采用多标量模型的感应电机简化模型预测转矩控制算法

Simplified Model Predictive Torque Control Algorithm of Induction Motor Using Multiscalar Model

模型预测转矩控制(model predictive torque control,MPTC)基于定子磁链和定子电流预测值间接获取转矩预测值,会造成转矩预测误差,引起转矩纹波。且其价值函数中,转矩与磁链的量纲不同,需引入加权因子加以调整。该文引入定子磁链与定子电流矢量的矢积和标积,选择新的状态变量,导出新的感应电机多标量模型,并将其与模型预测转矩控制相结合,提出一种多标量模型预测转矩控制方法。运用该文所提方法直接预测转矩和无功转矩,能够简化算法,减小转矩预测误差,提高转矩响应速度,抑制转矩纹波。此外,该文定义一个统一加权因子的价值函数,构成价值函数的转矩与无功转矩的量纲相同,避免繁琐的加权因子选取设计。最后,通过实验验证所提方法的有效性和可行性。

Model predictive torque control(MPTC)indirectly obtains the torque prediction value based on the stator flux and stator current prediction value,which will cause torque prediction error and torque ripple.In the cost function,the dimensions of torque and flux are different,so the weighting factor needs to be introduced for adjustment.In this paper,the vector product and scalar product of stator flux and stator current vector were introduced,some new state variables were selected,a new multiscalar model of induction motor was derived and combined with model predictive torque control,and a multi-scalar model predictive torque control method was proposed.Using the proposed method to directly predict the torque and reactive torque could simplify the algorithm,reduce the torque prediction error,improve the torque response speed and suppress the torque ripple.In addition,this paper defined a new cost function by using unified weighting factor.The cost function constituted with torque and reactive torque had the same dimension and weighting factor,which avoided the cumbersome a weighting factors design.Finally,experiments verified the effectiveness and feasibility of the proposed method.