基于神经网络风力发电机组载荷优化控制策略研究  

Optimal Control Strategy for Wind Turbine Load Reduction Based on Neural Network

作　　者：刘军[1] 陈东亮[1] 

机构地区：[1]西安理工大学自动化与信息工程学院,陕西西安710048

出　　处：《电气传动》2014年第2期59-63,共5页Electric Drive

基　　金：教育部博士点基金资助(20106118110009);西安市科技计划项目资助(CX1250)

摘　　要：为减轻大型风力机叶片在工作过程中受到的空气动力载荷均值,提出一种基于神经网络的风力机载荷优化控制策略。基于神经网络在线建立了风力机载荷数学模型;在功率控制的基础上,以减小载荷为目标,利用梯度法计算出控制量的修正量,从而实现功率与载荷综合优化控制的目的。将该控制策略应用于2 MW风机的非线性模型,用Matlab/Simulink进行了仿真,仿真实验结果表明,对比传统PI控制器,该控制策略能有效降低风机叶片的空气动力载荷均值。A load optimizing control strategy based on neural network model to reduce mean aerodynamic fatigue load of blade for largescale wind turbine was proposed. The mathematic model of wind turbine was set up by neural network online. And load reduction was adopted as another control on the basis of power control. The negative gradient of load is calculated as the modifying signal. Then the negative gradient signal is added to the output of PI controller to accomplish the goal of power and load optimization. The control strategy was applied to nonlinear model of 2 MW wind turbine in Matlab/Simulink. The results show that the controller is effective in reducing mean value of aerodynamic load of blade, comparing to traditional PI control.

关 键 词：风电机组 载荷优化 神经网络 梯度法 

分 类 号：TK89[动力工程及工程热物理—流体机械及工程]