五桥臂逆变器-双异步电机调速系统的零矢量交错模型预测控制

发布时间：2019-05-05 06:09

【摘要】：五桥臂逆变器-双异步电机调速系统采用模型预测控制相比于传统的矢量控制具有更好的动态性能,易于实现系统的全局性能优化,但却存在稳态性能差、计算量大难以实现、无法消除公共桥臂开关动作需求限制等问题。对此,提出一种零矢量交错模型预测控制(ZVI-MPC)方法,将空间矢量的概念与矢量合成原理引入到模型预测中,在一个采样周期中对每台逆变器均作用两个电压矢量(一个非零矢量和一个零矢量),利用无差拍理论计算两逆变器的非零矢量作用时间,结合公共桥臂开关需求,调节各矢量作用时间并按照零矢量交错原则进行开关状态分布,以避免出现公共桥臂开关动作需求冲突。该方法在任意采样周期内只需考虑12种有效矢量组合,大大减少预测控制计算量。仿真和实验结果表明采用零矢量交错模型预测控制,两台电机可独立稳定运行,且均具有良好的动静态性能,相比于传统模型预测控制方法,本文所提方法稳态性能更理想,且计算量显著减小,易于实现。
[Abstract]:Compared with the traditional vector control, the five-leg inverter-dual asynchronous motor speed regulation system with model predictive control has better dynamic performance. It is easy to optimize the global performance of the system, but it has poor steady-state performance and is difficult to achieve with a large amount of computation. There is no way to eliminate the problem such as the limit of the action requirement of the common bridge arm switch. In this paper, a zero vector staggered model predictive control (ZVI-MPC) method is proposed, in which the concept of space vector and the principle of vector composition are introduced into model prediction. In a sampling period, two voltage vectors (one non-zero vector and one zero vector) are applied to each inverter. The non-zero vector action time of the two inverters is calculated by using the theory of non-difference beat, and the switching requirements of the common bridge arm are combined with the non-zero vector action time of the two inverters. Adjust the action time of each vector and distribute the switch state according to the principle of zero vector interleaving, so as to avoid the conflict of the action demand of common bridge arm switch. In this method, only 12 effective vector combinations are considered in any sampling period, which greatly reduces the computational complexity of predictive control. Simulation and experimental results show that using zero vector staggered model predictive control, the two motors can operate independently and stably, and both have good dynamic and static performance. Compared with the traditional model predictive control method, the steady performance of the proposed method is better than that of the traditional model predictive control method. Moreover, the computation amount is reduced significantly, and it is easy to realize.
【作者单位】：北方工业大学北京市变频技术工程技术研究中心;
【基金】：国家自然科学基金(51477003) 北京市自然科学基金(4152013) 北京市科技新星计划(Z161100004916049) 北方工业大学“长城学者后备人才培养计划”(XN070007)资助项目
【分类号】：TM343;TM464

【参考文献】