引用本文: | 杨玮林,胡官洋,许德智.基于连续控制集的永磁同步直线电机模型预测控制[J].控制理论与应用,2021,38(10):1671~1682.[点击复制] |
YANG Wei-lin,HU Guan-yang,XU De-zhi.Model predictive control of permanent magnet linear synchronous motor based on continuous control set[J].Control Theory and Technology,2021,38(10):1671~1682.[点击复制] |
|
基于连续控制集的永磁同步直线电机模型预测控制 |
Model predictive control of permanent magnet linear synchronous motor based on continuous control set |
摘要点击 1932 全文点击 612 投稿时间:2020-09-09 修订日期:2021-09-02 |
查看全文 查看/发表评论 下载PDF阅读器 |
DOI编号 10.7641/CTA.2021.00610 |
2021,38(10):1671-1682 |
中文关键词 永磁同步直线电机 二次优化 模型预测控制 连续控制集 |
英文关键词 permanent magnet linear synchronous motor quadratic optimizations model predictive control continuous control set |
基金项目 国家自然科学基金项目(61903158, 61973140), 江苏省自然科学基金项目(BK20180595)资助. |
|
中文摘要 |
永磁同步直线电机(permanent magnet linear synchronous motor, PMLSM)目前多被应用于直线牵引系统, 例
如轨道交通、无绳电梯等. 传统的永磁同步直线电机预测控制主要考虑有限控制集模型预测控制(finite-control-set
model predictive control, FCS–MPC), 在一个系统采样周期从备选的开关状态中选择一个相对最优的开关状态送入
逆变器中. 该方法的计算量通常随着预测步长的增加呈几何增长, 因而限制了其广泛使用. 本文针对PMLSM提出一
种基于二次优化的连续控制集模型预测控制(continuous-control-set model predictive control, CCS–MPC)策略. 该方
法在每个周期内选择两组开关状态送入逆变器, 表现为两个相邻电压矢量的合成, 因而可以达到更为平滑的控制效
果. 策略结合了FCS–MPC中的扇区划分原理, 将扇区中的两个相邻非零矢量和一个零矢量等效合成为二次优化的
最优控制矢量. 与此同时, 在二次优化的框架下CCS–MPC有效地避免了多步预测控制中计算量过大的问题. 仿真与
实验结果表明在相同条件下, 所提方法相较于空间矢量调制以及FCS–MPC能获得更好的PMLSM控制效果. |
英文摘要 |
Permanent magnet linear synchronous motor (PMLSM) is currently widely used in linear traction systems,
such as rail transportation, cordless elevators, etc. The traditional predictive control for PMLSM mainly considers the
finite-control-set model predictive control (FCS–MPC), which selects the optimal switch state among all the candidate
ones into the inverter during one sampling period of the control system. The computational burden of this method usually
exponentially increases with the prediction horizon length, which limits its practical applications. This paper proposes a
continuous-control-set model predictive control (CCS–MPC) strategy that is based on quadratic optimizations for PMLSM.
It selects two sets of switching states to feed into the inverter in each sampling period, which is equivalent to the synthesis
of two adjacent voltage vectors. Thus, the resultant control performance becomes smoother. The strategy combines the
principle of the sector division in FCS–MPC, and employ two adjacent non-zero vectors and one zero vector to synthesize
the optimal control vector obtained by quadratic optimizations. Meanwhile, CCS–MPC effectively avoids the problem
of excessive calculation in multi-step predictive control under the framework of quadratic optimizations. Simulations and
experiments reveal that under exactly the same conditions, the proposed method can achieve better PMLSM control performance
than the space vector modulation method and FCS–MPC. |
|
|
|
|
|