Comparisons of Linear Characteristic for Shape of Stator Teeth of Hall Effect Torque Sensor

Similar documents
Analysis and Experiments of the Linear Electrical Generator in Wave Energy Farm utilizing Resonance Power Buoy System

Torque Ripple Reduction Using Torque Compensation Effect of an Asymmetric Rotor Design in IPM Motor

Experimental Assessment of Unbalanced Magnetic Force according to Rotor Eccentricity in Permanent Magnet Machine

Parameter Prediction and Modelling Methods for Traction Motor of Hybrid Electric Vehicle

Robust optimal design of a magnetizer to reduce the harmonic components of cogging torque in a HDD spindle motor

Development of a new linear actuator for Androids

Water-Cooled Direct Drive Permanent Magnet Motor Design in Consideration of its Efficiency and Structural Strength

The initial magnetization curve shows the magnetic flux density that would result when an increasing magnetic field is applied to an initially

Analysis and Case Study of Permanent Magnet Arrays for Eddy Current Brake Systems with a New Performance Index

Finite Element Analysis of Hybrid Excitation Axial Flux Machine for Electric Cars

Loss analysis of a 1 MW class HTS synchronous motor

338 Applied Electromagnetic Engineering for Magnetic, Superconducting, Multifunctional and Nano Materials

Limited Angle Torque Motors Having High Torque Density, Used in Accurate Drive Systems

The Optimum Design of the Magnetic Orientation of Permanent Magnets for IPMSM under Hot Environments

Modeling and Design Optimization of Permanent Magnet Linear Synchronous Motor with Halbach Array

COGGING torque is one of the major sources of vibration

Equal Pitch and Unequal Pitch:

Physica C 468 (2008) Contents lists available at ScienceDirect. Physica C. journal homepage:

An Introduction to Electrical Machines. P. Di Barba, University of Pavia, Italy

Mathematical Modeling and Dynamic Simulation of a Class of Drive Systems with Permanent Magnet Synchronous Motors

Fachgebiet Leistungselektronik und Elektrische Antriebstechnik. Test Examination: Mechatronics and Electrical Drives

Proposal of short armature core double-sided transverse flux type linear synchronous motor

Design and analysis of Axial Flux Permanent Magnet Generator for Direct-Driven Wind Turbines

Projectile s Velocity Effect for Voltage Induced at Sensing Coil for Applying to Air Bursting Munition

Keywords: Electric Machines, Rotating Machinery, Stator faults, Fault tolerant control, Field Weakening, Anisotropy, Dual rotor, 3D modeling

1205. Optimal design of quadratic electromagnetic exciter

Design and Characteristic Analysis of LSM for High Speed Train System using Magnetic Equivalent Circuit

Mathematical Modelling of an 3 Phase Induction Motor Using MATLAB/Simulink

Cogging Torque Reduction in Surface-mounted Permanent Magnet Synchronous Motor by Axial Pole Pairing

Effect of Reaction Plate on Performance of Single-Side Linear Induction Motor in Different Speeds and Frequencies with Finite Element Method

Finite Element Model of a Magnet Driven Reed Switch

Optimal Design of a Permanent Magnetic Actuator for Vacuum Circuit Breaker using FEM

W. A. D. M. Weragoda and N.T. Medagedara

Measurement And Testing. Handling And Storage. Quick Reference Specification Checklist

Research Article Trial Application of Pulse-Field Magnetization to Magnetically Levitated Conveyor System

Levitation and Thrust Forces Analysis of Hybrid-Excited Linear Synchronous Motor for Magnetically Levitated Vehicle

Cogging Torque Reduction in Surface-mounted Permanent Magnet Synchronous Motor by Axial Pole Pairing

Definition Application of electrical machines Electromagnetism: review Analogies between electric and magnetic circuits Faraday s Law Electromagnetic

Electromagnetic Forming Process Analysis Based on Coupled Simulations of Electromagnetic Analysis and Structural Analysis

Efficient Search Method of Deperming Protocol for Magnetic Silence of Vessel

Keywords: Principle Of Escapement Mechanism, Tower Escape Apparatus, Mechanism Design.

3D FEM MODELS OF LINEAR ELECTRICAL MACHINES USED IN FAULT DETECTION STUDIES

Transient Analysis of Three Phase Squirrel Cage Induction Machine using Matlab

A Linear Motor Driver with HTS Levitation Techniques

Predictive Cascade Control of DC Motor

Nonlinear dynamic simulation model of switched reluctance linear machine

Design and Analysis of Interior Permanent Magnet Synchronous Motor Considering Saturated Rotor Bridge using Equivalent Magnetic Circuit

Effect of the number of poles on the acoustic noise from BLDC motors

International Journal of Advance Engineering and Research Development SIMULATION OF FIELD ORIENTED CONTROL OF PERMANENT MAGNET SYNCHRONOUS MOTOR

Digital Control of Permanent Magnet Synchronous Motors

Magnetic Quantities. Magnetic fields are described by drawing flux lines that represent the magnetic field.

Accurate Joule Loss Estimation for Rotating Machines: An Engineering Approach

Doubly salient reluctance machine or, as it is also called, switched reluctance machine. [Pyrhönen et al 2008]

CURRENT-CARRYING CONDUCTORS / MOVING CHARGES / CHARGED PARTICLES IN CIRCULAR ORBITS

III.Sources of Magnetic Fields - Ampere s Law - solenoids

the machine makes analytic calculation of rotor position impossible for a given flux linkage and current value.

IE1206 Embedded Electronics

PRINCIPLE OF DESIGN OF FOUR PHASE LOW POWER SWITCHED RELUCTANCE MACHINE AIMED TO THE MAXIMUM TORQUE PRODUCTION

Coupled magnetic equivalent circuits and the analytical solution in the air-gap of squirrel cage induction machines

Measurements in Mechatronic design. Transducers

RESEARCH ON REDUCING COGGING TORQUE IN PERMANENT MAGNET SYNCHRONOUS GENERATORS

Analysis of Half Halbach Array Configurations in Linear Permanent-Magnet Vernier Machine

Power density improvement of three phase flux reversal machine with distributed winding

Electrical to mechanical - such as motors, loudspeakers, charged particle deflection.

The next two questions pertain to the situation described below. Consider a parallel plate capacitor with separation d:

Abstract. 1 Introduction

MECH 3140 Final Project

Chapter 2: Fundamentals of Magnetism. 8/28/2003 Electromechanical Dynamics 1

IE1206 Embedded Electronics Le2

DESIGN AND ANALYSIS OF AXIAL-FLUX CORELESS PERMANENT MAGNET DISK GENERATOR

A New Moving-magnet Type Linear Actuator utilizing Flux Concentration Permanent Magnet Arrangement

Analytical Model for Permanent Magnet Motors With Surface Mounted Magnets

Course no. 4. The Theory of Electromagnetic Field

A DARK GREY P O N T, with a Switch Tail, and a small Star on the Forehead. Any

Variable Sampling Effect for BLDC Motors using Fuzzy PI Controller

Modelling and Analysis of Permanent Magnet Electrodynamic Suspension Systems

Version The diagram below represents lines of magnetic flux within a region of space.

IEEE Transactions on Applied Superconductivity. Copyright IEEE.

Synchronous Machines

Electromagnetism. Topics Covered in Chapter 14:

APPLICATION OF THE FINITE ELEMENT METHOD TO MODEL THE NONLINEAR VOICE COIL MOTION PRODUCED BY A LOUDSPEAKER MAGNET ASSEMBLY.

Research on Permanent Magnet Linear Synchronous Motor Control System Simulation *

ANALYSIS OF INDUCTION MOTOR WITH BROKEN BARS AND CONSTANT SPEED USING CIRCUIT-FIELD COUPLED METHOD

MAGNETIC FIELD PRODUCED BY A TILE PERMANENT MAGNET WHOSE POLARIZATION IS BOTH UNIFORM AND TANGENTIAL

Study on influence factor of braking torque in liquid-cooled eddy current retarder with a structure of two salient poles

Comparative investigation of permanent magnet linear oscillatory actuators used in orbital friction vibration machine

Analytical Calculation of Air Gap Magnetic Field Distribution in Vernier Motor

1 Written and composed by: Prof. Muhammad Ali Malik (M. Phil. Physics), Govt. Degree College, Naushera

3 d Calculate the product of the motor constant and the pole flux KΦ in this operating point. 2 e Calculate the torque.

2002 Prentice Hall, Inc. Gene F. Franklin, J. David Powell, Abbas Emami-Naeini Feedback Control of Dynamic Systems, 4e

Third harmonic current injection into highly saturated multi-phase machines

Static Characteristics of Switched Reluctance Motor 6/4 By Finite Element Analysis

Electrical Machines and Energy Systems: Operating Principles (Part 2) SYED A Rizvi

Student Exploration: Electromagnetic Induction

Optimization Design of a Segmented Halbach Permanent-Magnet Motor Using an Analytical Model

Chapter 7. Chapter 7. Electric Circuits Fundamentals - Floyd. Copyright 2007 Prentice-Hall

Maxwell s Equations:

Modelling of Permanent Magnets Track in the Finite Element Analysis of Linear Synchronous Motor

A Study on the Tube of Integral Propeller Shaft for the Rear-wheel Drive Automobile Using Carbon Composite Fiber

Thermal Simulation for Design Validation of Electrical Components in Vibration Monitoring Equipment

Transcription:

Journal of Magnetics 17(4), 285-290 (2012) ISSN (Print) 1226-1750 ISSN (Online) 2233-6656 http://dx.doi.org/10.4283/jmag.2012.17.4.285 Comparisons of Linear Characteristic for Shape of Stator Teeth of Hall Effect Torque Sensor Boram Lee 1, Young Sun Kim 2 *, and Il Han Park 1 1 School of Information and Communication Engineering, Sungkyunkwan University, Suwon 440-746, Korea 2 Department of Electrical and Electronic Engineering, Joongbu University, Geumsan 302-702, Korea (Received 25 November 2012, Received in final form 4 December 2012, Accepted 12 December 2012) Electric Power Steering (EPS) system is superior to conventional Hydraulic Power Steering (HPS) system in aspect of fuel economy and environmental concerns. The EPS system consists of torque sensor, electric motor, ECU (Electric Control Unit), gears and etc. Among the elements, the torque sensor is one of the core technologies of which output signal is used for main input of EPS controller. Usually, the torque sensor has used torsion bar to transform torsion angle into torque and needs linear characteristic in terms of flux variation with respect to rotation angle of permanent magnet. The torsion angle of both ends of a torsion bar is measured by a contact variable resistor. In this paper, the sensor is accurately analyzed using 3D finite element method and its characteristics with respect to four different shapes of the stator teeth are compared. The four shapes are rectangular, triangular, trapezoidal and circular type. Keywords : Electric Power Steering (EPS), torque sensor, Finite Element Method (FEM), shape design, hall effect 1. Introduction Nowadays, cars are one of the essential elements in our life. However, they are main culprits for environmental problems and energy waste. Therefore, reducing the environmental impacts of cars is one of the most urgent tasks to be dealt with. With various technologies being developed for the purpose, there are on-going efforts to replace existing auto parts with new ones that are more eco-friendly and energy-efficient. These methods seem to be feasible in the near future. As another method, electricity is utilized to enhance energy efficiency and reduce weight, making cars more environmentally friendly. Power steering system is set up to improve driver's convenience and steering accuracy. HPS (Hydraulic Power Steering) system, which is a kind of power steering system, uses hydraulic pressure, supplemental power from an engine as shown in Fig. 1. The system rotates wheels in accordance with driver s steering will. It operates a hydraulic pump with the power generated from the engine. This system has some weaknesses, such as too complicated and heavy structure; and, due to the use of engine power, low engine output and fuel efficiency. EPS (Electric Power Steering) system is in the making to solve the problems of HPS system. EPS system is an equipment to supplement driver's steering through an electric motor. It has many benefits. First, car weight will be reduced by the removal of complicated hydraulic equipments. Second, the environmental impact will lessen because it doesn't use oil. Third, fuel efficiency will be enhanced with an electric motor running only when steering happens [1]. EPS system consists of an electric motor, ECU (Electric Control Unit), torque sensor etc. In particular, a torque sensor is one of the core parts in the system. It measures driver s steering will and sends the results to the ECU, which, in turn, commands the electric motor to generate supplemental power. Many countries are studying EPS The Korean Magnetics Society. All rights reserved. *Corresponding author: Tel: +82-41-750-6279 Fax: +82-41-750-6402, e-mail: yskim@joongbu.ac.kr Fig. 1. HPS system and EPS system. 2012 Journal of Magnetics

286 Comparisons of Linear Characteristic for Shape of Stator Teeth of Hall Boram Lee, Young Sun Kim, and Il Han Park systems and some have succeeded in developing the systems. In this paper, the Hall effect torque sensor is accurately analyzed using 3D FEM. Normally, 2D FEM is used in Finite Element Analysis. In this case, it is necessary to assume that the length of the analyzed model is infinite in a certain direction. However, a torque sensor is small in size and of complex structure, which means that 2D method must be replaced with 3D method. Also, characteristic of flux density measured by Hall sensor should represents linear changes to rotating angle to catch an angle of steering wheel accurately. This characteristic depends on the shape of stator teeth. Finally, various types of stator teeth are proposed to achieve this purpose. The results can provide the various data to machine designers according to the shape of teeth and permanent magnet array. 2. Hall Effect Torque Sensor of EPS System EPS systems are designed to operate a motor with electric power generated from a battery and create steering torque through a reduction gear on the steering column. A torque sensor measures steering angles and controls a steering column according to the angles, providing stability at high speeds and steering convenience at low speeds. Fig. 2 illustrates the structure and process of an EPS system [2]. When a driver turns a steering wheel, torque is generated between output and input columns. Such torque is transformed into torsion angles by the torsion bar connecting output and input columns. At this point, the torque sensor detects the torsion angles and changes them into electric signals, which will be sent to the ECU. Then, the ECU adjusts torque according to the signals and operates a motor to offset the torque. As a result, the driver is able to control a steering system with less effort. A torque sensor is one of the core technologies in the EPS system for reducing steering effort. The sensor identifies torque a driver s strength to turn a steering wheel, directions of rotation, and steering angles. After that, it figures out necessary degrees and directions of supplemental power for changing driving conditions. A torque sensor is mainly divided into two functional parts. The one is for measuring torsion angles of the torsion bar, and the other is for effectively transmitting the torsion angles to an ECU. Largely, there are two types of torque sensor: contact and non-contact. In the beginning, contact types, which measure torsion angles of both ends of a torsion bar using a contact variable resistor, are most used. However, their signal processing is not accurate and the maintenance is difficult. So, now, non-contact types are preferred to contact ones. Noncontact types are, again, divided into some subtypes. Among them, inductance type is using coils to measure torque by inductance changes. Optical type is using optical devices. Hall type is using Hall Effect [3, 4]. Fig. 3 shows a Hall effect torque sensor, the analyzed model in this paper. A Hall effect torque sensor is composed of permanent magnets, stators, collector, and Hall IC. Permanent magnets generate magnetic flux. Stators draw such flux into one spot in a linear manner according to torsion angles. Hall IC identifies the magnetic flux in the collector. The mechanism of the torque sensor is as followed. The permanent magnet located in the base point creates magnet flux, which will pass between upper and lower teeth. In this case, magnetic flux density in the Hall IC is zero. However, if the permanent starts to rotate, the balance between the upper and lower teeth is broken, resulting in movement of magnetic flux to the Hall sensor. Fig. 2. (Color online) Structure and driving process of EPS system. Here, ECU requires accurate information on the torque applied by driver to the steering wheel. This torque signal usually comes from the shift measurement between two shafts linked by a linear torsion bar. Fig. 3. (Color online) Exploded view of the Hall effect torque sensor. Each segment of permanent magnet generate magnetic flux which are measured by Hall IC according to rotating angle of magnet.

Journal of Magnetics, Vol. 17, No. 4, December 2012 287 The Hall sensor identifies measures torsion angles. The Hall IC needs a Hall type sensor, which transforms magnetic flux into voltage signals by using Hall effect. 3. Formulation for 3D Finite Element Analysis Ampere s circuital law, irrespective of the displacement current, is defined as (1) [5] H = J where H is magnetic field intensity [A/m] and J is current density [A/m 2 ]. With a permanent magnet, magnetic flux density B [T ] can be calculated using (2). (1) B = μ 0 ( H + M ) (2) where μ 0 is permeability of vacuum [H/m], M is magnetization of the permanent magnet [A/m]. When M is redivided by magnetization M 0, which constantly exists in the permanent magnet, (3) will result. M = χ H + M 0 (3) where χ is magnetic susceptibility. When (3) is applied to (2), the final form of magnetic flux density can be expressed by using the residual magnetic flux density,, as [6]. B = μ 0 ( H +χ H + M 0 )=μ 0 (1+χ) H +μ 0 M 0 = μ H + B r (4) (5) can result from the relation between magnetic flux density and magnetic vector potential. B = A (5) Where A denotes the magnetic vector potential including x, y and z-components. Accordingly, when (4) and (5) are applied to (1), a governing equation of the magnetic field including the permanent magnet will be produced. ν( A B r ) = J (6) Where ν is magnetic resistivity [m/h]. When the above governing equation is formulated by tetrahedron element using 3D Finite element method, equation (7), a 12 by 12 element matrix will result as follow [7-10]. Here, K ij are matrix coefficients related to the structure and material properties of the system and F i is forcing term including the current density, volume of tetrahedral mesh and residual magnetic flux density. B r (7) The individual elements of (8) can be expressed using tetrahedron volume V e, permeability μ and coefficients b i, c i, d j in (9). And details of coefficients include a coordinate of the vertex of a tetrahedron. K xx = 1 µ -- 1 ----------- ( c 36V e i c j + d i d j ), K xy = 1 µ -- 1 -----------c 36V e i b j, K xz = 1 µ -- 1 ----------d 36V e i b j 1 K yx = -------------b 1 µ 36V e i c j, K yy = 1 µ -- 1 ----------- ( b 36V e i b j + d i d j ), K yz = 1 µ -- 1 ----------d 36V e i c j 1 1 K zx = -- -----------b µ 36V e i d j, K zy = 1 µ -- 1 -----------c 36V e i d j, K zz = 1 µ -- 1 ---------- ( b 36V e i b j c i c j ) (9) Forcing terms for each component are represented as follows. F xi = 1 μ --1 6 ( B ry d i B rz c i ) + 1 4 --V e J xi F yi = 1 μ --1 6 ( B rz b i B rx d i ) + 1 4 --V e J yi F zi = 1 μ --1 6 ( B rx c i B ry b i ) + 1 4 --V e J zi b i = ( 1) i { y j ( z k z l ) + y k ( z l z j ) + y l ( z j z k )} c i = ( 1) i { z j ( x k x l ) + z k ( x l x j ) + z l ( x j x k )} d i = ( 1) i { x j ( y k y l ) + x k ( y l y j ) + x l ( y j y k )} where i, j, k, l = 1, 2, 3, 4 (cyclic permutation) 4. Analysis of Hall Effect Torque Sensor (8) (10) (11) 4.1. Magnetic field profile of permanent magnet When analyzing the torque sensor, it is important to know the characteristics of magnetization distribution for accurate analysis. A samarium cobalt magnet was used as an analysis model. The magnet is 20 [mm] in outside diameter, 15.5 [mm] in inside diameter, 4 [mm] in height. And the number of poles in the ring-type permanent magnet is 16 segments. The advantages of samarium cobalt are good quality related to temperature, and high remnant magnetic flux density and coercivity. On the other hand, the disadvantages are insufficient resource reserves, thus difficulty in securing it. Residual magnetic flux density of samarium cobalt is 1.149 [T] and coercivity 450 [ka/m]. The result can be illustrated as in Fig. 4 which shows the

288 Comparisons of Linear Characteristic for Shape of Stator Teeth of Hall Boram Lee, Young Sun Kim, and Il Han Park Fig. 4. (Color online) (a) Magnetic flux density distribution of permanent magnet, (b) Tetrahedral mesh and (c) magnetic flux density distribution of torque sensor in 3D finite element analysis. Fig. 5. (Color online) Calculated magnetic flux density profile along the surface of the permanent magnet. Fig. 6. (Color online) Magnetic flux density at the Hall sensor according to rotation angles. finite element and magnetic flux density distribution of the torque sensor when the permanent magnet rotates 11 compared with reference position. When magnetic flux heads for z-axis, the magnetization distribution is bottomup for N pole and top-down for S pole. When we measured the magnetic flux density 0.5 [mm] away from the surface of the permanent magnet, the result, as shown in Fig. 5, was obtained. 4.2. Magnetic field profile of torque sensor Torque sensor used steering apparatus decide the magnitude and direction of supporting power sensing torque and rotating angle of steering wheel. So, the torque sensor using permanent magnet needs accurate analysis because the characteristic is sensitive to small movement and its structure is very complex. Also, torque sensor should generate changes of magnetic flux density to the rotating angle in specific operating ranges. Magnetic field analysis of assembled Hall effect torque sensor was performed by using 3-dimensional finite element method. Fig. 6 shows the values of magnetic flux density at the Hall IC according to rotation angles of the permanent magnet of the torque sensor, ranging from 0 to 360. It was found that the values of magnetic flux density are similar to sinusoidal curve. Fig. 7 shows the magnitude of magnetic flux density around the collector part using 3D arrow expression when the torque sensor rotates 11. Especially, the enlargement of the magnetic flux density profile when the torque sensor rotates from 22.5 to 22.5 is shown in Fig. 8(a) to know the linearity of flux density characteristic. Also, Fig. 8(b) shows more specific range with reference line. In range of operating angle about 4.5, the characteristic of the torque sensor should satisfy linearity within this angle. Here the result of numerical analysis did satisfy the linearity in above ranges. 4.3. Shape design of stator teeth Because torque sensors need accurate response characteristics, the following performance enhancements are

Journal of Magnetics, Vol. 17, No. 4, December 2012 289 Fig. 7. (Color online) Magnetic flux density distributions around the collector according to the rotation angle. Magnitude of flux density is the maximum when the permanent magnet pole and stator teeth are aligned exactly. Fig. 8. (Color online) (a) Magnetic flux density of specific range ( 22.5 ~22.5 ) and (b) compared view with linear reference. required at the Hall sensor of collector. First, linearity of the flux characteristic should be large to get accurate information from the sensor. Secondly, linear range should be wide to assure the stability of the sensor. Finally, the magnetic flux density should be high to measure flux easily at the Hall IC. The above enhancements, we proposed four kinds of stator teeth prototypes. The four prototypes named trian-

290 Comparisons of Linear Characteristic for Shape of Stator Teeth of Hall Boram Lee, Young Sun Kim, and Il Han Park magnetic flux density. 5. Conclusion Fig. 9. Design shape of teeth for numerical analysis: (a) triangular type, (b) circular type, (c) trapezoidal type and (d) rectangular type. This paper performed the characteristic analysis and shape design of stator teeth of Hall effect torque sensor used electric power steering system using 3D finite element method. Hall effect torque sensor using permanent magnet has complex structure and its characteristic is very sensitive to small mechanical movement. So, it is necessary for torque sensor to have linearity to operate wheels of the car accurately. After 3D finite element analysis of torque sensor, conclusions were obtained. First, the magnetic field of a torque sensor with complex 3D structure is accurately analyzed and characteristic curves of output flux at the collector are compared according to four shapes of stator teeth. Secondly, linear output flux profile of torque sensor with range of shift angles 4.5[ ] ~ +4.5[ ]. Finally, these results can predict the characteristics of torque sensor for various shape and permanent magnet array. Also, we will realize optimal design in a follow-up paper using the prototype that obtained the best result. References Fig. 10. (Color online) Magnetic flux density of various stator teeth shapes to compare its characteristic. gular type, circular type, trapezoidal type and rectangular type are illustrated in Fig. 9. Magnetic field analysis for these types is performed to compare with each other. Under the same condition for size, the researcher carried out the analysis of the 4 prototype. As a result, magnetic flux density varies according to the type of stator teeth as seen in Fig. 10. Among the 4 prototypes, the circular type achieved the highest values in linearity, linear range, and [1] C. H. Hu, Second International Symposium on IITA 3, 436 (2008). [2] A. W. Burton, IEEE Contr. Syst. Mag. 23, 30 (2003). [3] K. Miyashita, T. Takahashi, and S. Kawamta, IEEE Trans. Magn. 26, 1560 (1990). [4] D. Angleviel, D. Frachon, and G. Masson, SAE 2006-01- 0939 (2006). [5] W. H. Hayt, Engineering Electromagnetics, McGraw-Hill, New York (1989). [6] S. J. Salon, Finite Element Analysis of Electrical Machines, Kluwer Academic, Norwell, MA (1995). [7] M. A. Alhamadi, R. Wang, and N. A. Demerdash, IEEE Trans. Magn. 27, 5016 (1991). [8] P. Campbell, M. V. K. Chari, and J. D'Angelo, IEEE Trans. Magn. 17, 2997 (1981). [9] N. A. Demerdash, T. W. Nehl, F. A. Fouad, and O. A. Mohammed, IEEE Trans. Power Apparatus Sys. 100, 4104 (1981). [10] D. H. Im, Finite Element Method in Electrical Engineering, Dongmyeongsa, Seoul (1987).