得最优的初始权值和阀值。将上述方法应用于某风电场风功率超短期多步预测中,结果验证了所提模型的多步预测性能明显优于Volterra预测滤波器和BP神经网络。As wind power penetration in power system increases continuously, a good prediction method is essential to providing accurate and reliable results. In order to improve the wind power prediction accuracy, a super-short-term wind power prediction model is proposed based on genetic algorithm-Volterra neural network（GA-VNN） model using the chaos characteristic of wind power time series, mathematically deducing and analyzing the consistency between Volterra functional model and back propagation（BP） neural network in structure. The practicable Volterra functional model and BP neural network are combined together, learning the advantages of both, overcoming the difficulty in obtaining the high order kernel function of Volterra functional model, and designing a learning algorithm of the GA-VNN model of chaotic time series, in which BP neural network is optimized by utilizing the global optimization of GA algorithm to obtain the best initial weights and thresholds. The GA-VNN model of chaotic time series is applied to the super-short-term multi-step prediction of wind power, and the experimental results show that its estimated performance is obviously superior to both Volterra filter model and BP neural network and satisfactory results are achieved.

关 键 词：混沌时间序列 BP神经网络 GA算法 Volterra泛函模型 风功率超短期多步预测 

分 类 号：TM715[电气工程—电力系统及自动化]