双馈风电机组网侧控制器参数辨识的频域方法

A Frequency-Domain Based Method to Identify Parameters of Grid Side Converter Controller for Doubly Fed Induction Generators

提出了基于频域方法解耦辨识双馈风电机组网侧控制器参数的新方法。首先根据双馈风电机组网侧控制器模型以及发电机端口到网侧控制器的电压方程,推导出网侧控制器的解耦模型,该模型可实现d轴和q轴参数的解耦辨识。然后根据解耦模型传递函数的幅值灵敏度及相位灵敏度,分析了网侧控制器参数的可辨识性及辨识的难易度。通过在输入参考信号上叠加伪随机信号,根据输入信号的自功率谱及输入、输出信号的互功率谱,获得传递函数的频率序列。最后应用Levenberg-Marquardt优化算法对网侧控制器各参数进行了辨识。50次辨识值的方差表明了参数辨识结果的稳定性,辨识结果表明了所提方法的可行性。

A new frequency-domain based method is proposed in this paper to identify the parameters of grid side converter controller for doubly fed induction generators （DFIG）. Firstly, the d-axis and q-axis decoupling model of grid side converter controller is deduced based on the grid side converter controller model and electromagnetic equation between generator terminal voltage and grid side converter voltage. Based on this model, the d-axis and q-axis model parameters can be estimated separately. Secondly, the identifiability and sensitivities of each parameter are analyzed based on transfer function of the decoupling model. Finally, by superposing pseudo random binary signals （PRBS） onto each reference input, the parameters of grid side converter controller are estimated with Levenberg-Marquardt optimization algorithm. The variance of 50 estimation results shows the estimation error of each parameter, and the estimation results validate the method proposed in this paper.