一种表面式永磁同步电机有限状态集模型预测直接转矩控制

Finite control set model predictive direct torque control of surface permanent magnet synchronous motor

针对常规有限状态集模型预测直接转矩控制成本函数需要设计磁链和转矩控制权重系数的问题,设计了将磁链和转矩控制转换为相对误差率控制的成本函数,从而消除了权重系数。建立了基于定子磁链坐标系的表面式永磁同步电机预测模型,并以逆变器生成的7个基本电压矢量作为备选电压矢量,仿真验证了提出成本函数的可行性。仿真发现:由于受备选电压矢量自身限制,转矩动态变化下,磁链会出现较大的波动。基于电压矢量作用规律,给出了可同时满足转矩和磁链控制要求的备选电压矢量集合。仿真结果表明:该电压矢量集合可有效抑制动态下的磁链波动。为了提高模型预测控制的实时性,对备选电压矢量数目和预测控制模型进行了简化,并基于STC89C51单片机平台,实验对比了不同备选电压矢量集合和预测模型预测控制的运算时间和运算结果。仿真和实验结果表明了采用精简的备选电压矢量集合和简化模型可在控制性能基本不变的前提下,有效减小运算负担,提高系统实时性。

Aiming at the problem of weight factor design of cost function in model predictive direct torque control( MP-DTC),cost function without weight factor is proposed. The control of stator flux and torque was converted to the control of stator flux and error rates. Based on MP-DTC model of surface permanent magnet synchronous motor( SPMSM),7 basic voltage vectors of inverter were used as candidate vectors.Simulation results testify the effectiveness of the proposed strategy. Simulation results show that,due to limitation of candidate vectors,stator flux has high ripples at dynamic state. Based on effects of voltage vectors on stator flux and torque,candidate voltage vectors set was given to satisfy the control of stator flux and torque at the same time. Simulation results show the proposed candidate vectors set can suppress stator flux ripple. In order to improve real-time performance of MP-DTC system,candidate vectors set and predictive model were simplified. And MP-DTC using different candidate voltage vectors sets and predictive models were carried out using STC89 C51 chip. Simulation and experimental results show the simplified candidate voltage vectors sets and predictive models can decrease calculation burden and has equivalent control performance.