introduction to magnetic bearings - qip08-handouts · • magnetic bearings are free of contact and...
TRANSCRIPT
1
Introduction to Magnetic Bearings
Jagu Srinivasa Rao, (Research Scholar)
Department of Mechanical EngineeringIndian Institute of Technology Guwahati
December, 2008
Lecture presented in Quality Improvement Program (QIP’08) at Indian Institute of
Technology Guwahati
Overview of the Presentation
• Introduction
• Design of Active Magnetic Bearings
• Control Engineering of Magnetic Bearings
• Control of Rotor by using Magnetic Bearings
• Conclusions
Introduction
• An active magnetic bearing (AMB) system supports a rotating shaft, without any physical contact by suspending the rotor in the air, with an electrically controlled (or/and permanent magnet) magnetic force.
• It is a mechatronic product which involves different fields of engineering such as Mechanical, Electrical,
Control Systems, and Computer Science etc.
Test Apparatus for rotor control
Eight-Pole Radial Magnetic-Bearing
Radial Magnetic Bearing
Horizontal shaft Vertical shaft
Rotor shaft
Upper AMB
Lower AMB
Rotor Disc
Coil WindingLeft AMB
Thrust Magnetic Bearing
Left AMB
Ro
tor s
haft
Typical Actuator – Controller unit of an AMB
Introduction to Active Magnetic Bearings
2
Working principle of magnetic bearing
Electro magnet
Sensor
Controller
Power Power
AmplifierAmplifier ff
Roto
r
Introduction to Active Magnetic BearingsAdvantages of Magnetic Bearings
• Magnetic Bearings are free of contact and can be utilized in
vacuum techniques, clean and sterile rooms, transportation of aggressive media or pure media
• Highest speeds are possible even till the ultimate strength of
the rotor
• Absence of lubrication seals allows the larger and stiffer
rotor shafts
• Absence of mechanical wear results in lower maintenance costs and longer life of the system
• Adaptable stiffness can be used in vibration isolation, passing critical speeds, robust to external disturbances
Classification of Magnetic Bearings
According to •control action
– Active– Passive– Hybrid
•Forcing action – Repulsive – Attractive
•Sensing action– Sensor sensing– Self sensing
•Load supported
– Axial or Thrust– Radial or Journal– Conical
• Magnetic effect– Electro magnetic– Electro dynamic
• Application– Precision flotors– Linear motors
– Levitated rotors– Bearingless motors– Contactless Geartrains
Applications of Magnetic Bearings
•Turbo molecular pumps
•Blood pumps
•Molecular beam choppers
•Epitaxy centrifuges
•Contact free linear guides
•Variable speed spindles
•Pipeline compressor
•Elastic rotor control
•Test rig for high speed tires
•Magnarails and maglev systems
•Gears, Chains, Conveyors, etc
•Energy Storage Flywheels
•High precision position stages
•Active magnetic dampers
•Smart Aero Engines
•Turbo machines
Fields of Applications of Magnetic Bearings
•Semiconductor Industry
•Bio-medical Engineering
•Vacuum Technology
•Structural Isolation
•Rotor Dynamics
•Maglev Transportation
•Precision Engineering
•Energy Storage
•Aero Space
•Turbo Machines
– Electromagnetic field
– Lorenz force
Electromagnetism
3
Electromagnetism
When a charged particle is
at rest it won’t emit electromagnetic waves rather it is surrounded by
electrostatic field
When the charged particle is in uniform motion (i.e. the
motion with uniform velocity in a direction) the electrostatic field is
associated with magnetostatic field.
3d electrostatic field surrounding a
charged particle
Magnetostatic field
Electromagnetism
When the particle is in accelerated motion then
the magnetic field will be oscillating.
In electromagnetic waves both the electric and magnetic fields are
oscillating and harmonic.
The electric and magnetic fields are generated by electric charges
Charges generate electric fields
Movement of charges generate magnetic fields
The electric and magnetic fields interact only with each other
Changing electric field acts like
a current, generating vortex of
magnetic field
Changing magnetic field
induces (negative) vortex of
electric field
Feed back loop of electromagnetism
The electric and magnetic fields produce forces on
electric charges
Electric force which is
generated by the electric field and is in same direction as
electric field
magnetic force which is
generated by the magnetic field and is perpendicular both
to magnetic field and to
velocity of charge
The electric charges move in space
The electric charges move in
space when they are acted
upon by field forces
The electric and magnetic The electric and magnetic fields are generated by fields are generated by
electric chargeselectric charges
The electric and The electric and magnetic fields magnetic fields
interact only with interact only with each othereach other
The electric and magnetic The electric and magnetic fields produce forces on fields produce forces on
electric chargeselectric charges
The electric The electric charges move in charges move in space when they space when they are acted upon by are acted upon by
field forcesfield forces
Feed back loop of electromagnetism
The four fundamental forces
Strong nuclear force
which holds atomic
nuclei together
Weak nuclear force
which causes
certain forms of
radioactive decay
The four fundamental forces
Electromagnetic force
Which is caused by
electromagnetic fields on
electrically charged
particles
Gravitational force
Which causes the
masses to attract
each other
4
The four fundamental forces
All the other forces are derived from these four fundamental forces
Electro-magnetic force is one of these
four fundamental forces
1 2
3
04c
q qf
rπε= r
Force between two electrically charged particles
Coulomb force (Static)
cf1q
r
2q
Lorenz force (Dynamic)
1 2 1 2
3 2 3
0 04 4l
q q q qf
r c r
γ γ
πε πε
×= + ×
r v rv
If q1=q then
( )q= + ×F E v B
2
2 3
04
q
c r
γ
πε
×=
v rB2
3
04
q
r
γ
πε=
rE
-7 2
02
0
1= 4π×10 N/A
cµ
ε=
Electric and magnetic components of Lorenz force
; = r r
12 28.854 10 C / J-mε −0 = ×
( )2
1
1 /v c
γ =−
Electric flux; Magnetic flux;
Lorenz factor;
Magnetic permeability of vacuum;
Electric permeability of vacuum;
2
2 23
1
10
v
c
×≤ ≤
v B
E
Three conclusions:• Magnetic component of Lorenz force is at least smaller by a factor of 1023!
But we don’t face the effect of electric field in conductors because protons
and electrons are equal in number and generate equal and opposite electric fields
canceling each other
• Protons have no motion with reference to conductor and there won’t be magnetic component from them. Thus the magnetic component observed is
the relativistic effect of electrons only
• When the conductor is moving with reference to another frame both the
protons and electrons will move with the same velocity thus the relativistic
effects due to the velocity of conductor will be cancelled out
Comparison Electric and magnetic components of Lorenz force
Effective Lorenz force in macro calculations
For macro calculations Lorenz force is
reduced to the form
( )q= ×F v BB
v
F
wB
Lorenz force acts perpendicular to both velocity
of charged particle and magnetic flux
Relations between E and B
0
q
ε∇ ⋅ =E
t
∂∇ × = −
∂
BE
0 0t
µ ε∂
∇ × = + ∂
EB J
0∇ ⋅ =B
Gauss’ Law for linear
materials
Gauss’ Law for
magnetism
Faraday’s law of magnetic induction
Ampere’s law and
Maxwell's extension
0
1
S Vqdv
ε⋅ =∫ ∫E ds
0S
⋅ =∫ B ds
L St
∂⋅ = − ⋅
∂∫ ∫E dl B ds
0 0L S t
µ ε∂
⋅ = + ⋅ ∂
∫ ∫E
B dl J ds
These relations are called simplified Maxwell's relations who formulated
the original relations from previous works
5
Design of magnetic
actuator
– Bearing magnet
– Magnetic circuit
– Coil
De
sig
n m
eth
od
olo
gy o
f m
ag
ne
tic
be
ari
ng
sys
tem
s
yes
Specifications
Mechanical design
Magnetic actuator design
Control system design
Simulation
Experimentation
Performance O.K?
Performance O.K?
End
Performance O.K?
yesno
yes
no
no
Magnetic bearing system design
Mechanical design
Magnetic actuator design
Control system design
Modal frequencies
Bearing magnet design
Coil design Sensor design
Controller design
Power amplifier design
Topology
Load estimation
Magnetic circuit design
Admissible coil temperature
Number of turns
Winding scheme
Coil head
Position sensing
Velocity sensing
Current sensing
Flux sensing
Stiffness
Damping
Balancing
Stability
Losses
Self sensing
Areas involved in the design of magnetic bearing systems
Bandwidth
Magneto mechanical systems
According to the known technology till According to the known technology till
now, magnetic bearings can be classified now, magnetic bearings can be classified
for their design according to the purpose for their design according to the purpose
of the levitated object asof the levitated object as
– Precision flotors (precision stages,
isolation bases, isolation springs)
• Levitation force
• Propulsion force
Magneto mechanical systems
A magnetic Precision Stage
Linear motors(Contactless sliders,
maglev trains and
conveyors)• Levitation force
• Propulsion force
Levitation force Propulsion force
Principle of a linear motor
Magneto mechanical systems
6
Levitated rotors(gas turbines, energy storage flywheels, high speed spindles, balancing and vibration control of rotors)
– Radial load
– Thrust load
Magneto mechanical systems
Rotor levitated by Radial and Axial Active Magnetic Bearings
Bearingless motors (canned pumps, compact pumps, blood pumps, spindle drives, semiconductor
process)
– Radial load
– Thrust load
– Torque
Magneto mechanical systems
Bearingless Motor
Contactless Gears and Couplers – Regulated torque
transmission
Magneto mechanical systems Linear systems from rotary systems
Design of a thrust magnetic bearingMacro Geometry of Thrust Magnetic Bearing
Inner wall
Outer wall
Back-wall
Coil
Space for coil
Space for shaft
Figure 1: Parts of Thrust Magnetic Bearing
7
Optimal design
Optimal design is carried out in two steps
• Modeling the magnetic circuit
– Determines the accuracy of achieving the objective
• Optimization of the parameters
– Determines the efficiency of the achieving the objective
Magnetic circuit
aR
lR
gR
Ni
Equivalent electric (dc) circuit representation
Magnetic circuit
Ni
φl
R
aR
gR
gap Levitated object
Actuator
Coil
φ
0
fp fp
r
l l
A AR
µ µ µ==
Magnetic circuit analogy with electric circuit
Electro Motive
Force (EMF) or
Voltage (V)
Magneto Motive
Force (MMF)
Electric circuitMagnetic circuit
Resistance (R)Reluctance (R)
Electric Current (i)Magnetic Flux ( )φ
Ideal magnetic circuit model
( ) Ampere's lawL SH dl J nda⋅ = ⋅∫ ∫
2g g a a s s
H l H l H l ni+ + =
or /B H H Bµ µ= =
al
gl
sl0 02 a s
g g a s
a s
B BB l l l niµ µ
µ µ
+ + =
0if is neglecteda sa s
a s
B Bl lµ
µ µ
+
0
2g
g
niB
l
µ=
H
J
Flux density is used to find the force exertedFlux density is used to find the force exerted
Extension of the ideal modelal
gl
sl0
2a g g i
K B l K niµ=
a 0
i
if K is added for
as core loss factor and K is added
as coil loss factor, then
a sa s
a s
B Bl lµ
µ µ
+
0
2
ig
a g
K niB
K l
µ=
The model reduces toThe model reduces to0B B+ ∆
0B B− ∆
0i i+ ∆
0i i− ∆
0
0
2 gAB BF
µ∆∆ =
Force by using flux density
Differential actuator
0
2( )g
Ni
A l xB
µφ=
−=
2
0
02
g
BF A
µ=
8
Linear Range
max satBα
min satBα
satB
Magnetic force, N
Magnetic f
lux d
ensity,
T
Hysteresis is assumed to be negligible while setting the linear range
Linear range of flux density
0.1005
10.05
0.0010
1600
7.95e5 for air
3.97e4 for Fe
0.026
Magnitude
Wb-
turns
Magnetic flux
linkage
TFlux density
WbMagnetic flux
A-
turns
Magneto
motive force
Vs/AReluctance
Vs/AmPermeability
Vs/AmPermeability
of vacuum
UnitsFormulaSymbolQuantity
R0
fp fp
r
l l
A wlµ µ µ=
µ0 rµ µ
0µ
2
0
1
cε
φ 0
2 2 ( )g
N i w lN i
R g x
µ=
−
B 0
2( )
Ni
A g x
µφ=
−
ni n i×
λ Nφ
Terminology used inmagnetic circuit
74 10π −×
19.84
804.2
804.2
0.0063
16e4
Magnitude
NMagnetic
force for diff
actuator
NMagnetic
force by flux
density
NMagnetic
force by
inductance
HNominal
inductance
H=Wb/AMagnetic
inductance
A/m2Current
density
UnitsFormulaSymbolQuantity
Different quantities used in
magnetic circuit
0L
2
0
0 2
xg
n wlL
l
µ=
=
L( )
2
0
2
g
n wl
i l x
µλ−
=
F2
0
2 g
L i
l
Ji i
A wl=
F2
0
02g
BA
µ
F ( )2 2
02
gA
B Bµ
+ −−
Design vector for optimal design
Known parameters areKnown parameters are
••Gap Gap
••Inner radius of the bearingInner radius of the bearing
••Outer radius of the bearingOuter radius of the bearing
Free parametersFree parameters
••Inner radius of the coil spaceInner radius of the coil space
••Outer radius of the coil spaceOuter radius of the coil space
••Height of the coil spaceHeight of the coil space
••Current density suppliedCurrent density supplied
All the other parameters are dependantAll the other parameters are dependant
70mmMaximum height of bearing120mmMaximum outer radius
of bearing
820mm3Maximum allowable coil
volume
4.0A/mm2Saturation current
density
0.85Packing factor1.2TRemnant flux density of
bias magnets
0.840Flux leakage factor1.00TSaturation flux density
1.072Actuator loss factor±10%Variation in the load
1.394Coil mmf loss factor±5%Variation in the gap
7.5g/cm3Specific gravity of
permanent magnet material
neodymium-iron-baron
2025NOperating load
8.91g/cm3Specific gravity of the
copper
4.00mmOperating air gap
7.77g/cm3Specific gravity of the
stator iron
25.00mmInner radius of the
bearing
ValueParameterValueParameter
Input parameters taken for the design of thrust magnetic bearing Eight pole radial magnetic bearing
Eight Pole AMB
9
Radial magnetic bearing
2
0
2
( )cos
4( ) 2
g i
a g
A K niF
K l
µ α=
The component of force will be at an angle of half of the angle between two poles
α
Three pole radial magnetic bearing
Three Pole AMB
Magnetic Circuit for three pole AMB
Coil design
• Admissible coil temperature is determined by
the choice of insulation type
• Number of turns are chosen such that it
generates maximum admissible magneto
motive force at the maximum current supplied
by the power amplifier
Coil
Winding scheme
Permanent magnetic bearings Permanent magnetic bearings
rB
aH BH
B
cH maxBH
10
MAGNETIC BEARINGS
CONTROL
Introduction
• Control is the process of bringing a system into desired path when it is
going away from it
• Earnshaw(1842) had shown that it is impossible to hover a body in all six
degrees of freedom by using permanent
magnets
• But it is possible to maintain the body in
equilibrium condition by active control
Types of control systems
• Open loop control systems
• The control in which the output of the system has
no effect on input is called open loop control
• Open loop control is used when the input is known
and there are no external disturbances
• An example of open loop control is washing
machine which works on time basis rather than the
cleanliness of clothes
( )G s( )U s ( )Y s
Types of control systems
• Closed loop control systems
• If the control maintains a prescribed output and
reference input relation by comparing them and
uses their difference as controlling quantity, it is
called feedback or closed loop control
• Temperature control of a room or a furnace is an
example of closed loop system
( )G s( )U s ( )Y s
( )H s
x
Classification of controllers
• According to control action controllers are
classified as:
• Two-position or on-off controllers
• Proportional controllers
• Integral controllers
• Proportional-integral controllers
• Proportional-differential controllers
• Proportional-differential-integral controllers
Classification of controllers
• Two-position or on-off controllers
• The output of the controller will be a
maximum or minimum according to the state of
error as below:
• are minimum and maximum values of
output0 1 and y y
0
1
( ) for ( ) 0
for ( ) 0
y t y e t
y e t
= <
= >
( )y t
( )e t
11
Classification of controllers
• Proportional controllers:
• The output of the controller is proportional to
the magnitude of the actuating error signal as
• By Laplace transformation
• is called proportional gain
( )y t
( )e t
( ) ( )py t g e t=
( )
( )p
Y sg
E s=
pg
• Integral controllers:
• In integral control action, the value of the
controller output is changed at a rate
proportional to the actuating error signal
• By Laplace transformation
•
• is called integral gain
( )y t
( )e t
( )( )
i
dy tg e t
dt=
( )
( )
igY s
E s s=
ig
Classification of controllers
0( ) ( )
t
iy t g e t dt= ∫(or)
• Proportional-Integral (PI) controllers:
• Control action is a combination of both
proportional and integral action
• By Laplace transformation
( ) 11
( )p
i
Y sg
E s T s
= +
0
( ) ( ) ( )
tp
p
i
gy t g e t e t dt
T= + ∫
Classification of controllers
proportional-differential (PD) controllers:
The control action is defined by
By Laplace transformation
( )(1 )
( )p d
Y sg T s
E s= +
( )( ) ( )
p p d
de ty t g e t g T
dt= +
Classification of controllers
proportional-Integral-differential (PID)
controllers:
It has the advantages of all three actions. So this is
the most common type of industrial controllers
Mathematical form of PID action is
By Laplace transformation
( ) 11
( )p d
i
Y sg T s
E s T s
= + +
0
( )( ) ( ) ( )
tp
p p d
i
g de ty t g e t e t dt g T
T dt= + +∫
Classification of controllers Control Design
An over all system
G(s)U(s) Y(s)
Transfer-function representation of a system
u(t) y(t)( )
( )
x t
x t
System Input Output
State-space representation of a system
( ) ( ) ( )Y s G s U s=
( ) ( ) ( )t A t B t= +y x u
12
Control Design
An over all system
SystemInput Output
Studying the behaviour of a system
KnownKnown unknown
UnknownKnown known
Studying the characteristics of a system
UnknownUnknown known
Designing of a control system of required behaviour
Methods of design and
analysis of controllers
Methods of design and analysis
Transfer-function method State-variable method
Transient and
steady state
Response
analysis
Root locus
analysis
Frequency
response
analysis
Linear-
quadratic
optimization
Pole-placement
analysis
(Classical control) (Modern control)
Pole-placement method and Linear-
quadratic optimization are the main
methods of design and analysis.
Steady state and transient response
analysis, Root locus analysis and
frequency response analysis are the
main methods of design and analysis
Analysis consists of system of n first
order differential equations.
Analysis consists of single higher
order differential equation
Time domain methodFrequency domain method
It is useful for nonlinear and
complex systems also.
It is useful for linear and simple
systems only
Used for multi input multi output
(MIMO) systems can be used for SISO
also
Used for single input single output
(SISO) systems
Modern control methodClassical control method
State-space methodTransfer-function
method
Mechanical and electro magnetic
stiffness
mf
mx
mg
Magnetic spring
Operating
position0x
Rotor
mechanical spring
Equilibrium
position
sf x
mg
0x
Mechanical spring stiffness
magnetic displacement stiffness
mf
mg
Magnetic spring
Operating
position0x
0iOperating
current
mi Instantaneous
current
0x
Magnetic Bearing Control
• Equilibrium and Operating points
• For a mechanical spring there will be an
equilibrium point where the force resisted by the
spring is equal to the force applied on the spring
• For electro magnets there will be a quantity of
current corresponding to position of the object and
force applied. At this point the gravity force and
magnetic force will be equal. A slight movement
form this point will cause indefinite movement of
the body. This point is called operating point
13
Linearization of current Linearization of displacement
Linearization at operating point
0x
0img
0 mx x x= −
0mi i i= −
if k i=
xf k x=
is the instantaneous currentm
i is the instantaneous positionm
x
Linearized formula around the operating point will be
( , )x i
f x i k x k i= +
xk is displacement stiffness
ik is current stiffness
x
i is the deviation of current
from operating current
is the displacement from the
operating position
where
f is instantaneous force
• Linearized equation is suitable for most of the
applications of magnetic bearings
• It is not valid in three occasions
• When the rotor touches the bearing magnet
• When there are strong currents such that magnetic
saturation of the material occurs
• When or very small currents there won’t be
levitation of the rotor because of very small
magnetic forces.
0x x=
0i i= −
Magnetic Bearing Control
m
Rotor
xf
k c
spring mass damper
system
Active magnetic
bearing system
x if k x k i= +
f mx=
By Newton's law
Combing above two equations we get
x imx k x k i− =
If controlling current i is zero then
0x
mx k x− =
Response of magnetic
bearing without control
And the response grows exponentially thus
the rotor may fall down or touch the magnet
Response of magnetic
bearing with control
If we supply controlling current i such that
then it becomes
( ) x
i i
k k ci x x x
k k
+= +
0mx cx kx+ + =
And the response is imitated to a spring mass damper
system by the magnetic bearing system
m
Rotor
xf
k c
spring mass damper system
14
( ) x
i i
k k ci x x x
k k
+= +
PD controller model
• The model is PD-controller with proportional
and differential feed back
• In design of controller we choose the stiffness
and damping to ensure the system come to
steady state in optimum time.
• The optimal stiffness suggested is
• The range of damping ratio for better systems
suggested is 0.1 to 1
x
i
k kP
k
+=
i
cD
k=
xk k=
ci Pe De= +
Controller
ci
ci i=
ii
k ++−
r
y
1/ m ∫ ∫
xk
f x x x
Amplifier
Sensor
y x=
Block diagram of PD controller with
current control
e
1
c
i
i Pe De
edtT
= +
+ ∫
Controller
ci
ci i=
ii
k ++−
r
y
1/ m ∫ ∫
xk
f x x x
Amplifier
Sensor
y x=
Block diagram of PID controller with
current control
e
loadf∆
Control of rotors by using
magnetic bearings
Topics to be covered
• Rigid rotor model
• Flexible rotor model
Differences between mechanical and
magnetic bearing models
• Stiffness is very high
thus the vibration of the rotor will be transmitted to foundation
• Damping is directly observed due to
hydrodynamic effects
• Stiffness is very low thus the rotor can rotate freely
about the principal axes of inertia which results in a vibration isolation system.
• As the rotor is free in the air there is no coulomb damping acting on the
system. The control law will have damping term.
Mechanical bearing model Magnetic bearing model
15
Rigid rotor model
Rotor mechanical bearing system
α Infinitesimal rotation about x axis
β Infinitesimal rotation about y axis
d
dt
αα = d
dt
ββ =
Ω Angular velocity of shaft
Rigid rotor model
Angular velocity vector can be expressed as
0
0
0
cos sin
sin cos
x
y
z
t t
t t
ω α β
ωω α β
ω
Ω + Ω
= = − Ω + Ω Ω
z
y
x
O z
y
x
Oz
y
x
z'
y'
z'
x'
z
y
x
[ ] [ ]T T
1 2 3 4x x x x x yβ α= = −x
If the variable vector is chosen as
Motion about x- axis Motion about y- axis
α
β
Rigid rotor model
Equations of equilibrium can be obtained as by using Lagrange’s principle
i
i i
d T TF
dt x x
∂ ∂+ =
∂ ∂
is the generalized force corresponding to variableth
iF i
( ) ( )2 2 2 2 2 2
0 0 0 0 0 0
1 1
2 2x x y y z zT m x y z J J Jω ω ω= + + + + +
Kinetic energy is expressed as
Rigid rotor model
Equations (1) can be expressed in matrix form by rearranging
( )M G C+ + =x x F
F can be expressed as
( )K N= − +F x
)
is the gyroscopic matrix )
is the damping matrix )
is the inertia matrix (
( -
(
T
T
T
G
C
M M M
G G
C C
=
=
=
)
is non-conservative force matrix )
is conservative force matrix (
( -
T
TN
K K K
N N
=
=
Rigid rotor model
• Conservative forces include
– forces due to stiffness
• Non-conservative or circulatory forces include
– Internal or structural damping
– Steam or gas whirl in turbines
– Seal effects
– Process forces such as in grinding
– Unbalance, etc
• Damping include
– Coulomb damping due to hydrodynamic effects
Rigid rotor model
• From Eq. (2) and (3) we get
• If the non-conservative and gyroscopic
forces neglected, we have
( ) ( ) 0K NM G C ++ + + =x x x
0KM C+ + =x x x
16
Natural modes
• The solution of the equations (5) gives
four modes, for there are four degrees
of freedom considered
Translation mode Rotation mode
Natural modes
Forward whirl Backward whirl
Forward nutation Backward nutation
Magnetic bearing model
• In a magnetic bearing if we neglect the
conservative, non-conservative, and
damping effects, we will have
• For small rotations gyroscopic effects
can be neglected and the equations in x
and y directions can be decoupled
GM + =x x F
M =x F
Weight considerations
mg
0ig mg k if = =0
cosig
mgk if
α= =
α
Imbalance considerations
( )2 cosme tfω θ= Ω Ω +
is the imbalance massm
is the eccentricity of
imbalance mass
e
e
Ω
is the angular position
of imbalance mass
θ
Magnetic bearing model
It can be written as
wherec gk
mx f f f fω−= + +
( )
0
2 cos
x
i
i
k
c
g
k x
k i
mg k i
me t
f
f
f
fω θ
=
=
= =
= Ω Ω +
17
Magnetic bearing model
• It will be
• i at any instant will be
( ) ( )2
0cos
x ik x k i i me tmx θ− Ω Ω += + +
( )2
0
cosx
i
k x me ti i
k
mx θ− Ω Ω += +
−
Rigid rotor with magnetic bearing
• Three steps involved:
– Formulation with respect to centre of gravity
– Transformation with respect to the bearing coordinates
– Transformation with respect to the sensor coordinates
z
x
y
O
Bearing
Sensor
Centre of gravity
Why with respect to sensor
coordinates
• Sensors cannot be
arranged directly in the magnetic actuator.
• This requires certain
gap between the magnet and the sensor.
• The displacements with respect to sensor
coordinates will be transformed to bearing coordinates
With respect to centre of gravity
• In slow role x and y directions can be decoupled
y
mx f
I pβ
=
=
BA
z
x
y
O
a b
c d
axf
bxf
p
xβ
fIn matrix form as
where
0,
0 y
M
m xM
I
f
p
β
=
= =
=
x f
x
f
With respect to bearing coordinates
• Forces are transformed as
ax bx
ax bx
f f f
p af bf
= +
= +
1 1,
f B
f B
ax
bx
T
fT
fa b
=
= =
f f
f
BA
z
x
y
O
a b
c d
axf
bxf
p
xβ
f
With respect to bearing coordinates
1
1 1
B B
B
B
a
b
b a
T
x
b aT
x
x
β
−
=
=
−=
−
=
x x
x
x
BA
z
x
y
O
a b
c d
ax
bx
p
xβ
• Displacement vector can be
transformed as
f
18
With respect to sensor coordinates
fBA
z
x
y
O
a b
c d
axf
bxf
p
xβ
S S S B BT T T= =x x x
S
d
cx
x
=x
1
1S
cT
d
=x
β
=x
0
0
s s b
S B
s
S B
S
S
T
T TT
T T
=
=
=x
x xx
dx
cx
State feed back
• The control vector is found by using control
law
• We do not know the velocity components
directly from sensors. So a state observer is
required to find the velocities
sF= −u x
s SC=x x
is the full state vector
is the vector from the sensor
s
S
x
x
1
s s b s
s b
A T A T
B B
−=
=s s s sA B+= ux x
State space form with respect to sensor coordinates
State feed back
• The whole closed loop system can be shown as
block diagram
s s s sA B+= ux x
sF= −u x
S sC=x x
sB + C∫− s
x
sA
F
d
dt
u
( )s s s sA B F−=x x
decides the closed loop
dynamics of the system
s sA B F−
sx Sx
Model at high speeds
• At high speeds the gyroscopic effects cannot
be neglected, thus the model becomes
• The displacements in x and y directions no
longer decoupled, so four forces and four
displacements should be taken into
consideration simultaneously.
• The same procedure is to be followed as for
the slow rotation
GM + =x x F
Model at high speeds
0 0 0
0 0 0
0 0 0
0 0 0
y
x
m
I
m
I
M
=
0 0 0 0
0 0 0 1
0 0 0 0
0 1 0 0
G
−
=
x
y
y
x
f
p
f
p
−
=fB
a
b
a
b
y
y
x
x
=x
Conclusions on rigid rotor model
– There is an optimal design for each speed
– The optimal design at higher speed may not
be stable at lower speeds, for the gyroscopic
effects are reduced.
– The optimal design at zero speed may not be the optimal at higher speeds
– The gyroscopic effects will not destabilize the system which is stable at lower speeds.
– Further more the design at lower speeds is decoupled and easier to design. Decentralized designs for lower speeds can be implemented
19
Conclusions on rigid rotor model
– Thus for stability considerations and other
advantages systems are designed for
lower speeds and with decentralization
xa xa aF xu −=
xb xb bF xu −=
yb yb bF xu −=
ya ya aF xu −=
Decentralized control mode scheme
Flexible rotor model
• Rigid rotor can be
defined by two
points
• Flexible rotor has
infinite degrees of
freedom. One
cannot define
uniquely by some of
the points
Flexible rotor model
• Equation motion of
an Euler-Bernoulli
beam is given by
• The variable
separable form is
4 2
4 20
y yEI m
z t
∂ ∂+ =
∂ ∂L
( , ) ( ) ( )y z t Y z q t=
z
Flexible rotor model
• By substituting we get
• By rewriting we get
4 2
4 2
2
( ) ( )
( ) ( )
d Y z d q t
dz dtEI
m Y z q tω
= − =
4 24
4
( )( ) 0,
/
d Y zY z
dz EI m
ωβ β− = =
22
2
( )( ) 0
d q tq t
dtω+ =
Flexible rotor model
• By applying initial
conditions and solving we get the natural frequencies
• By substituting the Eigen values in (29) we get the Eigen functions
or model functions
ω
Lz
( )Y z
0ω
1ω
Rigid rotor modes
z
z
2ω
3ω
Flexible rotor modes
• The mode shapes or
modal functions
depend on the end
conditions
20
Actuator sensor location
• Sensor should not be set at nodes
• Sensor and actuator should not lie on
opposite sides of a nodeactuator
sensor
Actuator sensor location
• We can conclude that the sensor can be set at a place where we can get
information from each mode under
consideration
Modal reduction
• While designing a flexible rotor system, we
can not consider all the modes of the system
for they are infinite
• Thus we consider first n number of modes
corresponding to first n natural frequencies
and neglect the remaining modes
• If we study the effect of the reduced modes
we can find the number of modes which we
can consider without destabilizing the system
Modal reduction (mathematical representation)
• Mathematical model of the
– full system
– Divided system
– Reduced system
A B
C
= +
=
x x u
y x
[ ]
M M MR M M
R RM R R R
M
M R
R
A A B
A A B
C C
= +
=
x xu
x x
xy
x
M M M M
M
A B
C
= +
=
x x u
y x
Modal reduction
• The reduced modes
give three kinds of
effects on the system
called spillovers
– Control spillover (By the input)
– Interconnection spillover
(By the parameters of the system)
– Observation spillover (on
the estimated output)
Input System Output
Control spillover
Interconnection spillover
Observation spillover
[ ]
M M MR M M
R RM R R R
M
M R
R
A A B
A A B
C C
= +
=
x xu
x x
xy
x
Modal reduction
Block diagram of effect of model reduction
∫
MA
MB MC+ +Mx yu
∫+ Rx
RA
RB RC
RMA MRA
Control spillover
Interconnection spillover
Observation spillover
Modeled modes
Unmodeled modes
21
Conclusion on flexible rotor control
• Modal reduction is studied to consider
the number modes to be taken into
consideration for having stable control
• Mechanical design is studied for finding
the sensor actuator locations
Conclusions
• Magnetic bearings advantages and applications have been discussed
• Electromagnetism and Control system technologies have been introduced
• Design of thrust and radial magnetic bearings have been studied
• Control of a rotor by rigid rotor and flexible rotor models have been studied
Schweitzer, G., Bleuler, H. and Traxler, A., 2003, “Active Magnetic Bearings: Basics, Properties and Applications of Active
Magnetic Bearings”, Authors Working Group, www.mcgs.chreprint.
Chiba, A., Fukao, T., Ichikawa, O., Oshima, M., Takemoto, M. and Dorrell, D.G., 2005, “Magnetic Bearings & Bearingless
Drives”, Newnes, Elsevier.
Maslen, E., 2000, “Magnetic Bearings”, University of Virginia.
Groom N.J. and Bloodgood, V.D. Jr., 2000, “A
Comparison of Analytical and Experimental Data for a Magnetic Actuator”, NASA-2000-tm210328.
Bloodgood, V.D. Jr., Groom, N.J. and Britcher, C.P., 2000, “Further development of an optimal design approach applied to
axial magnetic bearings”, NASA-2000-7ismb-vdb.
Further References
Anton, V.L. , 2000, “Analysis and initial synthesis of a
novel linear actuator with active magnetic suspension”, 0-7803-8486-5/04/$20.00 © 2004 IEEE
Chee, K.L., 1999, “A Piezo-on-Slider Type Linear
Ultrasonic Motor for theApplication of Positioning Stages”, Proceedingsof the 1999IEEE/ASME.
Shyh-Leh, C., 2002, “Optimal Design of a Three-Pole
Active Magnetic Bearing”, IEEE TRANSACTIONS ON MAGNETICS, VOL. 38, NO. 5.