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

net/publication/271699784

Linear Quadratic Regulator controller for Magnetic Levitation System

Article in International Journal of Computer Science and Software Engineering · January 2015

CITATIONS READS
6 2,425

3 authors:

Sadaqat Ur Rehman Muhammad Junaid

University of Salford China University of Mining and Technology
79 PUBLICATIONS 1,356 CITATIONS 45 PUBLICATIONS 166 CITATIONS

SEE PROFILE SEE PROFILE

Obaid Ur Rehman
Sarhad University of Science & IT
48 PUBLICATIONS 576 CITATIONS

SEE PROFILE

All content following this page was uploaded by Muhammad Junaid on 17 January 2019.

The user has requested enhancement of the downloaded file.

International Journal of Computer Science and Software Engineering (IJCSSE), Volume 4, Issue 1, January 2015
ISSN (Online): 2409-4285 www.IJCSSE.org Page: 6-9

Linear Quadratic Regulator controller for Magnetic Levitation System

Engr. Sadaqat Ur Rehman1, Engr. Muhammad Junaid2 and Engr. Obaid Ur Rehman3

1, 2, 3
Department of Electrical Engineering, Sarhad University of Science and IT, Peshawar, Pakistan

1
[email protected], [email protected], [email protected] and [email protected]

ABSTRACT system refers to the sideward force that requires moving

This paper explains Magnetic Levitation system of a train which the train on the track. Propulsion system uses an
comprises of guidance track made with magnets. The main electrically powered motor in the guide way which appears
objective is to design a proper controller that can suspend and to be the favoured option for high speed Maglev trains.
propelled the train on a guidance track made with magnets. To Levitation system keeps the train suspended against the
perform the desired task state space model of Magnetic
Levitation system is derived. The response of the system is
gravity by forces of magnetic field. These systems have
simulated in MATLAB. The open loop response showed that the unstable open loop response, to make the response of the
derived model is unstable. Linear Quadratic Regulator (LQR) system stable feedback path was used. Linear Quadratic
controller is designed to analyze the system in closed loop. The Controller (LQR) was used to make the closed loop
controller showed improved performance for different tracks. response of the system stable [3].
Different types of realization techniques (minimal realization, Valer and Lia build a nonlinear model for magnetic
balanced realization, modal realization, observer canonical levitation system and proposes systems linearization
realization) are compared for minimum fragility in controller principle (the expansion in Fourier series and the
implementation. The difference among the different preservation of the first order terms) in order to linearize
realization controllers has been analyzed in detail for
rounding off error or truncation error and an optimal non
the acquired nonlinear model [4, 5]. Our interest here is to
fragile controller design has been presented. Different design a Non-fragile optimal controller so a linear
disturbances were imposed upon the simulated model. All the controller was designed to give safety and ride comfort to
results are analyzed in open and closed loops. The closed loop passengers inside the cabin of a train. The complete
response showed that the train is suspend and propelled on the mathematical derivation for the Magnetic Levitation train
track and the desired results were achieved successfully model in state space form is simulated in MATLAB /
SIMULINK [6]. Magnetic levitation is very useful in High
Voltage insulation testing [7]. The block diagram of the
Keywords: LQR, Magnetic Levitation System, Realization Magnetic Levitation train is shown in the Figure 1.
techniques, Bullet Train, MATLAB Simulink.

1. INTRODUCTION
Magnetic levitation systems have many varied uses such
as in friction less bearing, high speed maglev passenger
trains, levitation of wind tunnel model, vibration isolation
of sensitive machinery levitation of molten metal in
induction furnaces and the levitation of metal slabs during
manufacture (Laithwaite 1965, Jayawant and Rea 1965)
[1]. Here our objective is to design such a controller which
can suspend and propelled the Maglev train on the track.
The maglev train is based on three type of systems which
control the train moves on the track Guidance system,
Propulsion system and levitation system [2]. Guidance
Fig. 1. Block Diagram of Magnetic Levitation System
 International Journal of Computer Science and Software Engineering (IJCSSE), Volume 4, Issue 1, January 2015 7
Engr. S. U. Rehman et. al

2. MATHEMATICAL MODEL OF states). Balanced realization, Modal realization and

MAGNETIC LEVIATION SYSTEM Observer based canonical realization are the other
different techniques used to obtained a reduced and non-
State space model of Magnetic Levitation is derived as fragile model.
given in [4].

V= dx/dt (1) 4. RESULTS AND CONCLUSION

u= Ri + dL(x)i/dt (2)
This work is carried out on considering a magnetic
mẍ = mg-c(i/x)2 (3)
levitation system. The mathematical derivations are done
in state space form. For simulations MATLAB software is
Equation (2) indicates that L(x) is a nonlinear function.
used. The open loop response in MATLAB shows
Various approximate values are used to determine the
oscillations, large overshoot and required large settling
value of inductance for the Magnetic Levitation. If we take
time to damp. LQR controller/compensator was designed
the assumption that the inductance of the system varies
to obtain the desired response. LQR controller improved
with the inverse of the ball position
the performance of the system. The results obtained were
satisfactory. Different realization techniques are then used,
L(x) = L + L0x0/x (4)
by applying these techniques the controllers action is made
more efficient and the system is made highly stable and
Where L is the constant Inductance of the coil in the
non-fragile.
absence of the ball, L0 is the additional inductance
contributed by the presence of the ball
A. Open Loop Response
u(t)= iR+ d/dt (Lc + L0x0/x)i
u(t)= iR+ Ldi/dt - (L0x0i/x2)dx/dt

Substituting L0 x0 =2C [4], we get

u(t)= iR+ Ldi/dt – C(i/x)2dx/dt (5)

Linear Model of the System is

Fig. 2. Open loop Response of Magnetic Levitation System

The open loop response is unstable, thus the open loop

response is very uncomfortable for the passengers.
3. METHODOLOGY
By taking the Open loop response of the system, open loop B. LQR Controller Response
response of the system shows that the system is unstable.
Linear Quadratic Regulators are used to overcome the
disturbance effect and overcome the disturbing effect and
to improve the performance parameter and make the close
loop response of the system is stable, realization technique
is used to obtain a r educed and non-fragile model [3].

A. Realization Technique
In order to obtain a reduced and non-fragile optimal
controller different realization techniques are used.
Minimal realization (The realization is known as
"minimal" as it defines the system with least number of Fig. 3. Closed Loop Response after (LQR Controller)
 International Journal of Computer Science and Software Engineering (IJCSSE), Volume 4, Issue 1, January 2015 8
Engr. S. U. Rehman et. al

Thus the LQR Controller gives a better response and

settled the oscillations in the vehicle body quickly. After
the addition of LQR controller the system becomes stable.

C. Minimal Realization
For LQR controller no state has been reduced, the
controlled response has three states and after minimal
realization the states remain the same. Difference between
the LQR controller and minimal realization response was
plotted as shown in Figure 4 Fig. 6. Modal Realization of the Closed loop System (LQR Controller)

F. Observer Canonical Realization

Difference between the LQR controller and Observer

Canonical realized response was plotted as shown in the
Figure7.

Fig. 4. Minimal Realization of the Closed loop System (LQR Controller)

D. Balanced Realization
Difference between the LQR controller and balanced
realization response was plotted as shown in the Figure 5.

Fig. 7. Observer Canonical Realization of the Closed loop System (LQR

Controller)

A brief summary of all types of realization techniques was

given below in table 1. This table shows that observer
canonical realization gives the least error to controller
which represents the most optimal and most non-fragile
optimal controller technique.

Table 1: Realization Analysis for different controllers

Realization type LQR controller

Fig. 5. Balanced Realization of the Closed loop System (LQR Controller)
Minimal Realization 10-15
Balanced Realization 10-15
E. Model Realization Modal Realization 10-14
Observer canonical
10-16
Difference between the LQR controller and modal Realization
realization response was plotted as shown in Figure 6.
 International Journal of Computer Science and Software Engineering (IJCSSE), Volume 4, Issue 1, January 2015 9
Engr. S. U. Rehman et. al

For different input disturbances the LQR controller shows

better response. The LQR controller settles the oscillations Engr. Muhammad Junaid is currently serving as a
more quickly, reducing the oscillation and overshoot. The Lecturer (Power Engineering) and member/team leader,
designed LQR controller provides better handling ability Quality Enhancement Cell (QEC) at the Department of
for wide range of disturbances and provides better ride Electrical Engineering, SARHAD University of Science
comfort for passengers. & IT Peshawar, KPK Pakistan. He is a Power Engineering
Also observer canonical realization gives the least error to Specialist having done MS in Electrical Engineering with
controller which represents the most optimal and most the Specialization in Power System Engineering in 2015
non-fragile optimal controller technique. recently & B.Eng in Electrical Power Engineering in 2011
respectively. His Area of Research is in High Voltage
Engineering/ Power System Engineering.He is an active
REFERENCES member of IAENG HONG KONG, IEEE PES, IJENS-
RPG Researchers Promotion Group, Pakistan Engineering
[1] Walter Bariet and John Chiassont, “Linear and Non Linear Council (PEC) & Institute of Engineers Pakistan (IEP).
State space controller for magnetic levitation”, International Cell: +92-345-9150634
Journal of System Sciences, 1996.J. Clerk Maxwell, A
Treatise on Electricity and Magnetism, 3rd ed., vol. 2.
Oxford: Clarendon, 1892, pp.68-73. Engr. Obaid ur Rehman is an Assistant Professor at the
[2] NASA Florida Space Institute, “Magnetic Levitation Space Department of Electrical Engineering, SARHAD
Propulsion”. University of Science & IT Peshawar, KPK Pakistan. He
[3] M.Abid Saeed, Niaz Ali, Muhammad Waqas and Anees ur is Currently PhD Scholar at College of Electrical
rehman, “Design of a Non Fragile Optimal Controller Engineering, Zhejiang University, PR China
Realization for Vehicle Suspension Systems for Different
Road Profiles”, Sarhad University of Science and
Information Technology Peshawar Pakistan & National
Power control center NPCC, Pakistan, IOSR Journal of
Electronics and Communication Engineering, Feb 2014..
[4] Ishtiaq Ahmad and Muhammad Akram Javaid, “Nonlinear
Model & Controller Design for Magnetic Levitation
System”, University of Engineering and Technology Taxila
Pakistan, International Conference on signal Processing,
robotics and automation, 2010.
[5] Valer Dolga, Lia Dolga “Modeling and Simulation of a
Magnetic Levitation System” ANNALS of the ORADEA
UNIVERSITY. Fascicle of Management and Technological
Engineering, Volume VI (XVI), 2007.
[6] Carnegie Mellon, “Control Tutorial for Mat lab”, The
University of Michigan, 1997.
[7] Junaid Muhammad, Mehran Tahir, and Muhammad Aamir.
"HVAC Breakdown of R12, R12–Air Mixtures in Sphere gap
Electrodes and Proposed Composite insulating Liquid to
Reduce the Amount of Transformer Oil In Transformers."
(2013).

AUTHOR PROFILES:

Engr. Sadaqat Ur Rehman is currently serving as a

Lecturer & member, QEC at the Department of Electrical
Engineering, SARHAD University of Science & IT
Peshawar, KPK Pakistan. He has done MS in Electrical
Engineering with a Specialization in Electronics &
Communication Engineering in 2014 & BSc in Computer
Systems Engineering in 2011. His Area of Research is
Image/Video Processing. He is an active member of
Pakistan Engineering Council (PEC)
Cell: +92-333-9206903

View publication stats