基于极限学习机的风电机组变桨距系统辨识方法研究

Research on Identification Method of Variable Pitch System of Wind Turbine Based on Extreme Learning Machine

针对风电机组在运行过程中难以建立精确的数学模型的特点,将极限学习机应用在风电机组变桨距系统辨识中。通过对变桨距系统的动态过程进行分析,确定了变桨距系统辨识的输入输出。风速和桨距角作为极限学习机神经网络模型的输入,发电机功率作为极限学习机神经网络模型的输出。从而构建输入输出的样本集,对网络进行训练,当学习精度满足要求,确定网络隐层节点数,得出ELM神经网络变桨距辨识模型。仿真结果表明,ELM神经网络算法在变桨距系统辨识中具有比较高的辨识精度和效率。

In view of the fact that it is difficult to establish accurate mathematical model during the operation of wind turbine, the ultimate learning machine is applied to the identification of variable pitch system of wind turbine. By analyzing the dynamic process of the variable pitch system, the input and output of the variable pitch system are determined. The input of the system is wind speed and pitch angle, and the output signal is generator power. Thus the input and output sample sets are constructed and the network is trained. When the learning accuracy meets the requirements the number of hidden nodes in the network is determined and the ELM neural network variable pitch identification model is obtained. The simulation results show that the ELM neural network algorithm has high accuracy and efficiency in the identification of variable pitch system.