Online Parallel Estimation of Mechanical Parameters for PMSM Drives via a Network of Interconnected Extended Sliding-Mode Observers

Abstract

The accurate estimation of mechanical parameters is essential for the control optimization and condition monitoring of permanent-magnet synchronous motor (PMSM) drives. In this article, an online parallel estimation scheme is proposed based on a network of three interconnected extended sliding-mode observers. With a two-step mechanism derived from the developed network, this scheme can simultaneously achieve the accurate real-time observation of three mechanical parameters, including the viscous friction coefficient, the moment of inertia, and the load torque. The first step utilizes the presented network to precisely acquire the Coulomb friction coefficient by offering only one speed. With aiding from this estimated coefficient, the second step attains the high-precision online identification of the three concerned parameters by compensating the Coulomb friction torque to counteract its adverse effect. These two steps are implemented utilizing the identical network, thus, furthest avoiding increasing the estimation complexity. Besides, any additional knowledge other than the rotor speed and electromagnetic torque is no longer required for guaranteeing the asymptotic convergence of the estimation errors. Sufficient evaluations from simulations and experiments confirm the validity of the proposed scheme.