摘要
风速预测对风电场和电力系统的运行具有重要意义。为了提高风速预测的精度,提出了一种新的风速预测方法——基于粗糙集理论的遗传神经网络模型。由于影响风速预测的因素很多,利用粗糙集理论的属性约简对神经网络输入的影响因素进行约简,识别出与预测风速相关性较大的影响因素作为输入量,减少了神经网络的计算量;利用全局搜索能力强的遗传算法对神经网络的初始权值进行优化,克服了神经网络收敛速度慢和容易陷入局部极小的缺点。实例结果表明该算法能够有效地提高预测的速度与精度,证明了该方法在风速预测中的可行性和有效性。
Wind speed forecasting is very important to the operation of wind farms and power system.To improve the wind speed forecasting accuracy,a new method-genetic algorithm neural network based on rough set theory is proposed in the paper.Due to many factors influencing the wind speed forecas-ting,the attribute algorithm of the rough set theory is used to select the influencing factors input into the neural network and identify factors with high correlation with wind speed forecasting which are used as input,thus reducing the work and calcula-tion time of the network.The genetic algorithm with ability of strong global search is used to optimize the initial weights of the neural network to overcome the shortcomings of the BP algori-thm,such as slowness in training speed and convergence to local minimum.The forecasting results of some actual examples show that the new method for wind speed forecasting can improve the speed and the accuracy of prediction,and it is also feasible and effective in the wind speed forecasting.
出处
《电网与清洁能源》
2012年第9期82-87,共6页
Power System and Clean Energy
基金
国家863高新技术基金项目(2008AA05Z216)~~
关键词
风速预测
粗糙集理论
遗传算法
神经网络
wind speed forecasting
rough set theory
genetic algorithm
neural network
