Modeling and State Feedback Controller Design of Tubular Linear Permanent Magnet Synchronous Motor

International Journal of Power Electronics and Drive System (IJPEDS)
Vol. 7, No. 4, December 2016, pp. 1410~1419
ISSN: 2088-8694, DOI: 10.11591/ijpeds.v7i4.pp1410-1419  1410
Journal homepage: http://iaesjournal.com/online/index.php/IJPEDS
Modeling and State Feedback Controller Design of Tubular
Linear Permanent Magnet Synchronous Motor
Hossein Komijani1
, Saeed Masoumi Kazraji2
, Ehsan Baneshi3
, Milad Janghorban Lariche4
1,2
Faculty of Electrical and Computer Engineering, University of Tabriz, Tabriz, Iran
3
Department of Electrical Engineering, Islamic Azad University, Neyriz, Iran
4
Abadan School of Medical Sciences, Abadan, Iran
Article Info ABSTRACT
Article history:
Received Apr 26, 2016
Revised Nov 30, 2016
Accepted Dec 13, 2016
In this paper a state feedback controller for tubular linear permanent magnet
synchronous motor (TLPMSM) containing two gas springs, is presented.
The proposed TLPMSM controller is used to control reciprocating motions
of TLPMSM. The analytical plant model of TLPMSM is a multi-input
multi-output (MIMO) system which is decoupled to some sub single-input
single-output (SISO) systems, then, the sub SISO systems are converted to
sub-state space models. Indeed, the TLPMSM state space model is decoupled
to some sub-state spaces, and then, the gains of state feedback are calculated
by linear quadratic regulation (LQR) method for each sub-state space
separately. The controller decreases the distortions of the waveforms.
The simulation results indicate the validity of the controller.
Keyword:
Multi input multi output system
Single input single output
State feedback controller
system
Tubular linear permanent
magnet synchronous motor
Copyright © 2016 Institute of Advanced Engineering and Science.
All rights reserved.
Corresponding Author:
Hossein Komijani,
Faculty of Electrical and Computer Engineering,
University of Tabriz, Tabriz, Iran.
Email: h.komijani90@ms.tabrizu.ac.ir
1. INTRODUCTION
Tubular linear permanent magnet synchronous motors are used in industrial application such as the
linear compressors, elevators, pumps, electromagnetic valve actuators, vibrators, and the industrial robots
[1-3]. This kind of linear motors has some advantages such as good dynamic respond, accurate position,
and reliability [4]. The TLPMSM reciprocating motion is used in drilling applications inoil industry
efficiently. TLPMSM has two gas springs that acts like a linear hammer, particularly suitable for drilling the
hard rocks. This kind of motor can transmit electrical power from an electric source to drill without using any
mechanical intermediate. Existence of two gas springs in both sides of the piston increases the efficiency
significantly and decreases the drilling time [18]. Several control techniques are used for controlling the
TLPMSM reciprocating motion. To control the position, speed, acceleration, and force of linear permanent
magnet synchronous machine, modeling, dynamic analysis, and parameter estimation have been studied in
[4-7]. Prescribed closed-loop speed control method offered an accurate realization in [8]-[9]. The sensor-less
control method for a miniature application has been validated in [3]. An advanced scheme, based on
neuralnetwork, has been proposed in [10] to compensate for sudden variations of the load. The TLPMSM
which is used in this work has been introduced in [11] and [12]. Analysis of the force performance of this
machine has been carried out in [11].
Displaying system, using its transfer function, defines the input-output behavior of the system inthe
time domain.The state space equations, not only has the input-output data but also prepares comprehensive
information of inner system structure to control engineers. Displaying transfer function in the frequency
domain is used in classical designs and SISO compensators and its efficiency is limited in analyzing

IJPEDS ISSN: 2088-8694 
Modeling and State Feedback Controller Design of TLPMSM (Hossein Komijani)
1411
and designing MIMO systems. Displaying system in space state universalize analyzing and designing of
SISO systems to MIMO systems [21].
In this paper, TLPMSM modeling and state feedback controller have been considered to have good
dynamic responses. A state space that increasingly exploited in diverse fields in order to modeling and
analyzing systems, is a mathematical model originated form some first-order differential equations describing
a physical system. A state space includes a set of state variables, input variables, and output variables. The
variables are defined in vector form. Moreover, if the considered system is a linear and time-invariant
system, the equations can be expressed in matrix form. State feedback can arbitrarily assign the closed loop
system poles at any position but has no effect on the system zeroes [19]-[20].
Permanent
Magnet
Dril String
External Spring
Casing
Iron
Stator Winding
Gas Spring
Drill Bit
Drill String
External
Spring
Casing
Electric Force
Friction between
Piston and Casing
Equivalent Spring
Piston
0
Displacement
PistonDisplacement
Casing
Friction between
Surrounding and Casing
Force From Drill Bit
Figure 1. TLPMSM Figure 2. Equivalent model of TLPMSM
The rest of paper is organized as: Overview of the TLPMSM model and the equivalent circuit are
given in section II. In section III, the TLPMSM mathematical model is presented. Section IV presents the
theory and design of state feedback controller. The simulation results for the controlled system with state
feedback controller in full load mode are presented in section V. Finally, section VI presents summary and
conclusions.
2. MATHEMATICAL MODELING OF TLPMSM
The TLPMM prototype is shown in Figure 1. The explained TLPMSM is included a case with the
stator winding, piston with permanent magnets, and two gas springs on both sides of the case [13].Figure.3
shows the equivalent circuit of TLPMSM stator winding. To facilitate simulations, the equivalent form of
spring and dampers is used for modeling gas springs and friction between the case and piston. The whole
system hits the rock with a constant speed. Each collision makes some cracks between rocks and makes some
more space between drill beat and rock, causes better performance for next hitting. Themiddle point of the
piston isdefinedas an origin for the case. The piston and the positive case displacement direction are shown in
Figure 2. According to the mentioned equivalent circuit, piston oscillates between gas springs [13].
Gas springs are assumed as linear elements, while the stroke length of piston is small enough
compared with the length of gas springs, therefore, the characteristic function of gas springs is obtained as
follows:
(1)
where is the gas spring force, is the equivalent stiffness coefficient formatting and y is the spring
displacment.

 ISSN: 2088-8694
IJPEDS Vol. 7, No. 4, December 2016 : 1410 – 1419
1412
The electromagnetic force, friction force between case and piston, and the spring force are the three
main forces that are subjected to the piston as illustrated in Figure 4. The following equation in vertical
direction can be expressed using Newton’s second law:
̈ ̇ (2)
Figure 3. Equivalent electric circuit of stator
winding
Figure 4. Force analysis for piston
where is mass of piston, is friction coefficient between piston and casing, is stiffness coefficient of
equivalent spring, ̇ ̇ ̇ is piston velocity respective to the case, is
piston displacement respective to thecase, is the piston displacement and is the case displacement.
The subjected forces to case are shown in Figure 5. According to Figure 5, case is subjected to four forces:
i. Counter force by the piston-gas spring-damper subsystem
ii. Friction force between the case and neighboring
iii. External spring force
iv. Force from rock subjected to case because of impact.
Equation in vertical direction is obtained as (3):
̈ ̇ (3)
wher e is the mass of casing, is the friction coefficient between casing and surroundings, is stiffness
coefficient of external spring, is the external spring force, ̇ is the friction force between the
case and neighboring, is acounter force by rocks because of impact which is assumed as follows:
{ (4)
Using the Lorentz force equation, the electromagneticforce is obtained as:
(5)
where is stator winding current, B is flux density around stator winding.
Voltage equation for the circuitshown in Figure 3 is acquired from Kirchhoff's voltage law (KVL):
(6)
where is stator’s winding input voltage, R is resistance of coil, and L is inductance of coil,
isback EMF of stator winding. Using Faraday’s law, back EMF of stator winding, is
obtained as:
̇ ̇ (7)

1413
where D is diameter of chamber, and N is number of turns of stator winding. State space variables are
determined:
, ̇ , , ̇ , .Using equations (1) to (7), the state space model
is obtained as:
[
̇
̇
̇
̇
̇ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
(8)
3. STATE FEEDBACK CONTROLLER DESIGN
Undoubtedly, most control methods are based on the mathematical model of physical systems. In
classic methods, transfer function in the frequency domain is used to design of control systems such as
frequency response and Root-Locus method. These models are presented simple SISO physical and industrial
systems properly, also approximate the input-output behavior of systems with acceptable accuracy [14].
Precision study of complicated industrial systems needs more integral models. For controlling these
kinds of system with optimal function, progressive control system designs are required.. Description of state
space of a system gives a complete view of internal system structure. State variables describe internal
dynamics of the system. This model shows how state variables have mutual effect on each other, how input
signals affect state variables, and how can compute the output response with various synthesizing of state
variables. Models of SISO systems can be universalized to MIMO systems easily with this method [15]-[16].
The significant advantage of the modern control system analysis and design according to classical control
systems is their usage in the MIMO and the time-invariant (TI) systems, however, the classical control
system usage is only in SISO-LTI systems [17]. The other advantage of using the state space model is the
convenience of close loop system function optimization. Therefore, optimal control systems can be designed
in space state. The most utilization of state space concept in modern control are the pole assignment and the
stabilization of systems that implemented with state feedback variables. In this paper, a closed-loop state
feedback controller is designed for a TLPMSM. Generally, the state space system equation can be written as:
̇
(9)
where t is time variable, X(t) is state vector, U(t) denotes the controlling signal, Y(t) is output vector and A,
B, C, D are constant matrixes. The values of TLPMSM prototype parameters areexpressed in Table 1. With
these parameter values, the poles of system as shown in Figure 5.
-fnet(t)
d . fc
yc
Kext.yc(t)bc.ẏc(t)
Figure 5. Force analysis for casing
Table 1. TLPMM prototype parameters

 ISSN: 2088-8694
1414
Parameter Value Unit
Flux density around stator winding (B) 0.42 T
Diameter of chamber (D) 0.0476 M
Number of turns of stator winding(N) 157 -
Period of collision (T) 0.005 S
Mass of piston ( ) 1.68 Kg
Mass of casing ( ) 5.14 Kg
Friction coefficient between piston and casing ( ) 0.005 -
Friction coefficient between casing and surroundings ( ) 0.001 -
Stiffness coefficient of equivalent spring ( ) 7.93 e4 N/M
Stiffness coefficient of external spring ( ) 2.578 e4 N/M
Resistance of Coil (R) 1.20 Ω
Inductance of Coil (L) 0.00119 H
(10)
The state space controlling signal is shown as:
(11)
where matrix k is the state feedback gain and R(t) is the reference input signal. With this assignments, the
equation (8) is converted as follows:
̇ (12)
The appropriate pole assignment can be obtained by setting matrix k to the proper value. The sate
feedbackschematic model for this approach is shown in Figure 6.
This paper focuses on designing a closed-loop state feedback controller for stable output respond in the full-
load mode for a TLPMSM. Therefore, the closed-loop controller design objective is considered to track a set
of desired references. The transfer function can obtain from state space model as:
(13)
According to (9), this system has two inputs and one output, therefore, the transform function matrix is
obtained as follows:
[ ]
(14)
For designing state feedback controller, these two separate transfer functions are converted to space
states separately (each array of G matrix is converted to one state space) as shown in Figure 7. Indeed, the
state space of TLPMSM is converted to two sub-state spaces (SS1 and SS2). Each sub-state space is a SISO
system which can becontrolled by the gain of state feedback controller separately. Designing state space
controller for SISO system is much easier than MIMO systems.
In the next step of controller designing, the optimal gain of state feedback is calculated for obtained
state spaces including both of two separate state spaces using linear quadratic regulator (LQR) method in
sense of optimal control. The LQR algorithm is a well-known procedure of finding the pertinent gain of state
feedback controller [22]. Briefly, in LQR, for a continuous time system, the state feedback
law minimizes the quadratic cost function of:

1415
CCBB
AA
-k
Y(t)X(t)Ẋ (t)U(t)R(t)
ʃʃ∑∑∑∑
+
input1
input2
ss1
ss2
Figure 6. State feedback control model Figure 7. Separating state space model of
TLPMSM to two state spaces
∫ (15)
k is calculated as:
(16)
where is computed by associated Riccati equation as follow:
(17)
where R and Q are symmetric positive matrixes. State feedback gains can be achieved using LQR as follows:
[ ]
[ ] (18)
where is the state feedback gain for first state space and is the gain of second space state. With these
gains, the poles replacement is done as:
(19)
It is clear that the poles on the right side of real axis in (9) have been displaced to the left side of real axis
which makes the system stable. This replacement makes enhancement in the system response.
4. SIMULATION RESULTS
In this section, the results of TLPMSM simulation and the designed state feedback controller are
presented. The system is considered in the full-load condition. In simulations, beats are happened in TI mode.
The transferred energy equation can be shown as follows:
̇ ̇ (20)
where the terms of ̇ is the kinetic energy in casing before the collision, and the terms of
is the potential energy in external spring before the collision. The overall left-hand side of (20)
is the system energy before collision. The term of ̇ is the kinetic energy in casing after the
collision, is the potential energy in the external spring after the collision, is the transferred
energy to the rock during the collision, and the overall right side term of (20) is the system energy after the
collision.

 ISSN: 2088-8694
1416
The effective energy incomes to rocks is only the kinetic energy. Potential energy remains in the
system and has no direct effect to the rocks. The ideal beat happens when in (20). In this case, the
whole kinetic energy andthe potential energy are restored in the system. So the energy transfer equation is:
̇ (21)
Therefore, the most of energy transferring rate is based on the motion of the case. Base on collection
(2) to (6), the case motion is a function of the input current. By utilizing state feedback controller, input
current walks away from instability mode, makes the moment of case movement, according to most
energytransfers to rocks at the hitting moment. So the system acts on full-load mode if each beat takes place
in the mentioned point. The input voltage of TLPMSM is shown in Figure 8. The amplitude of input voltage
is 12V and the frequency is even with the natural frequency of mass-spring-damper.
Input current in the moment of hitting to the rock, before and after using space feedback is
illustrated in Figure 9. According to Figure 9, the current waveform in full-load mode by using state feedback
controller is similar to current waveform in no-load mode that shows good functioning and good system
dynamic response of the system via state feedback controller. It is also illustrated that the most of stored
energy in the case is transferred to the rock, and the little bit of energy stored in the spring is stored for
displacement of the case in this period. Back EMF before and after exerting state feedback controller is
shown in Figure 10. Electromagnetic force before and after exerting space feedback is illustrated in
Figure 11.
Indeed, this force is the driving piston force that is called electromagnetic force. By using proposed
controller, this force is similar in both cases of full-load and no-load conditions, shows the proper behavior of
system under space feedback controller.
Figure 8. Input voltage
(a) (b)
Figure 9. Input current: (a) before and (b) after exerting state feedback controller
0 0.2 0.4 0.6 0.8
-15
-10
-5
0
5
10
15
Time(s)
Voltage(V)
0 0.1 0.2 0.3 0.4
-3
-2
-1
0
1
2
3
Time(s)
InputCurrent(A)
0 0.1 0.2 0.3 0.4
-3
-2
-1
0
1
2
3
Time(s)
InputCurren(A)

1417
(a) (b)
Figure 10. Back EMF (a) before and (b) after exerting state feedback controller
(a) (b)
Figure 11. Electromagnetic force: (a) before and (b) after exerting state feedback controller
(a) (b)
Figure 12. Piston velocity respect to case velocity: (a) before and (b) after exerting state feedback controller
The piston velocity respect to the case velocity, with and without state feedback is shown in
Figure 12. According to Figure 12(a), the system cannot keep its stability without using state feedback
controller.Therefore, these velocities cannot coordinate together and the efficiency is descended drastically.
The system keeps stability by using space feedback controller, and acts in ideal functioning point as shown in
Figure 12(b).
0 0.1 0.2 0.3 0.4
-10
-5
0
5
10
Time (s)
BackEMF(V)
0 0.1 0.2 0.3 0.4
-5
0
5
10
Time(s)
BackEMF(V)
0 0.1 0.2 0.3 0.4 0.5 0.6
-50
0
50
Time(s)
fe(N)
0 0.1 0.2 0.3 0.4 0.5 0.6
-50
0
50
Time(s)
fe(N)
0 0.1 0.2 0.3 0.4 0.5 0.6
-20
-10
0
10
20
Time (s)
Vp-Vc(m/s)
0 0.1 0.2 0.3 0.4 0.5 0.6
-15
-10
-5
0
5
10
15
Time(s)
Vp-Vc(m/s)

 ISSN: 2088-8694
1418
5. CONCLUSION
In this paper, a new state feedback controller for TLPMSM is proposed. The concept of TLPMSM
with two gas springs and state feedback control has been achieved in details. Functioning and dynamic
responseare studied in full-load condition. The TLPMSM state space model is decoupled to some sub-state
spaces. Indeed, the MIMO model of TLPMSM is decoupled to some sub-SISO systems, and then, these
sub-SISO systems are converted to sub-state space models, and the gains of state feedback are calculated by
LQR method for each sub-state space separately. Results are encouraging and confirm the effectiveness of
proposed state feedback method. By using of this state space controller for explained TLPMSM, nearly the
whole energy is transferred to the rock in the collision moment, and in this situation, high efficiency is
achieved.
REFERENCES
[1] Souissi A, Abdennadher I, Masmoudi A. Analytical Prediction of the No-Load Operation Features of Tubular-
Linear Permanent Magnet Synchronous Machines. IEEE Transactions on Magnetics. 2016 Jan; 52(1): 1-7.
[2] Huang X, Tan Q, Wang Q, Li J. Optimization for the Pole Structure of Slot-Less Tubular Permanent Magnet
Synchronous Linear Motor and Segmented Detent Force Compensation. IEEE Transactions on Applied
Superconductivity. 2016 Oct; 26(7): 1-5.
[3] Lu H, Zhu J, Guo Y, Lin Z. A miniature short stroke tubular linear actuator and its control. In Electrical Machines
and Systems, 2007. ICEMS. International Conference on 2007 Oct 8; 1680-1685.
[4] Tavana NR, Shoulaie A. Pole-shape optimization of permanent-magnet linear synchronous motor for reduction of
thrust ripple. Energy Conversion and Management. 2011 Jan 31; 52(1): 349-54.
[5] Chen X, Hu J, Chen K, Peng Z. Modeling of electromagnetic torque considering saturation and magnetic field
harmonics in permanent magnet synchronous motor for HEV. Simulation Modelling Practice and Theory. 2016
Aug 31; 66: 212-25.
[6] Sun Y, Wu X, Bai L, Wei Z, Sun G. Finite-time synchronization control and parameter identification of uncertain
permanent magnet synchronous motor. Neurocomputing. 2016; 207: 511–518
[7] Hama T, Sato K. High-speed and high-precision tracking control of ultrahigh-acceleration moving-permanent-
magnet linear synchronous motor. Precision Engineering. 2015 Apr 30; 40: 151-9.
[8] Ummaneni RB, Nilssen R, Brennvall JE. Demonstration model of a linear permanent magnet actuator with gas
springs. In Electrical Machines, 2008. ICEM 2008. 18th International Conference on 2008 Sep 6; 1-4.
[9] Lu L, Chen Z, Yao B, Wang Q. A two-loop performance-oriented tip-tracking control of a linear-motor-driven
flexible beam system with experiments. IEEE Transactions on Industrial Electronics. 2013 Mar; 60(3): 1011-22.
[10] Naso D, Cupertino F, Turchiano B. Precise position control of tubular linear motors with neural networks and
composite learning. Control Engineering Practice. 2010 May 31; 18(5): 515-22.
[11] Lu Q, Huang L, Ye Y, Huang X, Fang Y. Design of a novel permanent magnet linear synchronous motor with
segmented armature core for ropeless lifter. COMPEL: The International Journal for Computation and
Mathematics in Electrical and Electronic Engineering. 2016 Mar 7; 35(2): 556-71.
[12] Wang T, Liang H, Jiao Z, He P, Yan L. Dynamics modeling and load analysis of linear motor for LEHA system. In
Fluid Power and Mechatronics (FPM), 2015 International Conference on, 2015 Aug 5; 1128-1133.
[13] Zhang S, Norum L, Nilssen R. Analysis of tubular linear permanent magnet motor for drilling application. In
International Conference on Electric Power and Energy Conversion Systems, 2009. EPECS’09 2009 Nov 10 (pp.
10-12).
[14] Rudra S, Barai RK, Maitra M. Nonlinear state feedback controller design for underactuated mechanical system: A
modified block backstepping approach. ISA transactions. 2014 Mar 31; 53(2): 317-26.
[15] Y.Li, X. Zhou, Observerbased state feedback Hinfinity control for networked control systems, TELKOMNIKA
Indonesian Journal of Electrical Engineering, 2014; 12(8), 6144-6152
[16] Sutha S, Thyagarajan T. Eigenstructure assignment based Multiobjective Dynamic State feedback controller design
for MIMO system using NSGA-II. In Modelling, Identification and Control (ICMIC), The 2010 International
Conference on, 2010 Jul 17; 870-875
[17] Chakraborty K, Dutta DK, Basak RD, Roy I. Temperature Control of Liquid Filled Tank System Using Advance
State Feedback Controller. Indonesian Journal of Electrical Engineering and Computer Science. 2015 May 1;
14(2): 288-92.
[18] Tiegna H, Amara Y, Barakat G. Overview of analytical models of permanent magnet electrical machines for
analysis and design purposes. Mathematics and Computers in Simulation. 2013 Apr 30; 90: 162-77.
[19] Li R, Yang M, Chu T. State feedback stabilization for Boolean control networks. IEEE Transactions on Automatic
Control. 2013 Jul; 58(7): 1853-7.
[20] Trégouët JF, Peaucelle D, Arzelier D, Ebihara Y. Periodic memory state-feedback controller: new formulation,
analysis, and design results. IEEE Transactions on Automatic Control. 2013 Aug; 58(8): 1986-2000.
[21] Flores JV, da Silva JM, Pereira LF, Sbarbaro DG. Repetitive control design for MIMO systems with saturating
actuators. IEEE Transactions on Automatic Control. 2012 Jan; 57(1): 192-8.
[22] Jafari J, Ghazal M, Nazemizadeh M. A LQR Optimal Method to Control the Position of an Overhead Crane. IAES
International Journal of Robotics and Automation. 2014 Dec 1; 3(4): 252.

1419
BIOGRAPHIES OF AUTHORS
Hossein Komijani did his master degree in Electrical Engineering, majoring in control systems
form Tabriz University, Tabriz, Iran. He is currently is a researcher in Intelligent systems in
Faculty of Electrical and Computer Engineering, University of Tabriz. His research interests are
Intelligent Control, Robotics, and Nonlinear Control.
Saeed Masoumi Kazraji is a PhD student in Power Electronic Engineering, in Tabriz University,
Tabriz, Iran. Hisresearch interests are Power Systems, Power Electronics, and Industrial
Electrical Motor Drives.
Ehsan Baneshiis an M.Sc student in Power Electronic Engineering in Islamic Azad University of
Neyriz, Neyriz, Iran. He is currently doingresearch in Power Systems, and Smart Buildings. His
research interests are Energy Optimization, and Photovoltaic Systems.
Milad Janghorban Lariche did his B.Scin Biomedical Engineering from Rajshahi Amirkabir
University of Technology Tehran Polytechnic, Iran. He is presently working as lecturer at
Abadan University Medical Sciences, Abadan, Iran. His research interests are Bioelectronic,
Renewable Energy, Industrial Motor Drive, and Automation

Modeling and State Feedback Controller Design of Tubular Linear Permanent Magnet Synchronous Motor

More Related Content

What's hot (20)

Similar to Modeling and State Feedback Controller Design of Tubular Linear Permanent Magnet Synchronous Motor (20)

More from IAES-IJPEDS (20)

Recently uploaded (20)

Modeling and State Feedback Controller Design of Tubular Linear Permanent Magnet Synchronous Motor