Online Estimation of Rotor Temperature in Induction Motors Based on Parameter Identification

This study proposes an in-service induction machine (IM) rotor's temperature estimation method. The proposed method is based on the noninvasive measurements of current and potential sensors. First of all, using the rotor slot harmonics superimposed on the stator's current, the rotor speed is extracted. Secondly, the first few cycles of the IM's starting process is used to mimic the locked-rotor test condition while the IM is in-service. Using the aforementioned analogy, the initial value of the rotor's resistance is calculated directly. A merit of the proposed method is that it requires significantly less time to estimate the rotor's resistance compared to the traditional methods in literature, and convergence speed of the parameter identification can be improved significantly. Consequently, the rotor resistance can be on-line identified via the Model Reference Adaptive (MRA) method. Thereby, the rotor temperature is calculated by the resistance-temperature linear relation of metals. Experimental validation is performed on a 5.5kW motor, and the estimated results agreed with the tested ones.