Abstract

The performance of electrical machines is ultimately limited by thermal issues. To facilitate the modelling and speed up the time to market of novel designs, the accuracy, efficiency and reliability of thermal models has to be improved. In the literature, the majority of thermal models of electrical machines are based on lumped-parameter thermal networks, which are both efficient and simple to implement. Accuracy is the main concern related to this approach. Moreover, the outputs of these thermal models are average temperatures, which are not a suitable constraint for the machine design. In this work we introduce an improved lumped parameter configuration based on a novel structure: the Distributed Losses and Capacitance (DLC) element. The DLC is able to accurately estimate local temperatures at steady-state or during transients with a high Biot number or high aspect ratios. We validate the approach experimentally, creating a thermal model of an external-rotor Surface-mounted Permanent Magnet (SPM) machine. The error in the estimation of the end-windings temperature is within 4%.

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