多参数失配下鲁棒型双馈风机模型预测电流控制

MPC control of robust type DFIG on multi-parameter mismatch

风机系统模型预测电流控制(model predictive current control,MPCC)是基于系统模型实现的,存在系统参数鲁棒性较差的问题。为解决该问题,提出了一种带有新型滑模趋近律(new sliding mode veaching law,NSMRL)的滑模观测器(sliding mode observer,SMO)结合MPCC策略。首先,建立基于电压定向的双馈风机数学模型;其次,建立无差拍模型预测控制系统;最后,构建一种改进的等速伸展项和指数到达项的滑模趋近律,该新型滑模趋近律包括系统状态变量和滑模面函数的功率项,可以以两种不同的形式来表示趋近律,新滑模趋近律的作用是使系统状态更快、抖振更小的收敛到滑模平面,从而提高参数鲁棒性。研究结果表明,该方法与无滑模MPCC和等速趋近律MPCC相比,能有效地抑制系统固有抖振,并提高系统状态到达滑模面的速度。同时,可将平均转矩脉动、电流脉动和相电流总谐波失真(total harmonic distortion,THD)降低10.54%、24.16%和5.12%。因此,所提出的NSMRL-SMO-MPCC方法能够通过在线补偿扰动实现参数失配情况下对电流的可靠预测,从而提升系统的参数鲁棒性。

The model predictive current control(MPCC)is based on the system model in DFIG,which has the problem of poor robustness of system parameters.In this paper,a sliding mode observer(SMO)combined with MPCC control strategy based on new sliding mode reaching law(NSMRL)is proposed.At first,the mathematical model of DFIG by voltage orientation is established.Then,the model of the deadbeat control system is established.Finally,An improved sliding mode reaching rate of isometric extension term and exponential arrival term is constructed,NSMRL includes the system state variable and the power term of the sliding mode surface function,and can represent the approach law in two different forms.The function of the new sliding mode approach rate is to make the system state converge to the sliding mode plane faster and with less buffeting,thus improving the parameter robustness.The results show that proposed method can suppress the inherent chattering of the system effectively and can also increase the speed of the system state reaches the sliding mode surface compared to that of the MPCC without sliding mode and MPCC with constant velocity approach rate methods.Specifically,the average torque ripple,current ripple and phase current THD can be reduced by 10.54%,24.16%and 5.12%,respectively.Therefore,the NSMRL-SMO-MPCC control strategy proposed in this paper can predict the current more reliable by compensating the disturbance online to improve the parameter robustness of the system while the parameters are mismatched.