See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/324921334

Analysis of Active Magnetic Bearings

Thesis · April 2018

CITATIONS READS

3 5,544

2 authors, including:

Mihir Pewekar

5 PUBLICATIONS   10 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Analysis of active Magnetic Bearings View project

Study and analysis of Non Pneumatic Tires View project

All content following this page was uploaded by Mihir Pewekar on 13 May 2018.

The user has requested enhancement of the downloaded file.

 Analysis of Active Magnetic Bearings

Submitted in fulfillment of the requirements of the degree of

Bachelor of Engineering

By
Mihir Mangesh Pewekar A-855
Anendra Gajendra Potawad A-856
Mohnish Rohidas Pujari B-863
Rishab Rajendra Rane B-865

Under the guidance of

Prof. Sambhaji D. Gaikwad

Department of Mechanical Engineering

MCT’s Rajiv Gandhi Institute of Technology,
Juhu-Versova Link Road, Versova, Andheri (W),
Mumbai-53.

UNIVERSITY OF MUMBAI
2017-2018
Scanned by CamScanner
Scanned by CamScanner
Scanned by CamScanner
 Acknowledgment

First and firmly we express our deepest gratitude and heartfelt thanks to our project guide Prof.
S. D. Gaikwad for their expert guidance, constant encouragement, constructive criticism and
inspiring advice throughout this course of project.

I grateful to my father Mangesh Pewekar for his constant guidance, immense moral and
emotional support. I am also grateful to my mother Swati Pewekar for her immense moral and
emotional support. I am grateful to my grandfather Dharmraj Pewekar for guiding me through
the process of creation of accurate windings.

I am grateful to Mukesh Kurlawala for providing us the tools and machinery required to
realize this project. A special thanks to Abhishek Dhanukar of Dhanukar Electricals Pvt.
Ltd. for guiding us and giving us a market overview.

I am extremely grateful to ANSYS Inc. for sponsoring the Academic Research Licenses. A
special thanks to Vanaja Narayanan for coordinating and officiating these licenses. I am also
grateful to EM Works Inc. for providing their Electromagnetic Simulation Suite under the
Education License. A special thanks to Mike Curic for officiating this sponsorship.

We are especially thankful to Dr. S. U. Bokade, Head of Mechanical Engineering Department

and Dr. Udhav Bhosle, Principal of the Institute for their administrative support and constant
encouragement during the various phases of our project work.

iv
 Abstract

Stress developed due to the contact between mating surfaces i.e. Contact stresses and forces are
the main factors inhibiting the further increase of efficiency of a system. Bearings are an
integral part of all major mechanical systems. Thus, it has become essential to eliminate such
induced losses. Also, due to the presence of contact stress, wear is induced, decreasing the life
of the bearing. Hydrodynamic bearings eliminate the contact losses but are subjected to the
viscous losses. Active Magnetic bearing is a contactless bearing technology that aims to
eliminate the drawbacks of roller bearings as well as journal bearings. Active Magnetic bearing
is based on fundamentals of electromagnetic levitation. The load bearing rotor is suspended at
the center of the stator. In current practice, this technology is being implemented in large scale
turbomachinery. Advantages of magnetic bearings are Lower or no frictional losses, absence
of auxiliary systems, no hot shutdown, Low maintenance, very high speeds can be attained and
Can handle variable loads. This project intents to analyze the existing practices in providing
suspension to the rotor, and to indigenously design an active magnetic bearing based on the
analysis. Another aspect of this project deals with the theoretical analysis of the active magnetic
bearing. This is achieved through the estimation of slew rate and specific Load capacity. The
estimation of force efficiency factor is achieved by means of experimental analysis on the
Active magnetic bearing. We also aim to simulate magnetic fluxes and the forces generated in
commercially available electromagnetic solvers based finite element method. The outcomes of
our analysis will aid in the shrinking of this technology as well as aid in implementation to a
wide range of mechanical systems.

v
 Table of Contents

Caption Page No.

Acknowledgement iv

Abstract v

List of Tables viii

List of Figures x

Chapter 1 – Introduction 1

1.1 Aim of the project 2

1.2 Organization of report 2

1.3 Current use of AMB technology 3

1.4 Characteristics of Active magnetic bearings 4

1.5 Market Analysis 7

Chapter 2 -Literature Survey 8

Chapter 3- Study of Magnetic Bearing Forces 13

3.1 Governing Laws of an Active Magnetic Bearing 13

3.2 Behavior of magnetic Force 15

3.3 Mathematical Model of Differential Actuator 17

3.4 Magnetic Materials 19

Chapter 4- Rotor dynamics 22

4.1 Introduction 22

4.2 Governing D.E. for Rotor. 23

4.3 Balance Quality Grade 26

vi
Chapter 5- Electronic Components and Circuits 29
5.1 Electronic Components 29

5.2 Control Techniques 31

Chapter 6- Design of Test Rig 35

6.1 Geometric Parameter Calculation 35

6.2 Load Capacity 38

6.3 CAD- Model of Test Rig 39

6.4 Electrical Circuits 41

Chapter 7- Practical Limitations and Dimension-Less Parameters 43

7.1 Specific Load Capacity 43

7.2 Inherent Instability 44

Chapter 8- FEA Simulation 47

8.1 Preprocessing 47

8.2 Solving 48

8.3 Post-processing and Results 49

Chapter 9- Experimentation 52

Chapter 10- Conclusion and Future Scope 55

10.1 Conclusions 55

10.2 Future Scope 56

References 58

vii
 List of Figures

Figure No. Caption Page No.

1.1 Typical Industrial Active Magnetic Bearing 4

2.1 Typical example of a control loop implemented in AMB 11

Typical schematic representation of Active magnetic
2.2 12
bearing elements
Comparison of the supporting forces to mechanical
3.1 17
Spring

3.2 One Pole model of Magnetic Bearing 17

3.3 Magnetic fields around conductors 19

3.4 B-H diagram, hysteresis loop 20

4.1 Free Body Diagram of the rotor 23

5.1 Arduino Uno Micro-controller 29

5.2 Schematic of Eddy current and Capacitive sensors 31

6.1 Winding constraints 35

6.2 CAD model of Test-Rig 40

6.3 Fabricated Model of Test-Rig 40

6.4 Actual core used in AMB 41

Equivalent electrical circuit of any two diametrically

6.5 41
opposite poles

7.1 Schematic representation of magnetic bearing 45

8.1 CAD Models for FEA analysis 48

8.2 Generated Mesh 48

8.3 Contour Result of magnetic flux for single pole actuation 49

viii
 Magnetic field intensity plot showing opposite polarity of
8.4 50
consecutive poles

8.5 Magnetic flux vector plot 50

8.6 Contour result of magnetic flux for complete actuation 50

8.7 Magnetic flux vector plot 51

9.1 Experimental Setup 52

9.2 Comparison plot of results 54

ix
 List of Tables

Table No. Caption Page No.

1.1 Questionnaire for market survey 7

4.1 Balance Grade and Its Applications 26

5.1 Comparison of different control techniques 34

6.1 Geometric Parameters 38

8.1 Mesh parameters 49

9.1 Theoretical Values of force generated 53

9.2 Experimental Values of force generated 54

x
 List of Abbreviations

AMB Active Magnetic Bearings

xi
 Chapter 1

Introduction

Bearings are an integral part of all major mechanical systems, they are used to reduce friction
when there is relative motion between two parts like shaft and housing. Generally, bearings are
classified in two categories rolling contact and sliding contact bearings.

In rolling contact bearings, stresses developed due to the contact between mating surfaces i.e.
contact stresses (hertz contact theory) and forces are the main factors inhibiting the further
increase of efficiency of a system. Due to the presence of contact stress, wear is also induced.
In Sliding Contact bearings, the viscosity of the lubricant induces the friction loss.

Thus, it has become essential to eliminate such induced losses. Magnetic bearing is a contactless
bearing technology that aims to eliminate the drawbacks of roller bearings as well as journal
bearings. Magnetic bearing is based on fundamentals of electromagnetic levitation. The load
bearing rotor is suspended at the center of the stator. In current practice, this technology is being
implemented in large scale turbomachinery.

1
1.1 Aim of the project

Our aim during project is to conduct an in-depth analysis of radial magnetic bearings. The aim
of project is present certain characteristic curves representing load, power and current.

Objectives:

• To design an Active Magnetic bearing.

• To optimize the power requirement.
• To study the materials that can be implemented to increase the efficiency.
• To analyze the performance at different loading conditions (steady state/transient).
• To determine the stability of the bearing system.
• To derive dimensionless factors to standardize the use of active magnetic bearing.

The objectives will be achieved by means of certain live experiments on an indigenously

designed test Rig system as well as results from commercially available electromagnetic
solvers based on finite element method. The outcomes of our analysis will aid in the shrinking
of this technology as well as aid in implementation to a wide range of mechanical systems.

1.2 Organization of Report

We begin here with the introduction to the basic concept underlying the magnetic bearings, the
types and sub types. The advancements that have propelled the research and the paradigms
associated with this revolutionary concept. Then we move on to its general uses and the
advantage over conventional alternative used. A part of this chapter is dedicated to the market
analysis and survey we performed prior to the beginning of the project. This chapter includes
the current state of magnetic in India and the scope of this technology in India.
In chapter 2, we discuss the basic concepts evolved because of years of investigation through
the literature survey. In chapter 3, we discuss in depth about the laws that govern the behavior
of the active magnetic bearings. In section 3.2, we introduce the mathematical model of the
basic a two-pole electromagnet from which the radial active magnetic bearing was born. In
chapter 4, the dynamics of the rotor and its effect on the bearing system is discussed. The rotor
dynamics is not studied, in depth in due course of this project. The balance quality of the rotor
is also discussed, which is important when viewing through an industrial perspective. In chapter
5, we discuss about the Electronic components and the circuits which are used as the heart of
the active magnetic bearings. In chapter 6, the design of the test rig is put forth with the design

2
considerations. The chapter 7 constitutes the practical limitations and non-dimensionless
factors that exist in the active magnetic bearing technology. The practical limitations give us
the direction of improvement. The dimensionless factors provide us direction towards
standardization of active magnetic bearings, so that magnetic bearings can also subjected to
selection design. The chapter 8 constitutes the visualization of the magnetic field through the
Electro-magnetic solvers. This analysis and visualization of the magnetic fields help in the
improvement of the geometrical configurations of the active magnetic bearings. The final
chapter constitutes the results and conclusions drawn from the analysis done. This chapter will
also constitute of the scope of work in future in this project. This chapter includes the current
state of magnetic in India and the scope of this technology in India.

1.3 Current use in industry

The various advantages of the magnetic bearing have led to applications

mainly in the five following areas:
• Vacuum and cleanroom systems: The bearings will not suffer from any mechanical wear
or give rise to any related contamination, and if necessary, the bearings can even be
arranged outside the vacuum container with field forces acting through the container
walls. The absence of aerodynamic drag losses and the low energy consumption of the
bearings is a welcome feature for flywheels for energy storage.
• Machine tools: A main advantage is the high precision that can be attained and the high
rotational speed with relatively high load capacity. This is useful for heavy-duty high-
speed milling of aluminum. The high speed is an essential requirement in the precision
grinding of small parts.
• Medical devices: A specific application is the use of magnetic bearings in an artificial
heart pump, or more precisely, in a left ventricular assist device intended to assist an
ailing heart in keeping the pumped blood at a desired rate, which is needed to provide
the circulatory requirements.
• Turbo-machinery: The main application area of AMB is turbomachinery. The area
covers small turbo-molecular pumps up to turbogenerators and compressors in the
Megawatt range. Turbo-generators in the 300 MW range are in an early planning stage
already. An advantage is the possibility of controlling and damping vibrations and
obtaining a well-defined dynamic behavior. Furthermore, it is possible to simplify
machine construction, as there are no bearing fluids, usually oil, that must be kept away
3
 from the process fluid by seals. Other important features that have been corroborated
by practical experience are the inherent means for self-control and diagnosis, the very
low maintenance costs, and the low energy consumption. With the availability of very
high efficiency power electronics, the need for turbo-generators running at a low 50/60
Hz speed or the necessity of coupling a high-speed gas turbo engine to a reduction gear
for driving a generator has decreased, and for high-speed machinery with high power
density, the AMB is the bearing of choice. Even for aero engines, generating just thrust
and electric energy for the all-electric airplane of the future, research on AMB
applications is going on.
• Superconducting bearings: The advances of superconducting bearings with their
inherent passive stability promise a future alternative to active magnetic bearings, see
Figure. 1.1. However, to achieve damping properties in a superconductive suspension
for rotating machinery the use of additional active dampers by AMBs may still be
necessary.

Figure 1.1 Typical industrial Magnetic bearing (photo courtesy SIEMENS)

1.4 Characteristics of Active magnetic bearings

In the following chapters, the most widely used bearing types: the active electromagnetic
bearing AMB, and to some extent the self-bearing motor, will be presented in more detail. First,
at this introductory level, some specific properties, which render the AMB particularly useful
for some applications, and may also open up new applications, will be summarized:
1. The property of being free of contact, and the absence of lubrication and contaminating
wear allow the use of such bearings in vacuum systems, in clean and sterile rooms, or
for the transport of aggressive or very pure media, and at high temperatures.

4
2. The gap between rotor and bearing amounts typically to a few tenths of a millimeter,
but for specific applications it can be as large as 20 mm. In that case, of course, the
bearing becomes much larger.
3. The rotor can be allowed to rotate at high speeds. The high circumferential speed in the
bearing – only limited by the strength of material of the rotor – offers the possibilities
of designing new machines with higher power concentration and of realizing novel
constructions. About 350 m/s are achievable, for example by using amorphous metals
which can sustain high stresses and at the same time have very good soft-magnetic
properties, or by binding the rotor laminations with carbon fibers. Design advantages
result from the absence of lubrication seals and from the possibility of having a higher
shaft diameter at the bearing site. This makes the shaft stiffer and less sensitive to
vibrations.
4. The low bearing losses, which at high operating speeds are 5 to 20 times less than in
conventional ball or journal bearings, result in lower operating costs.
5. The specific load capacity of the bearing depends on the type of ferromagnetic material
and the design of the bearing magnet. It will be about 20 6 Gerhard Schweitzer N/cm2
and can be as high as 40 N/cm2. The reference area is the cross-sectional area of the
bearing. Thus, the maximum bearing load is mainly a function of the bearing size.
6. The dynamics of the contact-free hovering depends mainly on the implemented control
law. The control is implemented by a microprocessor, which makes the design very
versatile. Thus, it is possible to adapt the stiffness and the damping, within physical
limits, to the bearing task and even to the actual state of operation and the rotor speed.
The terms stiffness and damping include the conventional static parts, known as spring
and damping constants, and the frequency dependent part, the dynamic stiffness. This
renders it possible, for example, to use the bearings for vibration isolation, to pass
critical speeds with no large increase in vibration amplitude, or to stabilize the rotor
when it is excited by non-conservative disturbances.
7. Retainer bearings are additional ball or journal bearings, which in normal operation are
not in contact with the rotor. In case of overload or malfunction of the AMB they must
operate for a very short time: they keep the spinning rotor from touching the housing
until the rotor comes to rest or until the AMB regains control of the rotor. The design
of such retainer bearings depends on the specific application and despite a variety of
good solutions still needs special attention.
8. The unbalance compensation and the force-free rotation are control features where the
vibrations due to residual unbalance are measured and identified by the AMB. The

5
 signal is used to either generate counteracting and compensating bearing forces or to
shift the rotor axis in such a way that the rotor is rotating force-free.
9. The precision with which the state of the rotor can be controlled, for example the precise
rotation about a given axis, is mainly determined by the quality of the measurement
signal within the control loop. Conventional inductive sensors, for example, have a
measurement resolution of about 1/100 to 1/1000 of a millimeter.
10. Diagnostics are readily performed, as the states of the rotor are measured for the
operation of the AMB anyway, and this information can be used to check operating
conditions and performance. Even active diagnostics are feasible, by using the AMB as
actuators for generating well defined test signals simultaneously with their bearing
function.
11. The AMB has the potential to be a key element in a smart machine. The AMB can make
use of its measured state information to optimize the operation of the whole machine. It
contributes to the overall process control and supports the safety and reliability
management.
12. The lower maintenance costs and higher life time of an AMB have been demonstrated
under severe conditions. Essentially, they are due to the lack of mechanical wear.
Currently, this is the main reason for the increasing number of applications in
turbomachinery. The maintenance and 1 Introduction and Survey 17 reliability
properties can be even further improved by making use of the smart machine concept.
13. The cost structure of an AMB is that of a typical mechatronics product. The costs for
developing a prototype, mainly because of the demanding software, can be rather high.
On the other side, a series production will lower the costs considerably because of the
portability of that software.
14.The design of an AMB for a specific application requires knowledge in mechatronics,
i.e. in mechanical and electrical engineering, and in information processing, in addition
to knowledge about the specific application area. Therefore, a close cooperation
between AMB producer and the manufacturer of application machinery (OEM) is
necessary.

6
1.5 Market Analysis:

With an objective of conducting an all rounded analysis of an Active Magnetic bearings, we

began by understanding the supply and demand of this type of mechanical system. This task
was achieved by means of a market survey. We visited Nagdevi street, which is the bearing hub
in Mumbai, in the month of October. The basic aim of conducting a market survey was:

• Level Familiarity of any type of magnetic bearings in the market.

• To find the source of demand, if any supply exists.

We initialized our survey with a basic set of questions given in table 1.1. Results of the analysis
showed us that this technology presently has no industrial scope. The momentary conclusion
that we can draw is that this situation can only be boosted by the emphasis on Academic
research in Magnetic bearings.

We got in touch with, SKF bearings which is the leading bearing manufacturer in the world,
though SKF has a magnetic bearing research facility, it is not available in India. From the
literature survey, we have found recently IIT Mumbai has conducted a research on Passive
magnetic bearings in collaboration with Larsen and Toubro research division, Mumbai [11].

Table 1.1 Questionnaire for market survey

Sr. No. Questions Result of Survey

1. Have you heard of Magnetic bearings? No, from 99% of the shops
visited
2. Do have demand for magnetic bearings? No demand from industry as
well as Academics
3. If yes, from whom and in what quantity? N/A
4. If yes, as an entrepreneur would try aid the debut No, from 99% of the shops
of such bearings in the Market? visited

1.5.1 Ability of fabrication Industry to aid research:

As discussed in previous section, we chose to pay emphasis on academic research. Now, we

evaluated the ability of the fabrication industry to aid research through its manufacturing
capability. This is necessary for the assurance of the feasibility of the project.

7
Indian markets have a lot of potential in terms of manufacturing capability. Mumbai City within
itself contains some prominent industrial areas which undertake such fabrication jobs. We
selected the Saki Naka- Marol Naka Industrial Area for our fabrication based on proximity from
our college.

The laminated cores which we require are manufactured through sheet metal presswork. This
would require creation of a die block. Another method which was found based on literature
review [2], was manufacturing the block by CNC milling machine. Saki Naka- Marol Naka
Industrial Area has the ability aid research in this area. This evaluation of the capability can
provide the initial push required to convince technical education bodies to venture into this
field.

8
 Chapter 2

Literature Review

In the previous chapter we have given the aim of project of on active magnetic bearing and its
industrial use. In this chapter we discuss the basic concepts evolved in the process of
development of such a technology.

Active magnetic bearings (AMBs), made of eight poles find their wider usage in flywheel
energy storage systems, turbomolecular vacuum pump, artificial heart blood pump, etc., due to
their built-in fault diagnostics, vibration free operation, and friction and wear [2] characteristics.
Each pair of pole makes an electromagnet consists of iron-core and copper-winding. On passing
the current through the copper wire of electromagnet, it attracts the rotor. Using four pairs of
electromagnets, the motion of rotor in horizontal and vertical (x and y) directions can be
controlled. The detailed description of working of electromagnet has been provided by
Schweitzer and Maslen [1].

An electromagnet have been shown in the Figure. 6.1, where Ws is the thickness of the outer
rim of the stator; dp is the inner diameter of the stator at the poles; L is the width of the bearing
along the shaft axis; l is length of the pole; Wp is the width of the pole; and a is the half angle
made by the two consecutive poles at the center of the bearing. The electromagnet is made of
N turns of copper wire having cross-sectional diameter dw. The A shaft of diameter Dshaft is
9
made of paramagnetic material and a ferromagnetic sleeve of thickness Wrotor is mounted on the
shaft.

Here, N is the number of turns in single coil, I is the current, AG is the area of air gap for one
pole, G is the air gap, and g is the loss due to magnetic flux leakage and fringe effect. The value
of g can vary from 0.6 (for very closely spaced poles with leakage) to 0.9 (for very well-spaced
with low leakage). We have calculated this factor in our experimental analysis.

µ 𝐴𝑔 (2.1)
𝐹= 𝜂 (𝑁𝐼)2 cos⁡(𝛼)
𝐺2

From Eq. (2.1), it can be inferred that designing a magnetic bearing with more number of
turns, higher current, and lesser air gap increases the magnetic force. But with more number of
turns, the volume of the electromagnet increases. Increase in the current is limited by wire
diameter and material saturation. Reducing the air gap reduces the reaction time for the
controller to react and increases the complexity of controller. Therefore, there is a need to
follow a systematic procedure [6] for designing the AMB.

Hsiao et al. [11] optimized the load carrying capacity and force slew rate by constraining the
bearing geometry. Hsiao et al. [11] did not consider the losses, such as copper loss and iron
loss, which constraint the performance of magnetic bearing. Further, Hsiao et al. [11] put a
constraint on the minimum distance between the tips of the poles, which does not ensure the
complete separation between the tips of winding. In the present study, both the aspects
(separation between winding, copper, and iron losses) have been accounted for maximization
of the magnetic force.

Shelke and Chalam [3] optimized the weight of the electromagnet and analyzed the losses
(copper and iron loss) considering different number of poles and concluded that eight pole radial
magnetic bearing reduces the total losses. As per Shelke and Chalam [3], on reducing outer
diameter of magnetic bearing copper loss increases. This may be due to higher current to be
provided in shorter stator radius compared larger radius stator to achieve same magnetic force.
On the other hand, increase in outer diameter of the stator increases the core losses because the
core loss proportionally increases with the increase in volume. Based on these considerations
Shelke and Chalam [3] suggested a wire of cross-sectional area of 1.48 cm2 (1.38cm diameter)
for 30A. Bakay et al. [8] explored the effect of angular velocity of rotor on copper and iron
losses of AMB. As per Yanhua and Lie [6], the iron loss is proportional to the thickness of
lamination and it is a higher-order function of flux density (B). For eddy currents, it is B2, and
for hysteresis it is B1.6. It is noteworthy that various researchers [8–10] analyzed the losses in
10
electromagnet but did not consider losses as a constraint. In the present study, the losses as
constraints have been accounted.

After reviewing the literature, it appears that researchers have either considered geometrical
constraints or analyzed the losses occurring in the electromagnet. Designing an efficient AMB
requires considerations of all the geometrical constraints and losses. In the present work,
constraint due to losses have been included by equating the heat generation and heat liberated
and an appropriate winding interference constraint have been considered. In the present
research, an effort has been made to maximize load carrying capacity using numerical method,
for the given outer diameter of stator, shaft diameter, axial length, and the maximum
permissible current.

Active magnetic bearings are a typical mechatronic product. They are composed of mechanical
components combined with electronic elements such as sensors, power amplifiers and
controllers which may be in the form of a microprocessor.

Figure 2.1 Typical example of a control loop implemented in AMB

Shelke and Chalam [3] optimized the weight of the electromagnet and analyzed the losses
(copper and iron loss) considering different number of poles and concluded that eight pole
radial magnetic bearing reduces the total losses. As per Shelke and Chalam, on reducing outer
diameter of magnetic bearing copper loss increases. This may be due to higher current to be
provided in shorter stator radius compared larger radius stator to achieve same magnetic force.
On the other hand, increase in outer diameter of the stator increases the core losses because the
core loss proportionally increases with the increase in volume. Based on these considerations,
Shelke and Chalam suggested a wire of cross-sectional area of 1.48 cm2 (1.38cm diameter) for
30A.

11
 Figure 2.2 Typical schematic representation of Active magnetic bearing elements

microcontroller with analog circuitry to condition the data transmitted over the interfaces.
Currently, most contact 14 less card solution involved two silicon devices: one like those used
in count cards, operator is microcontroller or memory device, as well as another device to
provide radio frequency interface.

12
 Chapter 3

Bearing Forces

3.1 Governing Laws of Active Magnetic Bearing

It is essential to study the laws that govern the working of the active magnetic bearings in
order to create a mathematical model of a magnetic bearing. The laws are as follows:

• Amperes’ Circuital Law-Consider a long thin wire carrying a steady current .

Suppose that the wire is orientated such that the current flows along the -axis.
Consider some closed loop in the x-y plane which circles the wire in an anti-
clockwise direction, looking down the -axis. Suppose that is a short straight-line
element of this loop. Let us form the dot product of this element with the local magnetic
field B. Thus,

(3.1)

where is the angle subtended between the direction of the line element and the
direction of the local magnetic field. We can calculate a for every line element

13
which makes up the loop . If we sum all of the values thus obtained, and take
the limit as the number of elements goes to infinity, we obtain the line integral

(3.2)

What is the value of this integral? In general, this is a difficult question to answer.
However, let us consider a special case. Suppose that is a circle of radius centred
on the wire. In this case, the magnetic field-strength is the same at all points on the loop.
In fact,

(3.3)

Moreover, the field is everywhere parallel to the line elements which make up the loop.
Thus,

(3.4)

or
(3.5)

In other words, the line integral of the magnetic field around some circular loop ,
centered on a current carrying wire, and in the plane perpendicular to the wire, is equal
to µ𝑜 times the current flowing in the wire. Note that this answer is independent of the
radius of the loop: i.e., the same result is obtained by taking the line integral
around any circular loop centered on the wire. Thus, Ampère's circuital law can be
written:
The line integral of the magnetic field around some closed loop is equal to the⁡µ𝑜 times
the algebraic sum of the currents which pass through the loop.

In forming the algebraic sum of the currents passing through the loop, those currents
which the loop circles in an anti-clockwise direction (looking against the direction of
the current) count as positive currents, whereas those which the loop circles in a
clockwise direction (looking against the direction of the current) count as negative
currents.

14
 • Gauss Law- Gauss’s law for magnetism states that no magnetic monopoles exist and
that the total flux through a closed surface must be zero. The Gauss’s law for magnetic
fields in integral form is given by:

∮B⋅da=0 (3.6)

where:

B is the magnetic flux

The equation states that there is no net magnetic flux (which can be thought of as
the number of magnetic field lines through an area) that passes through an
arbitrary closed surface S. This means the number of magnetic field lines that
enter and exit through this closed surface S is the same.

3.2 Behavior of Magnetic Force

As can be seen in Figure. 3.1 the magnetic force fm behaves in an opposite way to the
mechanical spring force fs: Whereas the latter decreases and even changes its sign with the
distance x, thus producing a repelling effect which stabilizes the rotor motion around the
equilibrium point (x0, s0), the magnetic force increases dramatically with decreasing air gap s
(constant bias current io assumed), which results in an unstable behavior. This basic instability
property is well known to everybody who, maybe as a child, has experienced playing with
permanent magnets or electromagnets. The mechanical stiffness of the suspension is equal to
the negative derivative of the suspension force with respect to displacement
k = −df /dx. (3.7)
Mathematically, the sign of the mechanical stiffness at the operating point (x0, i0, mg)
determines the stability of this equilibrium position. For an open loop magnetic bearing, this
mechanical stiffness is negative. Figure 2.2 also illustrates the quality of the dependency of the
magnet force fm on the air gap s and the coil current i. Basically, the magnet force is
proportional to the inverse of the square of the air gap on the one hand and proportional to the
square of the coil current on the other hand. For small air gaps or large coil currents respectively,
the magnetic flux in the iron path becomes saturated, which, in addition to the basic
characteristics, constitutes a further nonlinearity of the magnet force. Finally, the displacement
x will, of course, be limited geometrically to the size of the air gap.
15
Despite these strong nonlinearities a magnetic bearing system can usually be well controlled
by a linear control scheme. For this, the force/displacement and the force/current dependencies
of the magnet force fm have to be linearized at the operating point (xo, i0, mg) which, as stated
before, denotes the desired equilibrium position, i.e. fm (xo, i0) = mg. Figure 2.3 describes this
situation. In order to eliminate all operating point quantities from the resulting equations it
makes sense to introduce new variables for force f, current i and
displacement x as follows:
f = fm − mg (3.8)

i = i − io (3.9)

x = x – xo (3.10)

This yields the following linearized force/current and force/displacement relationship (2.2) at
the operating point (equilibrium position):

f(x, i) = −ksx + kii (3.11)

3.2.1 Comparison to standard Mechanical Spring

As discussed in the earlier section, many research papers have compared the bearing forces to
the forces provide by mechanical spring and have stated that they behave exactly opposite to
that of the spring. As seen from the figure the stiffness in the mechanical spring is positive,
hence the slope of the plot between deflection and restoring force is negative. But for the
magnetic bearing the stiffness is negative and hence the plot between deflection and restoring
force have positive slope.

16
 Figure 3.1 Comparison of the supporting forces to mechanical Spring

3.3 Mathematical Model of an Electromagnetic actuator

Figure 3.2 One pole model of Magnetic bearing

The conventional linearized model of the electromagnet set is it is a static gain from applied
current and rotor motion to a net force applied to the rotor:

17
 Figure 3.3 Equivalent Magnetic Circuit

𝑁 2 µ𝑜 𝑆 (3.12)
𝐿𝑜 =
2𝑔

𝐿𝑜 (3.13)
𝐹= cos(𝜃) 𝑖 2
2𝑔

Where

Lo is the inductance of the coils in series

g is the radial air gap

s is the area of the pole

µ𝑜 is the relative permebility of air

Figure 3.3 shows an “electrical” equivalent circuit for the magnetic circuit of the electromagnet.
In terms of MMF (voltage), flux (current) and reluctance (resistance), a constant (dc) magnetic
circuit can be treated in the same way as an electric circuit. The main difference is that magnetic
reluctance is an energy storage component rather than a loss component. The “dc voltage”
source Ni represents the MMF generated by the winding current. Rc and RI are the magnetic
reluctances in the C- and I-cores, respectively, and Rg represents the magnetic reluctance at the
airgap.

18
3.4 Magnetic Materials

Magnetic fields can be generated by moving charges (current), alternating electric fields, and
permanent magnets (molecular circular currents and electron spin). A rotation-symmetrical
magnetic field is generated around a straight conductor with a constant current i. The magnetic
field H is inversely proportional to the distance r from the conductor, and its direction is
tangential to concentric circles around the conductor (Figure. 3.3).

Figure. 3.3. Magnetic fields around conductors

Here, the magnetic field intensity is determined by the current density independently of the
medium. A contour integral has either a fixed value or else it vanishes, depending on whether
the path leads around the conductor or not.
If the integration path encompasses several current loops, as is the case with the air coil in Fig.
3.3b, then the integral of the current density J through the surface enclosed by the integration
path yields Ien, the enclosed current.
The formulation (3.5) is called the Ampere’s circuital law [11]. The magnetic field H and the
magnetic induction (flux density) B are linked by the constitutive law
Here, μ0 = 4π × 10−7 Vs/Am stands for the magnetic permeability of a vacuum. The relative
permeability μr depends on the medium upon which the magnetic field acts. For a vacuum, μr
equals 1 and is also approximately unity in air. The SI unit of the magnetic field H is A/m.
Electromagnetic Inductance
Electromagnetic inductance is in fact the inversion of the dynamic effect of magnetic fields, as
referred to at the beginning of this chapter. When a conductor is moved in a magnetic field, all
of its charged particles are moved, too. According to equation (3.1), a Lorentz force acts on
these particles, and they move perpendicularly to the field and the direction of the conductor.
The motion of the charged particles in the conductor corresponds to the electric current. A
current is always caused by an electric field. The electric field generating the induction current

19
is produced by a change in the magnetic flux which passes through the surface surrounded by
a conductor.

3.4.1 Properties of Ferromagnetic Material

When a magnetic field with a density H acts on a material, the magnetic flux density B
generated will be either higher or lower than the flux density μ0H generated in the vacuum,
depending on material properties. The part of B originating from the material itself is called
magnetic polarization M:
in which χm = μr – 1 is called the magnetic susceptibility. This describes the relationship
between the magnetic polarization and the flux density of the vacuum. Materials with χm < 0
(μr < 1) are called diamagnetic. They reduce flux density. Materials where χm > 0 (μr > 1) are
called paramagnetic. In some paramagnetic materials, coupling of the resulting atomic
magnetic moments can occur. When they are parallel, the material is called ferromagnetic. In
this kind of material μr is generally _ 1 and depends both on the size of the magnetic field and
the magnetic “history” of the material. In general, the parallel arrangement of the atomic
magnetic dipoles applies only to a limited space, referred to as Weiss’ domains. The transition
zones between these domains, where the atomic moments change from one privileged direction
to another, are called Bloch walls. The behavior of magnetic material is usually visualized in a
B-H diagram (Figure. 3.5). Thus, when an unmagnetized ferromagnetic sample is placed within
a homogeneous magnetic field with a value H (scalar) of increasing intensity, the magnitude of
flux density B increases rapidly along the new curve in the first quadrant, due to Bloch wall
shifts. Meanwhile, the domains having their privileged direction parallel to the field direction
expand – at the expense of the others.

Fig. 3.5. B-H diagram, hysteresis loop

If H keeps increasing, flux density increases, but slowly. Now, the so-called turning processes
take place, where the dipoles of the Weiss’ domains that are left after the wall shifts change
from their privileged direction in the field direction. Once all magnetic dipoles are parallel to
20
the outer magnetic field, saturation is achieved, and B only increases with slope μ0. When the
outer field is reduced to H = 0 flux density does not run reversibly along the original curve, but
irreversibly along a hysteresis loop. Before reaching value H = 0, only part of the turning
processes is reversed. The remaining flux density is referred to as remanence Br. If H is
increased in the opposite direction, further turning processes will take place first. Then,
remagnetization through wall shifts occurs, and B drops. The field intensity necessary to attain
B = 0 is called coercive field intensity Hc. With increasing intensity of the counter field, the
sample is magnetized until saturation in the opposite direction is achieved. By resetting the
field to zero and by increasing it subsequently in the original direction, saturation is achieved
again, and the hysteresis loop has thus been run through once.

3.4.2 Electrical Steel – (CRNO)

For the manufacture of our bearing we have used electrical steel- Cold Rolled Non – Grain
oriented. This is most commonly available and widely used in the field of electromagnetic
applications. The manufacturing of our model is discussed in section 6.3, here we only discuss
the material in general.
It is available in thin sheets of thickness 0.5 mm, these sheets are laminated in order to decrease
the eddy current loss also special steel tailored to produce specific magnetic properties:
small hysteresis area resulting in low power loss per cycle, low core loss, and
high permeability. Through our literature survey[1-5], generally the magnetic flux saturation is
taken as 1.6 T.

21
 Chapter 4

Rotor-dynamics

4.1 Introduction

As we move on to the stability of the rotor system, we have to consider and analyze the rotor
dynamics of the system. Rotor dynamics will basically include the instability occurring due to
eccentricity change of the center of gravity of the rotor, critical speeds of the shaft, possible
operating speeds, types of whirls and their relationship with the AMB.

Before going for further analysis of the rotor, we must initially make some assumptions for
feasibility of the formulations. The following assumptions has been made:

• The cross section of the rotor is uniform throughout the length.

• The rotor is only under dynamic condition.
• The rotor is a solid 3d geometry with smaller length, since we are considering
conditions under only one bearing i.e. width of the sleeve of the rotor will be equal to
the width of the bearing.
• Since the rotor has a smaller length, we will mainly focus on the radial instability i.e.
we will only consider the plane of cross section.

Now we will consider a rotor of a small finite length with some cross-sectional area having
mass ‘m’. The rotor is operating at some speed ‘ω’. At static conditions the rotor has CG at the
center of the geometry. But at dynamic condition, due to centrifugal force created, the CG of
22
the rotor gets shifted by some distance ‘e’ from the center of the rotor. Also, while rotating, the
rotor system gets translated by distance ‘u’ from the geometrical center of the system.

4.2 Stability of the Rotor

• Approach 1: By considering stiffness of the bearing system.

Here, we consider kx and ky as the stiffness of the bearing system in X and Y direction
respectively.

Figure 4.1 Free Body Diagram of the rotor

In X-direction,

𝑚ⅇ𝜔2 cos 𝜔𝑡 − 2𝑘𝜘 𝑥 = 𝑚𝑥̈ (4.1)

In Y-direction,

𝑚ⅇ𝜔2 sin 𝜔𝑡 − 2𝑘𝑦 𝑦 = 𝑚𝑦̈ (4.2)

Now, let,

𝑥 = 𝑋𝑐𝑜𝑠𝜔𝑡⁡𝑎𝑛𝑑⁡𝑦 = 𝑌𝑠𝑖𝑛𝜔𝑡 (4.3)

23
From all above equations we can formulate,

𝑚ⅇ𝜔2 (4.4)
𝑥= 𝑐𝑜𝑠𝜔𝑡
2𝑘𝑥 − 𝑚𝜔 2

And

𝑚ⅇ𝜔2 (4.5)
𝑦= 𝑠𝑖𝑛𝜔𝑡
2𝑘𝑦 − 𝑚𝜔 2

Also, from above equations, we can introduce,

2𝑘𝑥 2𝑘𝑦 (4.6)

𝜔1 = √ ⁡𝑎𝑛𝑑⁡𝜔2 = √
𝑚 𝑚

From equation (4.1), we can also say that

𝑠𝑖𝑛2 𝜔𝑡 + 𝑐𝑜𝑠 2 𝜔𝑡 = 1 (4.7)

𝑥2 𝑦2 (4.8)
+ =1
𝑋2 𝑌2

From this equation, we can see that it represents an ellipse. We can say that the rotor whirls in
an elliptical profile.

• Approach 2: By considering stiffness and damping of the rotor body.

Here, we only consider the stiffness and damping coefficient which opposes the motion. The
illustration of the considered system is given below. Here, xg and yg represents the translation
of the eccentricity from the center of the system and x c and yc represents the distance of the
C.G. of the rotor from center of the system.

For this system, we can write the general equation as,

24
 𝑚𝑥𝑔̈ = −𝑘𝑥𝑐 − 𝑐𝑥𝑐̇ (4.9)

𝑚𝑦̈𝑔 = −𝑘𝑦𝑐 − 𝑐𝑦𝑐̇ (4.10)

By considering the centrifugal unbalance,

𝑚ⅇ𝜔2 𝑐𝑜𝑠𝜔𝑡 = 𝑚𝑥̈ + 𝑐𝑥̇ + 𝑘𝑥 (4.11)

𝑚ⅇ𝜔2 𝑠𝑖𝑛𝜔𝑡 = 𝑚𝑦̈ + 𝑐𝑦̇ + 𝑘𝑦 (4.12)

(4.13)

By understanding above equations and the principles of unbalance,

𝑢 𝑟2
=
𝑒 √(1−𝑟 2 )2 +(2𝜉𝑟)2

Where,

r = frequency ratio

ζ = damping ratio

Conclusions:

• We have formulated the equations for rotor dynamics.

• We have got the critical speeds for the shaft.
• We have got steady state response for the rotor system.
• We have also got the profile of whirling of the rotor.

25
4.3 Balance Quality Grade

It is neither practical nor possible to remove all unbalance of a rotor. We must establish
balancing tolerances that are economically feasible and still low enough to provide satisfactory
operation. In establishing balancing tolerances, the type of equipment, operation speed, and the
weight of the rotor must be considered. The International Standards Organization has proposed
certain recommendations at to “The Balance Quality of Rigid Rotor Bodies”, their Specification
ISO DR1940. The following table gives the balancing grade for various types of rotors. To
calculate the balancing tolerances, the following formula can be used.

9549 × 𝑊ⅇ𝑖𝑔ℎ𝑡⁡𝑜𝑓⁡𝑡ℎⅇ⁡𝑟𝑜𝑡𝑜𝑟(𝑘𝑔) × 𝑄𝑢𝑎𝑙𝑖𝑡𝑦⁡𝐺𝑟𝑎𝑑ⅇ

𝑇𝑜𝑙ⅇ𝑟ⅇ𝑛𝑐ⅇ =
𝑆𝑝ⅇⅇ𝑑(𝑟𝑝𝑚)

In computing the balance tolerance, the maximum RPM that the rotor is to operate as is used
(not the balancing speed). For single-plane correction, the balancing tolerance as calculated is
used. Where two-plane balancing is required, it is normal to use one half of the balancing
tolerance in each plane.

Table 4.1 Balance Grade and its applications

Balance Grade G EW 1] 2] MM/S Rotor types

40 40 -Car wheels, wheel rims,

wheel sets, drive shafts.
- Crankshaft-drives of
elastically mounted fast
four-cycle engines (gasoline
or diesel) with six or more
cylinders4).
- Crankshaft-drives for
engines of cars, trucks and
locomotives.

16 16 -Drive Shafts (propeller

shafts, cardan shafts with
special requirement.

26
 - Parts of crushing
machinery.
- Parts of agricultural
machinery.
- Individual components of
engines (gasoline or diesel)
for cars, trucks and
locomotives.
- Crankshaft-drives of
engines with six or more
cylinders under special
requirements.

6.3 6.3 -Parts or process plant

machines.
- Marine main turbine gears
(merchant service).
- Centrifugal drums.

2.5 2.5 -Gas and steam turbines,

including marine turbines
(merchant service).
- Rigid turbo-generator
rotors.
- Rotors.
- Turbo-compressors.
- Machine-tool drives.

1 1 -Tape recorder and

phonographs (gramophone)
drives.
- Grinding-machine drives.
- Small electrical armatures
with special requirements
27
 0.4 0.4 -Spindles, disks and
armatures of precision
grinders.
- Gyroscopes.

In the simple case of an inboard rotor with small axial length, for which the couple unbalance
may be ignored, its unbalance state can then be described as a single vectoral quantity, the
unbalance U. To obtain a satisfactory running of the rotor, the magnitude of this unbalance (the
residual unbalance Ures) should not be higher than a permissible value Uper, i.e. Ures <= Uper .In
general, for rotors of the same type, the permissible residual unbalance Uper is proportional to
the rotor mass m: Uper ∼ m. If the value of the permissible residual unbalance is related to the
rotor mass, the result is the permissible residual specific unbalance eper, as given by the
following equation: eper = Uper/m. For rotors of the same type, experience shows that, in general,
the permissible residual specific unbalance value eper varies inversely with the service speed n
of the rotor: eper ∼ 1/n. Differently expressed, this relationship is given by the following
equation, where Ω is the angular velocity of the rotor at maximum service speed: eper⋅Ω =
constant.

28
 Chapter 5

Electronic Components and Control

5.1 Electronic Components

5.1.1 Controller
Manual Digital control in this project is performed through the micro-controller platform
Arduino.

Figure 5.1 Arduino Uno Microcontroller

The detail use and programming of the board is discussed in chapter 7.

29
5.1.2 Power Amplifier
The power amplifiers convert the control signals to control currents. Apart from the bearing
magnets the power amplifiers contribute most to the losses occurring in a magnetic bearing
system. For economic and technical reasons these losses must be kept as low as possible. In
industrial applications, switching amplifiers are used almost exclusively because their losses
are considerably lower than those of analog amplifiers. Admittedly, the switching may cause
electromagnetic disturbances. Because of their simple structure, analog amplifiers are usually
used for sensitive applications where switching disturbances would be a problem or for
applications requiring only very low power.

5.1.3 Displacement Measurement

When selecting the displacement sensors, depending on the application of the magnetic bearing,
measuring range, linearity, sensitivity, resolution, and frequency range are to be taken into
account as well as:
1. Temperature range, temperature drift of the zero point and sensitivity.
2. Noise immunity against other sensors, magnetic alternating fields of the electromagnets,
electromagnetic disturbances from switched amplifiers.
3. Environmental factors such as dust, aggressive media, vacuum, or radiation.
4. Mechanical factors such as shock and vibration.
5. Electrical factors such as grounding issues associated with capacitive sensors.

5.1.3 Capacitive Displacement Sensor

The capacity of a plate capacitor varies with its clearance. Using the capacitive measuring
method, the sensor and the opposing object to be measured form one electrode of a plate
capacitor each (Figure. 3.28). Within the measuring system, an alternating current with a
constant frequency runs through the sensor. The voltage amplitude at the sensor is proportional
to the clearance between the sensor electrode and the object to be measured, and it is
demodulated and amplified by a special circuit. Commercially available capacitive
displacement measuring systems are expensive, but they typically have extraordinary resolution
(for instance 0.02 μm at a measuring range of 0.5 mm). The bandwidth of the output signal
ranges between approximately 5 kHz and 100 kHz. The electrostatic charging of the contactless
rotor may cause interferences too. The sensors are sensitive to dirt which modifies the dielectric
constant in the air gap.
30
 Figure 5.2 Schematic of eddy current and capacitive sensors

5.2 Control Techniques

Another key issue is to establish the appropriate control strategies for a magnetic bearing-rotor
system, which seriously affects the overall performance of a bearing system.

5.2.1 PID
The proportional plus integral plus differential (PID) control scheme, which is a traditional
control scheme, has been used widely in RMB systems due to its ease of use. However, the
control performance of the RMBs system when using the conventional PID control scheme is
not very satisfactory due to the unmeasured parameters variations and unavoidable external
disturbances. Therefore, with the development of modern control theories, many types of
advanced control methods have been recently proposed for RMBs systems, such as fuzzy
control, sliding mode control, model predictive control, and fault tolerant control. These
advanced control methods not only enrich the control theory of RMBs systems but also improve
their performance in different aspects. A comparison of the performances of different control
strategies is shown in Table 5.1.

5.2.2 Fuzzy Control

Fuzzy logic is widely used in machine control. The term ‘‘fuzzy” refers to the fact that the logic
involved can address concepts that cannot be expressed as the ‘‘true” or ‘‘false” but rather as
‘‘partially true”. Due to the nonlinear properties and knowledge-based rules, fuzzy control (FC)
has advantages in applications to magnetic suspension systems, which itself is a highly

31
nonlinear electromechanical plant. FC is a nonlinear control strategy based on input-output
membership functions and rule bases. It describes a control in a qualitative and intuitive way
that emulates the heuristic rule-of thumb strategies of experts. Fuzzy logic can be applied to
diagnosis of magnetic bearing systems with knowledge-based expert systems. FC is employed
to improve the performance of RMB or to overcome the position dependent non-linearity of
magnetic bearings. Fault detection and supervision can apply fuzzy logic control concepts to
abnormal operating conditions.

5.2.3 Sliding Mode Control

In control systems, sliding mode control (SMC) is a nonlinear control method that alters the
dynamics of a nonlinear system by application of a discontinuous control signal that forces the
system to ‘‘slide” along a cross-section of the system’s normal behavior. The state-feedback
control law is not a continuous length of time. Instead, it can switch from one continuous
structure to another based on the current position in the state space. Hence, SMC is a variable
structure control method, which is usually used in the field of magnetic bearings due to its
associated robustness. Although SMC is fully self-adaptive against external disturbances and
parameter variations, the control precision and stability of the system will be affected by the
chattering. Therefore, the filtering method or elimination of disturbances and uncertainties can
solve the problem.

5.2.4 Model Predictive Control

The capability of controlling the coils current with a wide bandwidth is a crucial requirement,
as it directly impacts the dynamic performances of the RMB. For such systems, Model
Predictive Control (MPC) represents an attractive solution, due to its inherently fast dynamic
response, lack of modulation requirement, easy inclusion of nonlinearities and constraints of
the system, possibility of incorporating nested control loops in only one loop and the flexibility
to include other system requirements in the controller. MPC considers a model of the system
in order to predict its future behavior over a specific time horizon. Furthermore, due to the
complex driving conditions during the transitional state, the traditional model predictive control
algorithm with constant weight matrix cannot meet the requirement of improvement in the
RMB system. Therefore, a real-time weight tuning strategy can solve time-varying multi-
objective control problems, where the weight of each objective can be adjusted with respect to
different operating conditions.

32
5.2.5 Fault Tolerant Control
The rotor displacements should be accurately measured by sensors as the feedback signals for
the RMB controller. For the issue of safety, a position sensor fault may cause erroneous
feedback signals and system instability. An existing fault tolerant control solution is to
incorporate sensor redundancy, but the collocation of redundant sensors is expensive and
sometimes difficult due to space limitations. However, the self-sensing technology of the RMB,
in which rotor displacements are directly estimated from the coil voltages and currents, provides
a new and effective method for achieving the sensor fault tolerant control of the RMB. Because
the sensor-fault detection is only determined by the displacement output of the sensor, this fault
tolerant control method will not introduce extra system faults when the parameter estimator
itself fails. This causes the system to be more reliable than the method of using a simple
comparator for fault detection.

33
 Table 5.1 Comparison of Different control techniques
Control Advantages Disadvantages Methods to improve peformance
Techniques
PID Simple to be realized. Not suitable for unknown or Optimize the parameters of PID
unfixed model controller to adapt the changing of
model parameters.
FC Suitable for unknown or Control precision of the Unite other intelligent
unfixed model. system and the dynamic optimization algorithm to
quality of variation may be improve control performance.
reducing by simple fuzzy
information processing.
SMC Fully self-adaptive against Control precision and Filtering method or Elimination
external disturbances and stability of system will be of disturbance and uncertainty.
parameter variations. affected by the chattering,
caused by sliding mode
switch control.
MPC Strong robustness and can The model predictive Need continuous creation in
effectively overcome the control problem for predictive control strategy.
process of uncertainty, nonlinear and time-varying
nonlinear and shunt system has not been solved
resistance, and can process yet.
controlled variable and control
variables in various
constraints.
FTC Strong robustness and can Fault detection and Reduce failure source and
effectively overcome the identification of nonlinear dig into FTC algorithm in
process of uncertainty, system is difficulty and key. more depth.
nonlinear and shunt
resistance, and can process
controlled variable and control
variables in various
constraints.

34
 Chapter 6

Design of Test Rig

6.1 Geometric Parameters for design:

Geometric constraints ensure the physical non-intersection of the pole tips and sufficient gap
between the pole tips so that the wire can be wound onto the pole. The winding space constraint
limits the number of coils of any particular diameter. The excessive current in the coil is limited
by the pole magnetic saturation limits. The excessive current causes excessive losses. The
geometrical constraints include: (1) wire diameter due to current, (2) area of pole, (3) maximum
diameter of stator, (4) maximum diameter of rotor, (5) winding space, and (6) interference of
winding. The geometrical notations of the variables have been shown in Figure. 6.1.

Figure 6.1 Winding Constraints

35
1. Wire Diameter Constraint Due to Current:
Maximum current in the wire is limited by the area of the current carrying wire and
its current density. Where n is the maximum current density of copper wire which
depends on the cooling mechanism, shape of winding and insulation type, etc. In the
present study, the current density is limited to 5 A/mm2

𝐼 (6.1)
𝐽=
𝐴

4𝐼𝑚𝑎𝑥 (6.2)
𝑑𝑤 ≥ √
𝜋𝐽

The wire diameter is also constrainted by the winding space, this also depends on the
stacking effciency.

𝑊 = 𝑛⁡𝑑𝑤 (6.3)

𝑁 (6.4)
𝑛𝑜. 𝑜𝑓⁡𝑙𝑎𝑦ⅇ𝑟𝑠 =
𝑛
𝑁 (6.5)
𝐿= ⁡𝑥𝑑𝑤
𝑛

𝑎𝑟ⅇ𝑎⁡𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙ⅇ = 𝐿𝑤 ⁡ × 𝑊 (6.6)

𝜋 2 (6.7)
𝑎𝑟ⅇ𝑎⁡𝑟ⅇ𝑞. 𝑏𝑦⁡𝑤𝑖𝑛𝑑𝑖𝑛𝑔 = ⁡𝑑𝑤
4

𝜋 2 (6.8)
⁡𝑑𝑤
𝑢𝑡𝑖𝑙𝑖𝑠𝑎𝑡𝑖𝑜𝑛⁡ⅇ𝑓𝑓𝑖𝑐𝑖ⅇ𝑛𝑐𝑦 = ⁡ 4 = 0.785
𝐿𝑤 ⁡ × 𝑊

(6.9)
𝟒 × 𝑳𝒘 × 𝑾 × 𝟎. 𝟕𝟖𝟓
𝒅𝒘 ≤ √
𝝅

36
2. Area of the Pole:
Ferromagnetic materials are made of randomly distributed microscopic regions called
as magnetic domains (small magnets). The random distribution of the domains results
in zero net magnetic field. In the presence of externally applied magnetic field (H),
the magnetic domains align themselves and increase in the magnetic flux density (B)
of the ferromagnetic material. The maximum value of flux density (B) is limited by
the onset of material saturation (B¼Bsat). Saturation occurs when all of the domains
have been aligned. Once the effective saturation limit has been reached, further
increase in applied field (i.e., amp-turns) will result in minimal gain of force output.
The flux generated in the air gap of the magnetic bearing due to the current (I) and
number of turns (N) must be less than the saturation limit of the material hence the
constraint due to saturation is given by

𝐵𝑠𝑎𝑡 𝜇𝑜 (6.10)
𝑁𝐼 ≤
𝐺

𝜋𝐷𝜃𝑡 (6.11)
𝑆 = 2[⁡𝑙 × ( )⁡]
360

3. Pole Width:
Shaft is made of paramagnetic material to allow minimum field (minimize loss) to
pass through. A ferromagnetic sleeve is used to respond the magnetic field generated
by electromagnet. With this arrangement, there will be very little flux in the shaft
and the position sensors would not be affected by it. The radial thickness of this
ferromagnetic ring (Wrotor) should be equal to the pole width (Wp) so that the shaft
gets saturated only when the poles get saturated; thus, shaft is surrounded by ring
having thickness Wrotor of ferromagnetic material. The air gap (G) is kept very less
and so all the field is assumed to pass through the shaft path.

37
6.2 Load capacity Calculation

We have derived the force Equations for an electromagnet in section 3.3. Now we have to
incorporate those equations for the mathematical model and calculation of Radial active
magnetic bearings. Based on the above-mentioned constraints, we have selected the
geometric constraints as follows:

Table 6.1 Geometric Parameters

Parameter Value
Inner diameter of core 76 mm
Diameter of rotor 75.7 mm
Wire Diameter 0.5192 mm (24 AWG)
No. of turns 115
Pole Width 10 mm
Air Gap 0.35 mm
Area of Pole 724.24 mm2

Using Equations 3.6 and 3.7

𝑁 2 µ𝑜 𝑆 (6.12)
𝐿𝑜 =
2𝑔

𝐿𝑜 (6.13)
𝐹= cos(𝜃) 𝑖 2
2𝑔

𝑁 2 µ𝑜 𝑆 (6.14)
𝐹= cos(𝜃) 𝑖 2
2𝑔

From Equation (6.12),

2302 × µ𝑜 × 7.24 × 10−4

𝐿𝑜 = = 68.75⁡𝑚𝐻
2 × 0.35 × 10−3

From Equation (6.13),

38
 68.75 × 10−3
𝐹= cos(22.5) 𝑖 2
2 × 0.35 × 10−3

Keeping current as a variable, we get

𝐹 = 90.73⁡𝑖 2 (6.15)

As per [2] a force efficiency factor (η) is introduced. Now the modified equation becomes

𝐹 = η⁡90.73⁡𝑖 2 (6.16)

We have used a 24 AWG wire which has an ampacity of 3.5 amperes hence the load capacity
of the bearing is calculated by substituting the value of ampacity of the wire in the force
equation (6.13).

𝐿𝑜𝑎𝑑⁡𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 = 666⁡𝑁 (6.17)

But due to the losses (Joules effect) in wire i.e. the heating of the wire occurs. Hence the
operating capacity is always less than the load capacity.

6.3 Model of Test Rig

Construction of the test rig involves fabrication of parts and its positioning. The positioning
of components plays a pivotal role in the modularity of the test rig. The primary design
consideration of the test rig is to provide rigid mounting for the bearings. The figure 6.3
shows the test rig through the means of CAD model. The test rig contains the two active
magnetic bearings. This involves the one being manufacture with copper windings and the
other manufactures with aluminum windings. This is done to increase the specific load
capacity of the bearings. The details of this dimensionless factor are discussed in the next
chapter. The manual control circuits are discussed in section 6.4. The cad and the fabricated
model is shown in figure 6.3 and 6.4. The motor is treated as a shelf component and hence no
included in the fabricated model and design.

39
 Figure 6.2 CAD model of the Test Rig

Figure 6.3 Fabricated model of the Test Rig

The fabrication of the core involved the manufacture by non-conventional method. The design
of the core we have used is not standard as a practice in Active magnetic bearings. Our core of
the test rig was derived from a 24-pole induction motor core shown in Figure6.4. This was
modified to create an effective 8 pole core. The 8 poles were knocked off using CNC wire
cutting.

(a) (b)

Figure 6.4 Actual core used in the AMB (a) Unmodified core (b) Modified core

40
6.4 Electrical Circuits
The basic model for the current is required to be set up inorder to

Figure 6.5 Equivalent electrical circuit of any two diametrical opposite poles

Calculation of Current and Resistance values:

F2 – F1 = mg; ⸫ F2 = F1 – mg (7.1)

F1 = ηk(i1 )2 (6.19)

F1 = ηk(i1 )2 (6.20)

ηk(i2 )2 = ηk(i1 )2 − 𝑚𝑔 (6.21)

ηk(i2 )2 = ηk(i1 )2 − 𝑚𝑔 (6.22)

ηk(i1 )2 − 𝑚𝑔 (6.23)
⸫⁡(i2 )2 =⁡
ηk(i1 )2

mg (6.24)
⸫⁡i2 = √1 − i
F1 1

⸫⁡i2 + i1 = 𝐼 (6.25)

Now using equation (6.23) and (6.25), the values of the current in the diametrically opposite
poles can be found out.
41
Thus, this chapter concludes the design and fabrication of the test rig. The following chapter
discuss the practical limitations of the design of an active magnetic bearing. The consequent
chapters will consist of the FEA analysis and the experimental analysis of the Test Rig.

42
 Chapter 7

Practical Limitation and Dimension-Less Parameters

7.1 Dimension Less Parameter- Specific Load Capacity

In this project we have studied the dimensionless parameter Specific load capacity coined by
W.K.S Khoo [5]. Specific load capacity is defined as the ratio of Rms value of the load
supported to the mass of the bearings.

𝐹𝑟𝑚𝑠 •
𝑆𝑝ⅇ𝑐𝑖𝑓𝑖𝑐⁡𝐿𝑜𝑎𝑑⁡𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 =
𝑀𝑏𝑒𝑎𝑟𝑖𝑛𝑔

From equation (6.13)

𝑁 2 µ𝑜 𝑆 •
𝐹𝑟𝑚𝑠 = cos(𝜃) 𝑖 2 ⁡⁡
2𝑔

From equation (6.11) of area of pole,

𝜋𝐷𝜃𝑡 •
𝑆 = 2[⁡𝑙 × ( )⁡]
360

43
 𝑆𝑝ⅇ𝑐𝑖𝑓𝑖𝑐⁡𝐿𝑜𝑎𝑑⁡𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 •
𝜋𝐷𝜃
𝑁 µ𝑜 2[⁡𝑙 × ( 360𝑡 )⁡]
2
cos(𝜃) 𝑖 2
2𝑔
=
𝑀𝑏𝑒𝑎𝑟𝑖𝑛𝑔

We have attempted to derive the relation of specific load capacity to geometrical parameters
from the basic force equation. Hence, by selection of a desire load capacity (based on the
application) can directly point out the geometric parameters. Thus, a standardization of
magnetic bearings can be attempted.
The scope of this parameter is discussed in the final chapter in section 10.2.

Methods to Increase the Specific Load Capacity based on geometrical parameters (Equation
(7.4)):
• Increment of Axial Length
• Increment of diameter
• Decrement of air gap

7.1.1 Increment of specific load capacity through the mass reduction.

As discussed in section 6.3, we have constructed two AMB’s. One with copper winding and
another with aluminum windings. The density of Aluminum(2700Kg/m3) is very low compared
to copper(8960Kg/m3). The use of aluminum does not hamper the functionality of the bearings.
It only increases the geometric size, this is because the resistivity of aluminum is higher than
the copper. Hence, higher gauge wire is required support the same current.

𝑀𝑏𝑒𝑎𝑟𝑖𝑛𝑔 = 𝑚𝑐𝑜𝑖𝑙 + 𝑚𝑐𝑜𝑟𝑒 + 𝑚𝑟𝑜𝑡𝑜𝑟

7.2 Practical Limitations to Magnetic bearing

The concept of magnetic levitation is quite revolutionary, it has its advantages, but to make
progress, we must understand the limitation to an active magnetic bearing.

44
7.2.1 Inherent Instability

Magnetic levitation can be divided into electromagnetic and electro dynamical levitation. The
active magnetic bearing uses electromagnetic levitation (explained in section 3.2). The
attractive forces vary inversely to the square of air gap.
Now, consider a hypothetical case i.e. the absence of any control system. The figure 7.1 shows
the initial static condition of the bearing. If an external stimulus in downward direction in
applied on an active magnetic bearing the air gap reduced as depicted in figure 7.1(b).

(a) (b)
Figure 7.1 Schematic representation of magnetic bearing. (a) static condition
(8) Under External Stimulus

The force being inversely proportional to the square of the air gap. The force at the top pole
decreases and correspondingly at the bottom pole increases. Thus, the rotor starts moving
downward, ultimately sticking to the bottom pole. This behavior depicts the inherent instability
of the magnetic bearing.
Hence an active feedback loop is necessary to control levitation. Eric Maslen in his paper [4]
has addressed this issue through the calculation of Slew rate. This term is explained in the
consecutive section.

7.2.2 Slew Rate and Maximum Slew rate

The slew rate limitation causes:

1. Bearing force to change more slowly than expected
2. Introduces Phase lag
45
 3. Sudden loss of damping near the limit value of slew rate.

Based on the inherent instability of the active magnetic bearings the force slew rate for higher
displacements is calculated as follows.

µ𝐴𝑔 (7.5)
𝐹= 𝜂 (𝑁𝐼)2 𝑐𝑜𝑠⁡(𝛼)
𝐺2

𝑘1 𝑘2 (7.6)
𝐹𝑟𝑒𝑠 = 2
+
(𝐺 + 𝑥) (𝐺 − 𝑥)2

𝑑𝐹𝑟𝑒𝑠 𝑑 (7.7)
= − (𝐹2 ⁡–⁡𝐹1 )
𝑑𝑡 𝑑𝑡
𝑑 𝑘1 𝑘2
=⁡− ( 2
+ )
𝑑𝑡 (𝐺 + 𝑥) (𝐺 − 𝑥)2

𝑘1 = 𝜂(µ𝐴𝑔 )(𝐼1 𝑁)2 𝑐𝑜𝑠(𝛼) (7.8)

𝑘2 = 𝜂(𝐴𝑔 µ)(𝐼2 𝑁)2 𝑐𝑜𝑠⁡(𝛼) (7.9)

𝑑𝐹𝑟𝑒𝑠 −2𝑘1 𝑑𝑥 −2𝑘2 −𝑑𝑥 (7.10)

= −( 3
( )+ ( ))
𝑑𝑡 (𝐺 + 𝑥) 𝑑𝑡 (𝐺 − 𝑥)3 𝑑𝑡

𝑑𝐹𝑟𝑒𝑠 2𝑘1 2𝑘2 𝑑𝑥 (7.11)

=( 3
− 3
)
𝑑𝑡 (𝐺 + 𝑥) (𝐺 − 𝑥) 𝑑𝑡

𝑑𝐹𝑟𝑒𝑠 𝑑𝑥 (7.12)
= 𝑘1
𝑑𝑡 𝑑𝑡

𝑑𝐹𝑟𝑒𝑠 𝑑𝑥 𝑑𝛳 (7.13)
= 𝑘1
𝑑𝑡 𝑑𝛳 𝑑𝑡

𝑑𝐹𝑥 𝑑𝑥 (7.14)
= 𝑘1 𝜔
𝑑𝑡 𝑑𝛳

The force slew rate is limited because the magnet coils are inductive and the power supply to
the driving amplifier is at a fixed voltage. Eric Maslen [4] has derived this limitation
considering small displacements. The following is the equation.
𝑑𝐹𝑟𝑒𝑠 (7.15)
< 𝐾𝑖 [𝑉𝑠 − (𝐼𝑝 − 𝑖𝑝 )(𝑅𝑐 + 0.5 + 𝑅𝑓 )]/𝐿𝑐
𝑑𝑡

46
 Chapter 8

Finite Element Simulation

The finite element simulation is conducted to visualize the generation of magnetic flux and
visually understand the effects of change in geometry. The secondary aim was to verify the
non-conventional method of manufacture of magnetic bearings used would perform as
theorized. The finite elemental simulation was conducted in the Electromagnetic simulation
(EMS) suite of the software developed by EM Works Inc. This suite was run on ANSYS
workbench interface. The steps involved in this type of simulation remain similar to
conventional procedure i.e. Preprocessing, solving and postprocessing.

8.1 Preprocessing
The preprocessing step consists of the creation of basic CAD model of the component of the
bearings i.e. Core, coil and the rotor. The EMS suite is so programmed that it requires additional
pre-solving steps i.e. creation of Inner and outer air. This is since the software treats the outer
air as the external surrounding and the inner air from the reluctance path gets completed. In the
absence of the inner air the elemental matrix equations remain void and thus giving errors in
the results. The Figure 8.1 shows all the CAD models generated for the pre-processing step.
The inner air gets assembled in the empty space denoted in white shown in figure 8.1(c).

47
 (a) (b)

©
Figure 8.1 CAD Models (a)Assembled Magnetic bearing with outer-air (b) Inner air
(c)Assembly of the Bearing with coils

8.2 Solving
The solution is obtained through the Nonlinear Magneto-static solver used by the EMS suite.
The initial step of the solving process is the meshing of the body. To decrease the computation
time a medium size mesh was used. The type of solver used is Non-linear Newton Raphson
solver. The mesh Parameter used are summarized in table 8.

Figure 8.2 Generated Mesh

48
 Table 8.1 Mesh Parameters
Mesh Parameter Value
Global Size 8.00037 mm
Tolerance 0.008
Avg. number of elements per diagonal of 38
each solid body

8.3 Post-processing and Results

The results are calculated in two parts:
• Single pole – Reluctance path visualization
These results are generated to analyze the magnetomotive force per pole. The results
are displayed in Figure 8.3 to 8.5. These results also help in the visualization, that the
magnetic bearing works in accordance to Gauss Law.
• Complete Core Magnetization- This is performed to compute the magnetic flux
saturation. The magnetic flux saturation in the major limiting factor in active magnetic
bearings. These results are displayed in figure 8.5 and 8.6. The maximum saturation
flux is 1.57 Tesla.

Figure 8.3 Contour Result of Magnetic Flux for Single Pole Actuation

49
Figure 8.4 Magnetic Field intensity plot showing the opposite polarity of the consecutive
poles

Figure 8.5 Shows the Magnetic flux vector plot

50
Figure 8.6 Contour Result of Magnetic Flux for complete actuation

Figure 8.5 Shows the Magnetic flux vector plot

51
 Chapter 9

Experimental Analysis

The experimental analysis is performed on the bearing with the following aims:
• To find the Force Efficiency Factor (η). (discussed in section 6.2)
• To find the load vs current graph for the magnetic bearing.

Fig 9.1 The experimental set up (also shows the bearing under loaded condition)

52
The experimental schematic is shown in figure 8.1, we have conducted the experiments
independently on the AMB with copper windings and AMB with Aluminum windings. This
methodology of conduction of independent analysis provides us the confirmation that the
behavior of the magnetic bearing is independent of the material of winding.
The theoretical behavior of the AMB is governed by the equation (6.15). The table 8.1
summarizes the theoretical values.

Table 9.1 Theoretical values of Force Generated

Force Current
3.6292 0.2
14. 5168 0.4
32.6628 0.6
58.0672 0.8
90.73 1
130.65 1.2
177.83 1.4
232.68 1.6
293.96 1.8
362.92 2

Keeping the current constant (I) at 1A, the rotor was in contact with the upper pole. We placed
our magnetic bearing apparatus on two tall support platforms. Then, two pans were hanged at
each end of the rotor shaft. The rotor independently weighs 1kg; and for the experiment, load
in each pan was increased equally until the rotor lost contact with the pole, i.e. the stator. This
procedure was iterated for repeatability. The load, thus, obtained was F, which was between
50N to 60N, average value taken as 58.97N. based on equation (6.16) we have calculated the
force efficiency factor. Table 8.2 summarizes the results of the experiment.

𝐹 = η⁡90.73⁡𝑖 2

F 2 (8.1)
η= 𝑖
90.73

From table 8.1 the theorectical value for 1A is taken

53
 58.97 (8.2)
𝜂= = 0.65
90.73

Thus, the force efficiency factor was obtained. As discussed in section 7.3. the force efficiency
factor depends of various factors such as slot leaker and fringing effect.

Table 9.2 Experimental results

Force Current
2 0.2
10.5 0.4
20 0.6
40 0.8
60 1
84.92 1.2
111.5 1.4
150 1.6
195 1.8
220 2

Thus, to analyses the nature of the of the characteristic curve we constructed curve fit
plot comparing the theoretical and experimental results. The figure 8.1 shows this plot.
This plot was generated using MATLAB plotting and curve fitting tools

Figure 9.2 Comparison plot of the results (red line denotes experimental results and blue
line denotes the experimental results)

54
 Chapter 10

Conclusion and Future scope

9.1 Conclusions
The conclusion drawn are based on each chapter of this document.
• The Load vs current characteristic curve is parabolic in nature. The force efficiency
factor acts as a scaling factor to the curve.
• The practical limitations of the magnetic bearing are studied and an attempt to
theoretical analyze the slew rate for larger displacements is made. We can draw a
conclusion that the slew rate is proportional to the velocity in the direction of
motion.
𝑑𝐹𝑟𝑒𝑠 𝑑𝑥 (9.1)
= 𝑘1
𝑑𝑡 𝑑𝑡

𝑑𝐹𝑟𝑒𝑠 (9.2)
= 𝑘1 𝑉
𝑑𝑡

• The specific load capacity can be increase by reducing the mass of the bearings. The
specific load capacity is dependent of geometrical parameters. Directly proportion to
length and diameter of the core and inversely proportional to air gap.

55
 𝜋𝐷𝜃
𝑁 2 µ𝑜 2[⁡𝑙 × ( 360𝑡 )⁡]
cos(𝜃) 𝑖 2
2𝑔
𝑆𝑝ⅇ𝑐𝑖𝑓𝑖𝑐⁡𝐿𝑜𝑎𝑑⁡𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 =
𝑀𝑏𝑒𝑎𝑟𝑖𝑛𝑔
• The force efficiency factor depends on the slot leakage and the fringing effect.
• The finite elemental results have aided in the analysis but computing the values of the
saturation flux and aided in the selection of the material- Iron silicon steel Non – grain
Oriented.
• The current use of active magnetic bearings is limited to certain devices due to many
factors. But largely used in flywheel energy storage and power generations, where the
fluctuations are minimum.
• The Active Magnetic bearings are based on a simple two pole electromagnet which is
connect to an active feedback loop.
• The rotor dynamics are an essential part of the analysis as the rotor is levitated. Through
the basic theoretical analysis, governing equations are put forth. This is required as the
damping present is minimal.
• The electronic components are integral part of the bearing system. The different control
techniques are studied and compared. PID techniques is the most simple and robust.

9.2 Future Scope

Through this project we intent to give directions to the future research in the field of contact
less bearings. The future scope of this project is represented as directions of research.

• Direction 1- Increment of load Capacity while keeping geometric parameters

constant
This is direction can be achieved through metallurgical research i.e. implementing
material which have higher magnetic flux saturation. Hiperco being one such material
which has higher magnetic saturation of 2.36 T.

• Direction 2- Capacity to Handle Fluctuating Loads

The capacity to handle fluctuating loading conditions is governed by the SLEW RATE.
The slew rate as a parameter is discussed in the above chapter. The slew rate is
dependent on the controller feedback frequency. In chapter 5, we have studied the

56
 different method of control that can be implemented. Hence, a further research in
comparing the handling capacity of different types of control techniques is required.

• Direction 3- Standardization of Magnetic Bearings

In order to use a magnetic bearing for any application, a major task is its design. To
revolutionize this concept, it is essential to standardize the component. This can be
achieved through research on parameters like specific load capacity. Hence, while
incorporating a Magnetic bearing in any system a selection design can be performed.

• Direction 4- Setting up a research Lab in collaboration with Industry Giants

In section 1.5, we have discussed our market survey. In India, the potential growth of
this technology is high. Thus, we can say that setting up a research laboratory for such
a test rig will increase the chances of industrial growth of this technology in India. There
is a need for industry giants in India to venture in this technology. This can only be done
through the such kind of laboratories.

57
 References

[1] Schweitzer, G., and Maslen, H., 2009, Magnetic Bearings: Theory, Design, and
Application to Rotating Machinery, Springer-Verlag, Berlin, Germany.
[2] Dr. K.P. Lijesh; “optimization of eight pole radial active magnetic Bearing”, ASME
Journal of tribology; 2010.
[3] Shelke, S., and Chalam, R. V., 2011, “Optimum Energy Loss in Electromagnetic
Bearing,” 3rd International Conference on Electronics Computer Technology (ICECT),
Kanyakumari, India, Apr. 8–10, pp. 374–379.
[4] Eric H. Maslen, P. Hermann; “Practical limits to the performance of Magnetic
Bearings”, Journal of Tribology; 1989
[5] W.K.S Khoo, S.D. Garvey; “The specific load capacity of Radial Flux Magnetic
bearings”, IEEE transactions on magnetics, Volume 43 No.07; 2007
[6] Yanhua Sun, Lie Yu; “Analytical method for Eddy Current Loss in Laminated Rotors
with magnetic Bearings”, IEEE transactions on magnetics, Volume 38 No.02; 2002
[7] M. Necip Sahinkaya, Ahu E. Hartavi; “Variable Bias Current in Magnetic Bearings for
Energy Optimisation”, IEEE transactions on magnetics, Volume 43 No.03; Year
[8] Williams, R. D., Keith, F. J., and Allaire, P. E., 1990, “Digital Control of Active
Magnetic Bearings,” IEEE Trans. Ind. Electron., 37(1), pp. 19–27.
[9] Chang, H., and Chung, S. C., 2002, “Integrated Design of Radial Active Magnetic
Bearing Systems Using Genetic Algorithms,” Mechatronics, 12(1), pp. 19–36.
[10] Hsiao, F. Z., Fan, C. C., Chieng, W. H., and Lee, A. C., 1996, “Optimum
Magnetic Bearing Design Considering Performance Limitations,” JSME Int. J. Ser. C,
39(3), pp. 586–596
[11] Pranab Samanta, Harish Hirani,”Magnetic Bearing Configurations: Theoretical
and Experimental Studies”, IEEE transactions on magnetics, Volume 44 No.02; 2008
[12] Weiyu Zhang, Huangqiu Zhu, ”Radial magnetic bearings: An overview”,
Results in Physics 7 (2017) 3756–3766.

58
 59

View publication stats