基于神经网络的风力发电系统风速软测量  

Study on wind speed soft sensing based on neural networks for wind power generators

作　　者：刘震[1] 汪令祥[1] 

机构地区：[1]合肥工业大学电气与自动化工程学院,安徽合肥230009

出　　处：《中国电力》2010年第11期87-91,共5页Electric Power

基　　金："十一五"国家支撑计划资助项目(2006BAA01A20)

摘　　要：对风力发电并网系统的有效风速测量问题进行研究。鉴于神经网络可应用于非线性系统的模型与辨识,不受非线性模型类的限制,且可给出工程上易于实现的学习算法,提出基于神经网络的有效风速软测量。对实时采集的风力发电机组的风速样本参数集进行分析、训练及拟合,获得相应的有效风速计算网络。仿真结果表明,有效风速软测量可代替风速仪的作用,是一种非常有效的风速估计方法。Problems in speed measurements for wind power generators were studied. Considering that neural networks can be applied to modeling and identification of nonlinear systems without the restrictions of nonlinear models and give the learning algorithm for project to easily realize, a neural network based wind speed soft sensing was proposed. The wind speed sample set by real-time acquisition was analyzed, and the training and matching for an effective calculation network of wind speed were made. Simulation results show that the effective wind speed soft sensing can replaee anemometers and it is a very effective method for the estimation of wind speed.

关 键 词：风力发电 风速测量 神经网络 软测量 变桨距控制 

分 类 号：TM614[电气工程—电力系统及自动化]