Pseudo-three dimensional theoretical prediction of local temperature in an interior permanent magnet motor using a lumped parameter heat transfer model

Abstract

Interior permanent magnet (IPM) motors have been receiving much attention as special drive applications for industrial robots, servomechanisms, and electric vehicles because of their low cost and high efficiency. In general, magnets in IPM motors show the highest temperatures, particularly under high speed operating conditions, causing durability problems in motor performance. The purpose of this study was twofold. One goal is to propose effective thermal network modeling of circular electric motors by using a lumped-parameter network method, which is based on an axisymmetric motor structure. For this purpose, a pseudo-three dimensional thermal network model (TNM) is developed to reflect periodical repetitions of coil and yoke in the angular direction. The other is to conduct effective thermal analysis by manipulating key design and operating variables. As a validation process of the TNM model, temperatures at representative motor components were compared under the same operating conditions. Both theoretical and numerical results by the TNM and three dimensional computation results show minor temperature differences less than 5% in all parts of the motor. In subsequent applications, the TNM was utilized to predict maximum coil and magnet temperatures, derive heat transfer paths, and estimate heat losses through various directions of the motor. The TNM enables us to predict the heat distribution in various types of motors and derive optimized geometrical parameters in a convenient way.