同步(交流)
控制理论(社会学)
对偶(语法数字)
永磁同步电动机
磁铁
同步电动机
职位(财务)
计算机科学
模型预测控制
控制工程
控制(管理)
工程类
拓扑(电路)
电气工程
人工智能
文学类
艺术
经济
财务
作者
Xiuyun Zhang,Zhiwei Sun,Jianyu Chen,Zhibiao Zhao,Tao Liu,Weijie Hou
标识
DOI:10.1109/tpel.2025.3605365
摘要
Model predictive control (MPC) mitigates synchronous error accumulation, dynamic response mismatch, and precision-complexity trade-off in traditional multi-motor system. However, constrained numerical solutions in MPC present significant computational challenges. To address this, this paper proposes a predictive position synchronization control strategy based on geometric constraints for dual permanent magnet synchronous motor (PMSM) system. First, an incremental MPC framework is constructed, whose embedded integral action not only effectively compensates for short control horizons but also enables high-precision multi-step prediction. Second, a cost function incorporating tracking error, synchronous error and control increment is established to derive the current control increment. To reduce computational complexity, a constrained geometric mapping method is proposed: with control horizonNc= 1, current and voltage constraints are mapped to a 2D plane to form a feasible region, converting the optimal control increment solution into determining the tangent point between the cost function's elliptic surface family and this feasible region. Rapidly localizing the tangent via real-time query and calculation of analytical solutions significantly reduces online computational load. Experimental results show the proposed strategy can improve computational efficiency by approximately 47% and dynamic response speed by approximately 57% vs traditional QP strategy, resolving the precision-complexity trade-off in multi-motor synchronization while ensuring accuracy.
科研通智能强力驱动
Strongly Powered by AbleSci AI