Iron Pole Shape Optimization of IPM Motors Using an Integrated Method

Abstract

Abstract —An iron pole shape optimi zation method to reduce cogging torque in Interior Permanent Magnet (IPM) motors is developed by using the reduced basis technique coupled by finite element and design of experiments methods. Objective function is defined as the minimum cogging torque. The experimental design of Taguchi method is used to build the approximation model and to perform optimization. This method is demonstrated on the rotor pole sh ape optimization of a 4-poles/24-slots IPM motor. Index Terms—IPM motor, Cogging torque reduction, iron pole shape optimization, reduced basis technique, FEA, Design of Experiments (DOE) I. I NTRODUCTION Brushless permanent motors have been widely used in industrial applications because of their efficiency and power density. In particular, the spoke-type motor, which can concentrate flux from permanent magnets, has a high torque density per unit volume resulting from its high reluctance torque and structure for concentrating flux from permanent magnet. However, cogging torque and torque ripple can be increased due to a distortion of air gap flux density distribution in the IPM motors. This drawback is due to its asymmetric magnetic reluctance on the edge of the permanent magnet. There have been some developments for rotor pole shape optimization. Table I summarizes the literature by showing the type of optimization approach. Cogging torque reduction in IPM motors was performed using a rotor with flux barriers [1]–[3]. Eccentric pole design was proposed to compensate the armature reaction for reducing torque ripple [4]. To obtain better performance, the rotor pole shape was varied and divided into three parts. These consist of two end parts of eccentric surfaces and one uniform surface. Response surface methodology (RSM) was used during the design process [5]. The continuum shape design sensitivity formula and the finite element method are employed to calculate the sensitivity of flux linkage to the design variables, which determine the shape of iron pole piece [6]. They used B-Spline parameterization to optimize the design variables in order to provide the back-EMF waveform as close as possible to a sinusoidal form. The main challenge of current optimization methods, especially for complex one, is the number of design variables required for rotor pole shape optimization and the generality of the procedure. The first method has poor performance due to the loss in flux barriers. The second method is only suitable for motor rotation in one direction. Also, the later method is not a general approach to achieve the optimum rotor pole shape profile.