论文部分内容阅读
双馈式风力发电系统所处的工况环境往往较为复杂多变,受此影响,采用传统PI控制的网测变换器会出现抗干扰能力差、工作稳定性下降等问题。在传统PI控制器的基础上引入BP神经网络,并利用BP神经网络具有的自适应、自学习等特性对传统PI控制器参数进行实时在线整定,保证PI控制器的最优控制状态。仿真结果显示,BP神经网络PI控制器在复杂的工况下具有更高控制精度和鲁棒性,提高了网测变换器的抗干扰能力和运行效率。
The conditions of the doubly-fed wind power generation system are often complex and changeable. Affected by this, the network-based transducers adopting the traditional PI control may have problems of poor anti-interference capability and decreased work stability. Based on the traditional PI controller, the BP neural network is introduced, and the parameters of the traditional PI controller are adjusted online in real time by using the adaptive and self-learning features of the BP neural network to ensure the optimal control status of the PI controller. The simulation results show that the BP neural network PI controller has higher control accuracy and robustness under complicated working conditions and improves the anti-interference capability and operating efficiency of the network-based transducer.