Al-Baldawi: A Study on the Effects of Servovalve Lap on the Performance of a Closed…..

A Study on the Effects of Servovalve Lap on the Performance of a

Closed - Loop Electrohydraulic Position Control System
Rafa A. H. AL-Baldawi Yahya A. Faraj Rawand E. J. Talabani
Al-Mustansiriya University-College of Eng. University of Kirkuk \ College of Eng.
Mechanical Eng. Department. Petroleum Eng. Department.

Abstract
This paper deals with a closed–loop position control of a double acting and double–rod
actuator using an electrohydraulic servovalve (EHSV). This system is studied by using
symmetric critical center spool valve (zerolapped) and open center spool valve (underlapped).
The nonlinear dynamic behavior of each case is undertaken and simulated. The system is
modeled by using five state variables (piston position, piston velocity, actuator pressures, and
servovalve spool displacement) and is tested under different step inputs. The EHSV is modeled
with a first order differential equation. The closed-loop system stability is investigated by
introducing equilibrium state into Jacobian matrix and determining the eignvalues. Viscous
friction and compressibility of oil are included in the modeling of the system. Because the
electrohydraulic position servo system is not very sensitive to coulomb friction and piston
leakage they are neglected. The work showed that when the underlapped servovalve operates in
the underlap region, the hydraulic position control system has more stable operation and better
transient responses.
Keywords: Zerolap, Underlap, EHSV, Steady-State Characteristics, Dynamic Response, Position
Control, Modeling, Simulation.

– – – –

(Lap)
.
. (underlap) (zerolap)
, ) (State variable) .
. (
.
. Jacobian
.
(underlap)
.
Received 14 August 2008 Accepted 31 Dec. 2008
1
 Al-Rafidain Engineering Vol.17 No.5 October 2009
1. Introduction:
Fluid power control, that is the transmission and control of energy by means of a
pressurized fluid, is an old and well-recognized discipline. The growth of fluid power has
accelerated with our desires to control ever increasing quantities of power and mass with higher
speeds and greater precision. More specifically, where precise motion control is desired and
space and weight are limited, the convenience of small size-to-high power ratios and the ability to
apply very large forces and torques with fast response times, at the same time, achieve a high
degree of both accuracy and performance to make the hydraulic servomechanism systems the
ideal control elements. The demand to achieve more accurate and faster control at high power
levels has produced an ideal marriage of hydraulic servomechanisms with electronic signal
processing. Information could be transduced, generated, and processed more easily in the
electronic medium than as pure mechanical or fluid signals, while the delivery of power at high
speeds could be accomplished best by the hydraulic servo. This sophistication of electronic
devices and hydraulic devices into electrohydraulic servomechanisms rendered better hydraulic
systems, more efficient, more reliable, and faster equipments than ever before.
The key element in this family of mechanisms is the electrohydraulic servovalve. With a great
power gains, the servovalve acts as a power amplifier that converts a low-power electrical signal
into high-power hydraulic signal. These characteristics of electrohydraulic servo system make it
very attractive for many applications, such as the control of industrial robots, processing of
plastics, aircraft, satellites, launch vehicles, flight simulations, turbine control, and numerous
military applications[1]. Because of the nonlinearity and uncertainty parameters in hydraulic
systems (such as nonlinear flow/pressure characteristics, friction forces, flow forces and their
effects on the spool position and unknown external disturbances) many researchers provid a
nonlinear model of the hydraulic servo system control. A nonlinear model of electrohydraulic
velocity servo system is introduced by Jovanoic [2]. In this paper; the flow nonlinearities, internal
friction, oil compressibility, and valve dynamic (as first order transfer function) are presented.
Lyapunov–based design is used to develop a nonlinear controller. He shows a good agreement
between the analytical technique and experimental results. The work of Yun and Cho [3] has
considered unknown load disturbance as parameter uncertainty and designed a Lyapunove–based
controller to make the hydraulic system follows a given second order linear model. Rui [4] also
develops a nonlinear model for single–stage electrohydraulic servovalve and produced a
nonlinear controller based on backstepping approach. His work included the effects of frictions
(dry, Coulomb, and viscous), valve dynamic, and oil compressibility. Servovalve dynamics play
an important part on system behavior over a certain range of frequency response. Considerable
efforts had been gone into modeling the electrohydraulic servovalve, which suggested that,
depending on the frequency range of interest, a servovalve is best modeled by a first or second
order transfer function [5,6]. Mookherjee [7] has used a computer-aided design and sensitivity
analyses to study the effects of radial clearance, mismatch in the areas of the tractive air–gaps,
and port geometry on the valve performance in a single–stage servovalve. His work included a
nonlinear field modeling of hydraulic fluid in the spool valve and magnetic flux in the motor,
showed qualitative conformity with the results presented in Moog technical Bulletin.
This work is concerned with studying the performance of a hydraulic servo position control
system using zerolapped and underlapped spool servovalve. The work aims at studying the
steady-state characteristics of the EHSV (zerolapped and underlapped) such as coefficients of
valve flow gain, flow-pressure , and pressure sensitivity, determining and simulating a
mathematical model for an electrohydraulic position servo control system, and finding out the
influences of servovalve lap on the performance of the hydraulic position servomechanism.

2
 Al-Baldawi: A Study on the Effects of Servovalve Lap on the Performance of a Closed…..

2. Dynamic Models of Electrohydraulic System:

Figure (1) shows a schematic of the actuator\servovalve. As detailed by Merritt[8], the two-
stage electrohydraulic servovalve is comprised of a coil wrapped armature connected to a spool
by a spring used for force feedback. The spool acts as a control valve, that regulates flow into the
hydraulic actuator, which contains the two-actuator champers and a piston. The above control
system can be modeled as follows:

I. Servovalve Flow Equations:

The flow through the servovalve, shown in figure (2a), can be compared to the flow of a fluid
through a constricted point[9], as in figure (2b). The equation, which describes the flow through
the valve, can be derived from Bernoulli’s equation and the conservation of mass of the fluid as it
moves through the constricted orifice[9]. The volumetric flow rate Q through an orifice is given
by:
2
Q = C d A0 P1 P2 (1)
Where Cd is the discharge coefficient of the orifice, A0 is the area of the orifice, is the density of
the fluid, and P1 and P2 are the respective pressure on either side of the orifice. The area A0 varies
linearly with the position of the spool, in case of critical center spool valve, the orifice area is
only a function of spool position, A 0 = A 0 (xv) where xv is the valve spool displacement while
in open center valve, the orifice area is a function of spool valve displacement and valve
underlap, A 0 = A 0 (xv, U), where U is the valve underlap.

Electrohydraulic
servovalve
xv
Ps Pr Ps

Hydraulic
Actuator x

Chamber 2 Chamber 1

Ps – supply pressure
Pr– return pressure Spoo

Pist
Figure (1): Detail of actuator/servovalve[9]
on

-b- -a-
Figure (2): Flow through an office[9]

3
 Al-Rafidain Engineering Vol.17 No.5 October 2009
For the critical center spool valve, the flow equations are, as shown in figure (3a), (neglecting the
leakage flow rate):

2 PS P1 xv 0
Q1 Cd w x v P1 , P1 ,
P1 xv 0
... (2)
2 P2 xv 0
Q2 Cd w x v P2 , P2
PS P2 xv 0
where w , which is called the area gradient of the valve, is the rate of change of orifice area with
stroke (i.e. A = w xv). For open center spool valve the flow equations, as shown in figure (3b), are
(for underlap region only):

4
 Al-Baldawi: A Study on the Effects of Servovalve Lap on the Performance of a Closed…..

And for xv 0:
2 2
Q1 Cd w U xv PS P1 U xv P1 ,

2 2
Q2 Cd w U xv P2 U xv PS P2 ,

And for xv 0: ... (3)

2 2
Q1 Cd w U xv P1 U xv PS P1 ,

2 2
Q2 Cd w U xv PS P2 U xv P2 .

If a matched and a symmetrical spool valve is used , P1 = (Ps + PL) / 2 and P2 = (Ps - PL) / 2
can be used to write a general flow-pressure equations for both servovales in term of the supply
pressure (Ps) and the load pressure (PL)[5] as follows:
Zerolapped Servovalve:

Q1 Q2 QL Cd w x v 1 xv …(4)
PS PL
xv
Underlapped Servovalve:
Q1 = Q2 =QL = C d w (U 1 1 …(5)
xv ) PS PL (U xv ) PS PL

II. Servovalve Coefficients:

A general expression for the load flow is:
QL Q L x v , PL …(6)
If this function is expressed as a Taylor’s series about a particular operating point QL = QL1
and when the higher order infinitesimals is neglected, it can be written as:
QL QL
QL xv PL …(7)
xv 1
PL 1
The partial derivatives required are obtained by differentiation of the pressure-flow equation.
These partials define the two most important parameters of a valve.The flow gain is defined by:
QL
Kg …(8)
xv
and the flow-pressure coefficient Ks is defined by:
QL
Kc …(9)
PL
Another useful quantity is the pressure sensitivity Kc defined by:
5
 Al-Rafidain Engineering Vol.17 No.5 October 2009

PL
Kt …(10)
xv
which is related to the other quantity by the well-known relation from calculus:

PL QL xv Kg
or Kt …(11)
xv QL PL Kc

The coefficients Kg, Kc and Kt are called valve coefficients, that are extremely important in
determining stability, frequency response, and other dynamic characteristics. The flow gain
directly affects the loop gain constant in a system, therefore, it has a direct influence on system
stability. The flow-pressure coefficient directly affects the damping ratio of valve- motor
combinations, while the pressure sensitivity coefficient of valves is quite large, which accounts
for the ability of valve-motor combinations to break away large friction loads with little error [8].
The values of valve coefficients vary with operating point. The most important operating point
is the origin of the pressure-flow curves (i.e. QL = PL = xv = 0), as in figure 4, because system
operation usually occurs near this region and the valve coefficients evaluated at this operating
point are called the null valve coefficients [8].
Critical center spool valve coefficients
The valve coefficients of an ideal zerolapped spool valve (leakage flow is zero) can be
obtained by differentiation of equation. (4).
The flow gain is
1
Kg Cd w PS PL …(12)

xv

a- For zerolapped
servovalve

6
 Al-Baldawi: A Study on the Effects of Servovalve Lap on the Performance of a Closed…..

xv

b- For underlapped
servovalve

Figure (4): Servovalve pressure-flow characteristics for a supply pressure of 1000 pci

The flow pressure coefficient is :

1
Cd wxv PS PL
Kc …(13)
2 PS PL
and the pressure sensitivity is
2 PS PL
Kt …(14)
xv
The null valve coefficients are:

PS
Kg0 Cd w …(15)

K c0 0 …(16)

Kt 0 …(17)

Open center spool valve coefficients

By differentiating esquation (5), the valve coefficients can be obtained for underlapped valve
in the underlap region, and by evaluating the derivatives at QL = PL = xv = 0, the null valve
coefficients are:

7
 Al-Rafidain Engineering Vol.17 No.5 October 2009

PS
Kg0 2Cd w …(18)

Cd wU P S
K c0 …(19)
PS
2 PS
Kt 0 …(20)
U
III. Flow Into and Out of the Actuator:
The expression of the net flow into and out of the actuator consists of three terms:
1- Flow due to piston movement A dx/dt,
2- Flow due to laminar leakage across the piston, CLPL
3- Flow due to compressibility of the working fluid.
Thus the flow rate Q into the driving side of the actuator is:
( v 01 A1 x ) dP1 dx
Q1 A1 C L ( P`1 P2 ) ...(21)
dt dt
And the flow rate out of the actuator is:
( v 02 A2 x ) dP2 dx
Q2 A2 C L ( P1 P2 ) ...(22)
dt dt
where v01 and v02 are the initial volume of the driving and other sides of the linear actuator, A1 =
A2 = Ap, Ap is surface area of piton head. is the bulk modulus of the fluid, and CL is leakage
coefficient across the actuator head.
IV. Load Dynamic Equation:
Many researchers have written the load equation in different forms. The most general form of the
equation is[1, 9]:
d 2x dx dx
P1 P2 * AP M 2 F fc K S x Fd ...(23)
dt dt dt
where is viscous friction constant, Ffc is modeled coulomb friction force, KS is load spring
constant, M is the actuator and load mass, and Fd is the external disturbance.

V. Electrohydraulic Servovalve Equation:

It is often convenient in servo analysis or in system synthesis work to represent an
electrohydraulic servovalve by a simplified, equivalent transfer function[10]. If the effects of
hysteresis and flow forces on the servovalve are neglected, then the dynamic behavior of the
servovalve can be described by a first-order approximation[4], as follows:
1 Kv
xv xv u ...(24)

where xv is valve spool displacement, is time constant of the servovalve, Kv is servovalve gain,
and u is input current.

8
 Al-Baldawi: A Study on the Effects of Servovalve Lap on the Performance of a Closed…..

VI. Error Signal Equation:

It can be seen from the block diagram of the control system, figure (5), that the summing
amplifier differentially compares the input, and feedback voltage, which is obtained from the
input and feedback potentiometers respectively; the resulting signal being fed to the amplifier.
The output from the amplifier is used to move the spool of the electrohydraulic servovalve and
subsequently move the load to its required position. If a potentiometer (input and feedback) gain
(Kp) and a servo amplifier gain (Ga) are used, then the error current signal equation can be written
as::
u = Kp · Ga · (xdesired x) ...(25)
In order to close the system model, it becomes necessary to replace the system input u with
equation (25); therefore, equation (24) is rewritten as:
1 Kv
xv xv K p .Ga xdesired x ...(26)

x desired u xv
KP + Amplifier Servovalve Cylinder and the load x
Ga
-
x

KP

Figure (5): EHSV Position control System

2.1 Derivation of the State Equations:

A number of assumptions are made in order to simplify the dynamic model of the position
control system. They are as follows:
1-The effect of Coulomb friction is neglected and it is assumed that there are no external
disturbance and spring force.
2-The servovalve has a symmetrical spool, and the control is dominated to be within underlap
region, i.e. –U < xv< U for underlapped spool.
3- The piston is located in the center of the actuator so that v01 = v02 = vt / 2 = v0 and the
piston leakage is neglected.
4- The hydraulic pump delivers a constant supply pressure.
Five state variables are used to describe the system operation. The first four state variables
describe the operation of the hydraulic actuator: x1- the position of the piston (x), x2 - the velocity
of the piston (dx/dt), and x3 & x4- the pressures in the actuator chambers. The final state variable is
from the servovalve: x5 - the position of the control valve spool (xv). On the basis of the preceding
assumptions and the state variables definitions, the equations of the closed-loop control system
can be written as:

9
 Al-Rafidain Engineering Vol.17 No.5 October 2009
x1 x2
x2 x3 x 4 .A p .x 2 m

x3 Q 1 A p .x 2
(v0 A p x1 )

x4 Q 2 A p .x 2 ...(27)
(v0 A p x1 )
1 1
x5 x5 K v . Ga . K p . x desired x .
Where equations. (2 & 3) are used for Q1 & Q2.

2.2 System Stability:

It is expected that the stability criteria of the linear systems could be applied to nonlinear
systems, if the deviations from the equilibrium state are sufficiently small and the signals
involved are small, therefore, the nonlinearity has only a minor effect. Thus a linearization
principle can be used to determine the stability of nonlinear systems [11,12].
The equilibrium state xe of the nonlinear unforced, u = 0, autonomous system:
x f ( xe ) ...(28)
is asymptotically stable if the eigenvalues of the matrix:

f x
A ...(29)
x x xe

has negative real parts. It is unstable if at least one eigenvalue of A has a positive real part. It is
completely unstable if all eigenvalues of A have positive real parts[12]. [ f(x) x] denotes the
Jacobian matrix and equation (29) is called the linearized system of equation (28) about the
equilibrium state xe. The equilibrium state of the current hydraulic position servo control system
described by the equation (10) (for both zerolapped & underlapped), for zero input, is: (xe = x1 =
x2 = x5 = 0) &( x3 = x4 = PS / 2)

3. Simulation Results:
3.1 System Response:

The mass density of oil (the transmission medium used in hydraulic components) is a function
of both pressure and temperature = f (P,T). When one assumes a constant temperature and
neglects the change in mass density ( ) due to change in pressure, a constant value of (0.000078
Ib-sec 2/in4) can be considered for oil density. The data that had been used in the simulation are
shown in tables (1) and (2). The displacement step responses obtained from the simulation for
both critical and open center spool servovalves are shown in figure (4) with the step input signals
of 0.1in. & 0.3in. amplitudes. The pressure responses for both sides of the driving cylinder are
shown in figure (5) for zerolapped together with underlapped servovalve and the velocity
responses for the actuator are shown in figures (6).

10
Al-Baldawi: A Study on the Effects of Servovalve Lap on the Performance of a Closed…..

Table (1): HUM058-OBE Schneider two-stage servovalve parameters

Symbol Notation Value Unit
Rated flow at P 1018.3 pci Qrate 21.2 Liter/min.
Operating pressure ---- 145.74-3055 pci
Null leakage flow at P 1018.3 pci Qnl 0.742 Liter/min.
Rated input current ---- 20; 20 mA.
Servovalve gain Kv 0.955 in./A.
Area gradient w 0.513 in.2/in.
Typical spool travel xrate 0.0191 in.
Valve time constant 0.01 s

Table (2): System parameters

Symbol Notation Value Unit
Piston head inside diameter Di 2.5 in.
Piston rod outside diameter Dr 1.375 in.
Surface area of piston Ap 3.4238 in.2
Piston stroke Lt 6 in.
Current amplifier gain Ga 4 mA./V
Mass of piston and load M 1.1775 lb.s2/in.
Viscous friction coefficient 100 lb.s/in.
Supply pressure Ps 1000 pci
Mass density of hydraulic fluid 0.000078 lb.s2 /in 4
Effective bulk modulus 200,000 lb/in2.
Input and feedback (LVDT) gain Kp 5.25 V/in.

* * * * * * * *
*
* * * * * * * * * *
* *
* *
* *
* * *
* * *
* *
a- For 0.1 in. step Input.
F1111igure (6): Position Transient Response
a- For 0.1 in. step input. b- For 0.3 in. step input.
Figure (6): Position Transient Response

11
Al-Rafidain Engineering Vol.17 No.5 October 2009

12
 Al-Baldawi: A Study on the Effects of Servovalve Lap on the Performance of a Closed…..

3.2 System Stability :

According to the theorem presented in section 2.2, the eigenvalues of the mathematical model,
(equation.27) around the equilibrium state xe are shown in figure (9) and below.
For zerolapped servovalve system For underlapped servovalve system
1.47e001 5.84e 044
8.45e001 7.07e000
2.66e002 9.30e001
1.76e002 6.15e002i 4.25e001 6.21e002i
1.76e002 6.15e002i 4.25e001 6.21e002i

For zerolapped system For underlapped system

Figure (9): Eigenvalues of closed-loop electrohydraulic position control system.

4. Discussion:
x U
As the under lapped servo valve works in the under lap region (i.e. v ), the position
control system has dynamic results better than the same system with zero lapped servo valve. The
servo valve flow gain has a direct influence on the system stability[8] and because the under
lapped servo valve has a flow gain that is twice that of zero lapped servo valve, (eqs. 15 & 18 ),
in the under lap region, therefore, the system with under lapped servo valve is relatively more
stable as it can be noted in figure (9).The servo valve flow-pressure coefficient directly affects
the damping ratio of the hydraulic system[8]; the under lapped servo valve has a greater flow-
pressure coefficient than the zero lapped servo valve, (eqs. 16 & 19 ), therefore, the control
system with under lapped servo valve has a damping ratio less than the control system with zero
lapped servo valve (if the system is considered a linear system), so the position response of the
control system, with under lapped servo valve, reaches the steady state faster than the system
with zero lapped servo valve as shown in the figure (6) which also shows that the position
control system (of both servo valves) has a second order behavior for position response to a step
input with damping ratio > 1. It can be seen in figures (7) that both sides steady-state cylinder’s
pressure are the same and equal to 0.5 of PS for both spool valves. It can also be seen that the

13
 Al-Rafidain Engineering Vol.17 No.5 October 2009
pressure responses of the system with under lapped servo valve are faster than the system with
zero lapped servo valve (which is more oscillatory).

5. Conclusion:
The effects of servovalve lap on the performance of a closed - loop electrohydraulic position
control system are studied. The related equations of the control system are derived and simulated
using zerolapped and underlapped servovalve. Servovalve null coefficients are also determined.
The study showed that when the underlapped servovalve operates in the underlap region, the
position control system has more stable operation and better transient responses than the position
control system with zerolapped servovalve and the control system has a second order behavior
for displacement transient response to step input with >1.

5. References:
[1]. Vilenius, M. J.“The Application of Sensitivity Analysis to Electrohydraulic Position
Control Servos” ASME Journal of Dynamic Systems, Measurements, and control, Vol.
105, No. 2, pp. 77-82, June 1983.
[2]. Jovanoic M. “Nonlinear Control of an Electrohydraulic Velocity Servosystem”, University
of California, 2001.
[3]. Yun, J. S., and Cho H. S., “Application of an Adaptive Model Following Control
Technique to a Hydraulic Servo System Subjected to Unknown Disturbances” ASME
Journal of Dynamic Systems, Measurement, and Control, Vol. 113, No. 3, pp.479-486,
September 1991.
[4]. Rui L. “Nonlinear Control of Electro-hydraulic Servosystems: Theory and
Experiment”, Master degree thesis, University of Illinois, 1998.
[5] .Al-Baldawi R.A.H. “Optimum Performance of Electrohydraulic Position Control
Systems Using an Underlapped Servovalve”, Ph.D. Thesis, University of Wales College
of Cardiff. October 1990.
[6]. Martin, D.J. and Burrows, C.R. “The Dynamic Characteristics of an Electro-hydraulic
Servovalve” Trans of ASME, Journal of Dynamic System, Measurement and Control
December 1976, pp 395.
[7]. Mookherjee S. “Design and Sensitivity Analysis of a Single-stage Electro-hydraulic
Servovalve” Proc. of First FPNI-PhD Symposium Hamburg 2000, pp. 71-88.
[8]. Merritt, H. E. “Hydraulic control systems” John wiley and Sons, Inc., 1967.
[9]. Manhartsgruber B. “Application of Singular Perturbation Theory to Hydraulic Servo
Drives-System Analysis and Control Design”, Department of mechanics & Machine
Design, Johannes Kepler University.1998.
[10].Thayer W. J. “Transfer function for Moog servovalves” Technical Bulletin 103, Moog
servocontrols Inc. New York, December 1958, Rev. January 1965.
[11]. Ogata K. “Modern Control Engineering” 4th Edition Prentice-Hall, Inc., 2001.
[12]. Willems J. L. “Stability Theory of Dynamical Systems” Thomas Nelson and Sons Lid.
1970.
The work was carried out at the college of Engg. University of Mosul

14