On Synchronous Motor-Driven Mechanical Systems: Analysis and Linear Parameter-Varying Control

Abstract

The actuator dynamics problem is a challenging aspect of controller design considering mechatronic systems. While it is tempting to neglect such effects, industrial applications tend to bear the burden of nonlinear motor characteristics. Ideally, the motor time constant is negligible compared with the system dynamics, but when the system enters marginal operating regions, it may increase abruptly, preventing fast and smooth actuation and may lead to oscillations, reduced performance, or even instability. In its simplest form, the permanent magnet synchronous motor (PMSM) is equivalent to a linear system under coordinate transformation coupled with linearizing feedback. However, to account for the imperfections of field-oriented control (FOC), the motor model is redefined. This transformation leads to a less favorable flat output selection scheme, causing exact linearization to be a less viable option. Therefore, an alternative approach is proposed, in which the nonlinear motor character is mapped into a linear parameter-varying (LPV) control framework. The goal of this article is to design a robust control architecture boasting comparable performance to a baseline industrial solution while guaranteeing stability and improved robustness to parameter variations via a smoother actuation character. The presented method is validated on an electric power steering system equipped with a PMSM, where the high-level controller runs on AutoBox and all other elements are identical to a real-life system.