A thermal monitoring and parameter tuning scheme for induction machines

This paper presents a method for tracking the stator and rotor temperatures in three-phase induction machines. The proposed algorithm can be used to predict insulation failure in the machine and to tune the parameters of a thermal model for overload protection. Microprocessor-based thermal protection has become widely available as a viable replacement for the traditional, mechanical trip element. If the thermal condition of the machine changes, however, existing microprocessor-based devices have no means of recognizing or correcting the built-in model. The proposed scheme enables the protection unit to adapt to changes in the thermal conductance between the rotor core and the ambient atmosphere, which could be caused by a broken fan, clogged air vents, or other factors that affect machine cooling. The algorithm compares one rotor speed estimate derived from terminal measurements to a more accurate estimate obtained by digital analysis of the rotor slot ripple. The error between the two estimates provides a corrective input to the thermal model. By continuing this process over a period of time after startup, the thermal time constant of the rotor can be derived. Additionally, by combining this approach with trend analysis, a supervisory algorithm can detect incipient failures in the electrical integrity of the machine.