Proceedings of the Sixth International Conference on Research DOI: 10.

15439/2021R16
in Intelligent and Computing pp. 81–86 ISSN 2300-5963 ACSIS, Vol. 27

LQR Controller Design for Mini Motion Package

Electro-Hydraulic Actuator Control
NguyenVan Tan Huy Q. Tran*
School of Engineering – Technique Thu Dau Mot Faculty of Engineering – Technology Nguyen Tat
University Binh Duong, Vietnam Thanh University Ho Chi Minh City, Vietnam
[email protected] or 0000-0001-5929-6000 [email protected]

Phan Viet Hung Huynh Minh Phu

School of Engineering – Technique Thu Dau Mot School of Engineering – Technique Thu Dau Mot
University Binh Duong, Vietnam University Binh Duong, Vietnam
[email protected] [email protected]

Abstract—Electro-hydraulic actuators have been widely ap- signed based on the algebraic Riccati Equation with the qua-
plied in the industry because they have several major advan- dratic performance functional and the cost function. The
tages. In this paper, we focused on controlling the mini motion cost function corresponding to R and Q is equally important
package electro-hydraulics actuator. First, a mathematical
model of the electro-hydraulic actuator (EHA) was imple- for the control and the state variables as the output (piston
mented to apply the control process to the proposed system. position, piston velocity and piston acceleration). The con-
Second, we applied the linear quadratic regulator (LQR) con- troller can be tuned by varying the positive elements in the
troller to a linear model that is converted from the nonlinear Q, R matrix to achieve the desired response. A PID con-
EHA system. Finally, the numerical simulation results were troller is used for the numerical simulation process of the
performed in which the results obtained from the LQR con-
EHA system. The response signal in the LQR case is com-
troller were compared with the PID controller to show the su-
periority of the proposed solution. pared with the response signal in the PID case.
Index Terms—Linear control system; LQR controller; PID The important contributions of paper can be shortened as
controller, Electro-hydraulics actuator. follows:
 The mathematical model of the EHA system was
I. INTRODUCTION converted to the linear system for the application of
this controller
I N THE past decades, different types of electro-hydraulic
systems have been developed with different purposes for
industrial systems, such as heavy-duty equipment and some
 The gain K was calculated using Matlab software via
Q, R matrices, the algebraic Riccati Equation with
machining equipment. These systems have many great ad- the quadratic performance functional, and the cost
vantages such as fast response, high control precision and function
high output power based on position, force and pressure  A scale value factor N is also computed thanks to the
control signals. It is considered one of the most promising EHA’s parameters.
choices for modern industry and has become an indispens-
able control method in mechanical manufacturing, construc-  The successful numerical simulation results of the
tion machinery and defense. Moreover, the development of EAH system was respectively presented.
information technology has also greatly improved the qual-
ity of control process. II. EHA MODEL FORMULATION
To improve the performance of the electro-hydraulic actu-
The dynamics of EHA can be expressed the motion of the
ator (EHA) system, several EHA types were designed as
positioning piston via the mass of object Mp [4]:
shown in [1-3], and the use of conventional and advanced
controllers to improve control accuracy, including propor- m p p  
  F  F   S   S 
p sp frc 1 1 2 2 (1)
tional–integral–derivative (PID) controller [4-7], sliding
mode controller and adaptive back-stepping ccontrol [8,9],  and  are the mass, the position, the ve-
where m p ,  p ,  p , p
has been studied. Recently, many researchers have been fo-
cused on new approaches, such as fault-tolerant control and locity, and the acceleration of object Mp respectively.
fault-tolerant control compensation based on fault estimation S1 , 1 , and S ,  are the area, the pressure in two
2 2
using observer technology [4,10-12]. Research on linear
quadratic regulator (LQR) controller were also implemented chambers of the cylinder respectively.
for the EHA system [13-15]. Fsp , , Ffrc , and  are the external load force of the
In this paper, an LQR controller is applied to control the spring impact on the piston, the viscosity damping
linear EHA system. The gain K in the LQR controller is de-

* corresponding author

©PTI 2021 81
82 PROCEEDINGS OF THE RICE. THU DAU MOT, 2021

coefficient, the friction force, and unknown disturbance  xp 

2
 
respectively. 2e  Fbrk  FC  e
2
FC  xp   vst 
The spring force Fsp can be computed as:  tanh   
vcl  vcl  vst


2
 xp 
 Fsp   sp  
 
2 2ex 2p  Fbrk  FC  e
p  vst 
FC
  
where  sp is the stiffness of the spring. vst3 vcl
The friction force F can be presented as [4] Substitution equation (4), (5) into (7), we obtain
frc
as:
2
 vp 
  vp  vp 
Ffrc  2e  Fbrk  FC  e  vst 
 FC tanh     sp  S11  S22 
     pu
 vst  vcl    p
mp
p
mp
p
mp

where F , F , and B are breakaway friction, Coulomb

brk C v  
S  2
1 1  S222 


 
p
friction, and viscous friction coefficient, respectively. v p , mp mp
and vst are position velocity, Stribeck velocity threshold.  f  ,u
The hydraulic continuity equations for the EHA
system can be expressed as [4]: where

 
1  1 1i  S1 p     S1113i  S22 24i 

Equation (8) can be written as

 
 2  2  2i  S2 p   
 x  f ( x, u )  

  pump   p   where

x   x1T x2T x3T  , and x1  ; x2  ; x3 
where p p p

13i   pump  1v   3v   i Equation (9), can be rewritten as

 24i   pump   2v   4v   i
e e  
1  , and    
V
01  S1 p 
2
V02  S2 p   x1  
x   
x2 
x3 
e , i ,V01 , and V02 are the effective bulk modulus in  2  
each chamber, the internal leakage flow rate of the  x3     sp 
  m p  m  1 p u   2 p   
m p 
p
cylinder, the initial total control volumes of the first and  p p
the second chamber respectively [4].  f1 ( x, u ) 
1v , 2v , 3v , 4v , and  pump are the flow rate   f 2 ( x, u ) 
through the pilot operated check valve on the left, on the  f 3 ( x, u ) 
right, the flow rate through the pressure relief valve on the
left, on the right, and the pump flow rate respectively.  p ,
and  are the displacement and the speed of the servo
1 
 S11  S22 
; 2 
S  2
1 1  S222 

pump [4] mp mp
From equation (1) to (6) and the derivative both
sides of equation (1), we have: The appropriate equation written in a state-space
format of EHA system based on a Taylor expansion is
mp    sp  S11  S2  2 given as:
p p p


where 
x(t )  Ax(t )  Bu(t )  

where A, B matrices are calculated as:

 NGUYEN VAN TAN ET AL.: LQR CONTROLLER DESIGN FOR MINI MOTION PACKAGE ELECTRO-HYDRAULIC ACTUATOR CONTROL 83

 a11 a12 a13   b1  f 3 ( x, u )

a32 
A   a21 a22 a23  ; B  b2  ; u  
  x2
    x2 
2

 a31 a32 a33   b3   2 Fc tanh  x2   tanh    1  2
  vcl   cl    x2 
  x  6 2e e  vst   Fbrk  Fc 
v
where the parameters of the EHA system are provided  3   x2
 vcl 2 
3
 vst 
in Table 1  
 
 
Table 1. Basic parameters of the EHA system. x 
  2 
2

 Fbrk  Fc  3 S1  e S2 e
 4 2ee  x2   2  
 vst 

 v 5
 V  S x V02  S 2 x1
x1  0
Parameters Symbols Values Units st 01 1 1 x2  0
u 0
Piston cross-sectional S1 0.0013 m2
f 3 ( x, u )
area on the left a33 
Piston cross-sectional S2 9.4 10 -4
m 2 x3
area on the right  x 
2

The initial total control V01 2.09 10-4 m3 FC  tanh  2   1
 
2
x 
  2
Fbrk  FC  vcl 
   

volumes on the left
 2e e  vst 
The initial total control V02 4.006 10- m3 vst vcl
1
5
volumes on the right
x1  0
Mass of the control mp 10 kg x2  0
u0
object
e 2.9 108
2
Bulk Modulus Pa x 
  2   Fbrk  FC  2
 (2 2e e  vst 
  x2 x1  0
The stiffness of the  sp 2383 Nm
 vst
3
 x2  0
spring u0

The displacement of p 3.5 10-6 m3

III. LQR CONTROLLER DESIGN
the servo pump
LQR is a powerful method for designing controllers
The elements of matrix B is computed as: based on the optimal algorithm that minimizes a given
cost function. The parameter of the cost function is
f1 ( x, u ) f ( x, u ) determined by two matrices Q and R which is calculated
b1  b2  2
 u  u  via the algebraic Riccati Equation with the quadratic
f ( x, u ) S1  e
2
S2 e
2
performance functional and the cost function (12):
b3  3  
u V01  S1 x1 V02  S 2 x1

J (K ) 
1  T
2 0
 x (t )Qx t   uT (t )Ru(t ) dt

The elements of matrix A is computed as:
The optimal control signal is represented as
f1 ( x, u ) f1 ( x, u ) f ( x, u )
a11  a12   1 a13  1 0 u* (t )   Kx(t )
x1 x2 x3 
= 0; ;
Gain K is described as (14)
f ( x, u ) f ( x, u ) f ( x, u )
a21  2  0 a12  2  0 a13  2 1
x1 x2 x3 K  R 1 BT P 
; ;

f 3 ( x, u ) where P is determined by the algebraic Riccati Equation

a31  (15):
x1 x1  0
x2  0
u 0

   PA  AT P  Q  PBR 1 BT P  0  
S12  e S22  e
   x2
 V  S x  V  S x 2
2

 01 1 1 02 2 1  x1  0
x2  0 The state-space feedback Equation (16) can be built
u 0
as shown in Fig. 1.
 S13  e S23  e 
  *u
 V  S x 2 V  S x 2  x1  0  x  t   Ax  t   Bu (t )
 01 1 1 02 2 1  x2  0 
u 0
 y  t   Cx  t   Du  t   

We choose matrices C, and D is shown as:

84 PROCEEDINGS OF THE RICE. THU DAU MOT, 2021

C  I 33 ; D  0  response of piston position and velocity position are

shown in Fig. 2.

Q  I33 ; and R = 1 and gain: K1 =[1.000008 1.5746

-3
10 0.1]
Then, we continued obtain other simulations for
cases 2, 3, and 4 in Table 2. These cases correspond to
figures 3, 4, 5, respectively.


Fig. 1. Scheme diagram of the EHA system in state-space

Finally, matrices Q and R can choose as:

 Q  C T C  I 33 ; R  1 
 

IV. SIMULATION RESULTS

The obtained results from the simulation process of
the EHA system with parameters (Table 1) are shown in
Table 2 to verify the effectiveness of the suggested
techniques. The numerical simulation processes are
carried out and then the performance comparison between
LQR and PID controllers with the input u(t) Fig. 2. Position and velocity response with LQR controller (K1)
=0.2*ones(size(t)) was drawn.
First, simulation of the EHA system is done by the
LQR controller with chosen matrices Q and R. Step

Table 2. Experimental results.

Cases K, Q, R, N Status Setting time (s)

Q  I33 ; R=1, N=1
1 Unstable None
K1 =1.0000008; 1.5746 10-3 ; 0.1
Q  100000.I 33 ; R=1, N=1
2 Stable 18
K2 =316.2278 146.2361 316.2275

100000 0 0 
Q   0 0.001 0 
 0 0 0.005
3 Stable 1.35

3
R = 0.005; N = 4.472 10
K3 = 4.472 103 ; 4.683; 0.9997527
100000 0 0 
Q   0 0.005 0 
 0 0 0.005
4 Stable 0.25
4
R = 0.0001; N = 3.16 10
K4 = 3.162 104 ; 211.316; 7.0708
 NGUYEN VAN TAN ET AL.: LQR CONTROLLER DESIGN FOR MINI MOTION PACKAGE ELECTRO-HYDRAULIC ACTUATOR CONTROL 85


Fig. 6. Scheme diagram of the EHA system in state-space

The simulation results in the fourth case shown in

Fig. 8 which is also compared with position response for
case without change (Kp.= 27.52272; KI = 148.0119; KD
= 0.2) in the PID controller.
Fig. 3. Position and velocity response with LQR controller (K2)

Fig. 7. Position and velocity response with LQR controller (K3)

Fig. 4. Position and velocity response with LQR controller (K3)

Fig. 8 Position and velocity response with LQR and PID

controller (K4)
In summary, the simulation results in the first case
shown unstable position and velocity responses as shown
in Fig. 2. In the second case (Fig. 3), the response is more
stable when the Q value increases. The third and fourth
Fig. 5. Position and velocity response with LQR controller (K4)
cases showed the great effect of the Q and R parameters
The scheme diagram of the EHA model in Simulink through the obtained response signal as illustrated in Fig.
with N factor using the LQR controller can be described 4 and Fig. 5. We also see this influence in Fig.7 and Fig.8.
in Fig. 6. The simulation result in the third case is shown These figures shown the superiority of the proposed
in Fig. 7 which is compared with position response for controller thanks to achieved position responses.
case without change (Kp.= 27.52272; KI = 148.0119; KD V. CONCLUSIONS
= 0.2) in the PID controller.
The numerical simulation when applying the LQR
controller show that the feedback signal of piston position
in the EHA system achieved the expected results
compared to the PID controller. The overshoot of the PID
86 PROCEEDINGS OF THE RICE. THU DAU MOT, 2021

controller is greatly reduced and the response tim e of the [7] Skarpetis .M. G, and Koumboulis. F. N, "Robust PID controller for
electro — Hydraulic actuators," 2013 IEEE 18th Conference on
feedback signal is also improved as shown in the previous Emerging Technologies & Factory Automation (ETFA), Cagliari,
section. However, to achieve the desired control signal when Italy, (2013), pp. 1-5, doi: 10.1109/ETFA.2013.6648165.
using an LQR controller, it is important to be experienced in [8] Ahn, K.K.; Nam, C.N.D.; Jin, M. Adaptive Back-stepping Control of
the selection of parameters Q, R and parameter N. an Electrohydraulic Actuator. IEEE/ASME Trans. Mechatron. (2014),
19, 987–995.
ACKNOWLEDGMENT [9] Thomas. A. T, Parameshwaran. R, and et al “Improved Position
Tracking Performance of Electro-Hydraulic Actuator Using PID and
This research was supported by Research Foundation Sliding Mode Controller”, IETE Journal of Research, (2019),
funded by Thu Dau Mot University. DOI: 10.1080/03772063.2019.1664341
[10] Ahian, S.A., Truong, D.Q., Chowdhury, P. et al. Modeling and fault-
REFERENCES tolerant control of an electro-hydraulic actuator. Int. J. Precis. Eng.
[1] Gabriele . A, Andrea . V, “A design solution for efficient and compact Manuf. 17, 1285–1297 (2016). https://doi.org/10.1007/s12541-016-
electro-hydraulic actuators”, ScienceDirect, Procedia Engineering 106 0153-2
(2015) 8 – 16 [11] Mahulkar .V, Adams D. E, and Derriso .M, "Adaptive fault-tolerant
[2] Juliang . X , Qinyu . L, and et al “Theoretical and Experimental control for hydraulic actuators," 2015 American Control Conference
Analysis of the Hydraulic Actuator Used in the Active Reflector (ACC), Chicago, IL, USA, (2015), pp. 2242-2247, doi:
System” Hindawi Mathematical Problems in Engineering. Volume 10.1109/ACC.2015.7171066.
2018, Article ID 8503628, 13 pages, doi.org/10.1155/2018/8503628 [12] Phan .V. D, Vo .C. P., Dao .H. V. and Ahn .K. K., "Robust Fault-
[3] Shaoyang . Q, David . F, and et al “A Closed Circuit Electro- Tolerant Control of an Electro-Hydraulic Actuator With a Novel
Hydraulics Actuator with Energy Recuperation Capability”, Novel Nonlinear Unknown Input Observer," in IEEE Access, vol. 9, pp.
system solutions, 12th International Fluid Power Conference, 30750-30760, 2021, doi: 10.1109/ACCESS.2021.3059947.
Dresden (2020) [13] Pourebrahim .M, Ghafari ,A. S, and Pourebrahim .M, "Designing a
[4] Tan, N. V.; and Cheolkeun, H.; “Experimental Study of Sensor Fault- LQR controller for an electro-hydraulic-actuated-clutch model," 2016
Tolerant Control for an Electro-Hydraulic Actuator Based on a Robust 2nd International Conference on Control Science and Systems
Nonlinear Observer”. Energies, MDPI, Open Access Journal, Energies Engineering (ICCSSE), Singapore, (2016), pp. 82-87, doi:
(2019), 12, 4337; doi:10.3390/en12224337 10.1109/CCSSE.2016.7784358.
[5] Tan, V. N.; and Cheolkuen, H.; “The Actuator and Sensor Fault [14] Tan .V .V, Olivier .S, Luc .D, Enhancing roll stability of heavy
Estimation Using Robust Observer Based Reconstruction for Mini vehicle by LQR active anti-roll bar control using electronic servo-
Motion Package Electro-Hydraulic Actuator”. Intelligent Computing valve hydraulic actuators. Vehicle System
Methodologies, Proceeding of 15th International Conference, ICIC Dynamics, Taylor & Francis, (2017), 55 (9), pp.1405-1429. Doi:
2019. Nanchang, China, August 3–6. Part III, Pp. 244-256. (2019). 10.1080/00423114.2017.1317822.
[6] Norlela . I, Mazidah . T, and et al, “PID Studies on Position Tracking [15] Priyanka .E, Maheswari .C, Thangavel .S. “Remote monitoring and
Control of an Electro-Hydraulic Actuator”, International Journal of control of LQR-PI controller parameters for an oil pipeline transport
Control Science and Engineering 2012, 2(5): 120-126, DOI: system”. Proceedings of the Institution of Mechanical eers, Part I:
10.5923/j.control.20120205.04 Journal of Systems and Control Engineering. (2019); 233(6):597-608.
doi:10.1177/0959651818803183