Abstract

Thermal analysis is a key aspect in electrical machine design and for those operating with severe duty where a high-reliability thermal analysis is required. This paper presents a comprehensive analysis of techniques for calibrating generic thermal model of electrical machines. The proposed approach combines two existing experimental methods commonly used in thermal testing of electrical devices. The first method uses a short-duty transient excitation to derive the winding thermal parameters, whereas the second method uses a steady-state excitation to estimate the thermal parameters related to the heat transfer from the winding. Both experimental methods are based on a well-defined heat source provided by a constant current (dc) excitation. For reference, alternative existing methods for calibrating thermal models, are usually combining the winding thermal data predictions and a set of experimentally derived temperatures from dc tests. Such approach, however, might be inadequate if not supported by extensive experimental data. This frequently leads to a significant rate of inaccuracy, thus having a significant impact on the reliability of temperature predictions. The main advantage of newly proposed approach is the reducing of the number of variable parameters and a more systematic calibration process. Additionally, a detailed test procedure for generic thermal model calibration is presented and discussed.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call