Induction machine deep-bar and thermal models for sensorless IRFOC application

Abstract

Thermal and deep-bar (skin) effect models have been created that can be used together for parameter estimation in a vector control application. A thermal lumped-parameter model that accurately determines the stator and rotor representative temperatures is proposed and experimentally verified. Analytical and lumped-parameter solutions for deep-bar effect are generalised for any rotor bar shape and a ladder network approximate solution is finally selected for its good precision against simplicity characteristic. Once both models are verified, the relationship between both phenomena is found, and a global estimation is included into the speed estimator of an indirect rotor field oriented control (IRFOC) scheme. This scheme has been experimentally tested on a 1.5 kW inverter-driven induction motor system showing the effectiveness of the proposed estimation in both the steady and non-steady state.