On-Line Identification of Winding Resistances and Load Torque in Induction Machines

Abstract

Rotor and stator resistances along with load torque are typically uncertain quantities in induction machines. The machine heating makes the winding resistances vary during operation whereas the load torque strictly depends on the application. All those variables need to be on-line estimated to improve the drive performances and in particular to minimize the power loss at steady state. A new adaptive observer is designed in this brief. It is able to exponentially estimate the motor fluxes and to identify the aforementioned critical parameters from stator currents/voltages and rotor speed measurements. In contrast to other solutions proposed in the literature, rotor and stator resistances are not estimated on the same time scale. New insights on the behavior of an intuitively inspired observer are thus given through a detailed stability proof, which does not rely on linearization arguments around constant operating conditions. Persistency of excitation conditions, which only depend on exogenous signals to the estimation error system, are analyzed in detail and a clear physical interpretation is presented. Key features of the proposed solution are the overall simplicity of the estimation scheme, the low dimension of the regressor matrix (being exactly related to the number of unmeasured or uncertain quantities) and the exponential convergence to zero of the estimation errors. Simulations confirm the correctness of all the mathematical derivations. Experimental results show the effectiveness of the proposed approach in implementing an advanced version of the indirect field oriented control scheme: the uncertain rotor flux modulus reference that minimizes the power loss at steady state can be actually estimated and imposed.