ASMADO‐based refined anti‐disturbance low‐speed control of PMSM

Abstract

Permanent magnet synchronous motors (PMSMs) have expectant low‐speed servo performance. However, complex nonlinear disturbances restrict the performance and lead to speed fluctuations, especially for small inertia PMSMs where the current‐loop cannot be used in special operating conditions. This paper divides complex disturbances into periodic determinable disturbances and other indeterminate disturbances according to their characteristics. A refined anti‐disturbance control (RADC) method is proposed to enable the inner loop to compensate for disturbances in targeted manners. Then ideal low‐speed servo can be achieved using a simple outer loop controller. The proposed RADC consists of two parts. One is a backpropagation neural network based periodic disturbances compensator that is trained using the data from an iterative learning controller. The other is an adaptive sliding‐mode‐assisted disturbance observer that rapidly observes and compensates the residual disturbances. The convergence of the overall algorithm is analyzed. The effectiveness of the proposal is also verified by experiments.