UPFC的神经网络PI控制方法研究

A Neuron Network PI Controller for UPFC

为了满足工程上广泛应用的PI控制器的性能要求,又能克服一般PI控制器参数整定和调整的困难,基于dqo同步旋转坐标系下的UPFC串联侧的数学模型,采用BP神经网络PI控制器分别对有功和无功进行控制,使PI控制器的参数可以自动整定,省去了繁琐的人工整定过程,并能随系统的变化在线调节,使有功和无功达到较好的控制性能。

A BP neuron network PI control scheme is presented for a mathematics model of the seriespart of the UPFC in this paper, and the parameter of the PI controller can be found automatically and adjusted online. Simulation results prove that the scheme improves the self- adaptive capability of PI controller, and lowers interaction of the real and reactive power.