神经网络-滑模控制在STATCOM系统中的应用

Application of GA optimized RBF network-sliding model controller for AC servo system

研究了基于STATCOM无功补偿的交流系统电压波动和闪变调节与控制问题,建立了STATCOM在d-q坐标系的数学模型,提出了基于直接电流控制策略的神经网络-滑模控制方法:采用电压外环滑模控制和电流内环神经网络-滑模控制的双环控制策略,克服了传统PI控制不能自适应调整控制参数的缺点,实现了系统的自适应控制。数值仿真结果验证了本方法的有效性,在电压波动和闪变恶劣条件下STATCOM能够快速响应、控制品质好,交流系统能够稳定工作,控制性能比PID控制好。

The STATCOM reactive compensation based on voltage fluctuations and flicker of AC system is studied.The mathematical model of STATCOM in the d-q coordinate system is established and the neural network-sliding mode control based direct current control method proposed.In this new control system,the sliding mode control is applied to the outer voltage loop and the neural network-sliding mode controls employee for the inner current loop.By doing so,the lack of addictiveness of PID controller can be overcome.The numerical simulation model is established to evaluate and validate the proposed control method.The analysis shows that the proposed controller is effective for the AC system control under severe disturbance,and STATCOM can quickly compensate the voltage fluctuations and flicker.Additionally,the control performance of the proposed method is superior to the PID controller.