中文摘要：

双馈异步风力发电机（DFIG）网侧换流器的主要任务是保持直流母线电压的稳定、输入电流的正弦性和输入功率因数的恒定.一般在设计控制器时,常对电机模型进行一些理想假设,忽略掉一些次要的影响因素.在不同的运行工况下,电机参数会或多或少地发生变化,导致控制器的跟踪效果变差.为解决设计网侧换流器控制策略时由于参数变化带来的扰动不确定问题,提出了网侧换流器电流内环的基于BP神经网络整定的PI控制及电压外环的PI控制的复合控制策略.通过MATLAB/Simulink软件仿真分析,结果表明,相比于传统的PI控制,直流母线电压在电流内环的基于BP神经网络整定的PI控制及电压外环的PI控制的复合控制策略下能更快地达到给定值.因此,本文所提出的电流内环的基于BP神经网络整定的PI控制及电压外环的PI控制的复合控制策略能很好地实现对换流器输出电压和输入电流的有效控制,减小了由于参数变化等原因对换流器造成的不利影响,提高了系统的鲁棒性,具有实际意义和应用价值.

英文摘要：

The main task of doubly-fed induction wind-power generator（ DFIG） grid side converter is to maintain the DC bus voltage stability,ensure inputting sinusoidal current and control input power factor. Due to the design of the controller which carry out some of the idealized assumptions,ignore some minor factors of motor,the motor parameters at different operating conditions will have more or less change,which lead to deterioration in tracking performance of controller. To solve the grid-side converter control issues strategy that parameter change brings perturbations,this paper presents a grid side hybrid control strategy which based on BP neural network tuning PI control in the inner current loop and PI control in the voltage outer-loop. Through MATLAB / Simulink software simulation,compared to conventional PI control,the DC bus voltage reached a given value more quickly under the composite control strategy based on inner BP neural network tuning PI control and voltage outer PI control. The simulation results show that the control strategy,which based on current inner BP neural network tuning PI control and voltage outer PI control,can well realize the effective control of the inverter output voltage and input current. It＇s more important this method can not only reduce the adverse effects on the grid side converter due to the variations parameters and other factors,but also increases the robustness of the system. It has an important actual significance and practical application value.