In this study, a new approach based on adaptive dynamic programming (ADP) is proposed to control permanent magnet synchronous motors (PMSMs). The objective of this study is to control the torque and consequently the speed of a PMSM when an unknown load torque is applied to it. The proposed controller achieves a fast transient response, low ripples and small steady-state error. The control algorithm uses two neural networks, called critic and actor. The former is utilised to evaluate the cost and the latter is used to generate control signals. The training is done once offline and the calculated optimal weights of actor network are used in online control to achieve fast and accurate torque control of PMSMs. This algorithm is compared with field oriented control (FOC) and direct torque control based on space vector modulation. Simulations and experimental results show that the proposed algorithm provides desirable results under both accurate and uncertain modelled dynamics. Although the performance of FOC method is comparable with ADP under nominal conditions, the torque and speed response of ADP is better than FOC under realistic scenarios, that is, when parameter uncertainties exist.

中文翻译：

---

基于自适应动态规划的永磁同步电动机最优转矩控制

在这项研究中，提出了一种基于自适应动态规划（ADP）的新方法来控制永磁同步电动机（PMSM）。这项研究的目的是在施加未知负载转矩时控制PMSM的转矩，进而控制其速度。所提出的控制器实现了快速的瞬态响应，低纹波和较小的稳态误差。控制算法使用两个神经网络，称为评论家和演员。前者用于评估成本，而后者用于生成控制信号。一旦离线就进行训练，并且将所计算的actor网络的最佳权重用于在线控制，以实现对PMSM的快速准确的转矩控制。将该算法与基于空间矢量调制的磁场定向控制（FOC）和直接转矩控制进行了比较。仿真和实验结果表明，该算法在准确的和不确定的动力学模型下均提供了令人满意的结果。尽管在标称条件下FOC方法的性能可与ADP相媲美，但在实际情况下（即存在参数不确定性时），ADP的转矩和速度响应要优于FOC。