Al-Khwarizmi

Engineering
Journal
Al-Khwarizmi Engineering Journal, Vol. 9, No. 3, P.P. 80- 99 (2013)


Pressure Control of Electro-Hydraulic Servovalve and Transmission
Line Effect

Jafar Mehdi Hassan* Yiqin Xue** Majid Ahmed Oleiwi***
Ahmed Fouad Mahdi****
*Department of Mechanical Engineering / University of Technology
**Institute of Mechanical Engineering / Cardiff School of Engineering /UK
*** Department of Control and System Engineering / University of Technology
**** Department of Automated Manufacturing Engineering / Al-Khawarizmi College of Engineering
*Emal: JafarMahdi1951@yahoo.com
**Emal: xue@cf.ac.uk
***Emal: moleiwi@yahoo.com
****Emal: afmkridi@hotmail.com

(Received 17 March 2013; accepted 31 October 2013)


Abstract

The effected of the long transmission line (TL) between the actuator and the hydraulic control valve sometimes
essentials. The study is concerned with modeling the TL which carries the oil from the electro-hydraulic servovalve to
the actuator. The pressure value inside the TL has been controlled by the electro-hydraulic servovalve as a voltage
supplied to the servovalve amplifier. The flow rate through the TL has been simulated by using the lumped element
electrical analogy method for laminar flow. The control voltage supplied to servovalve can be achieved by the direct
using of the voltage function generator or indirect C
++
program connected to the DAP-view program built in the DAP-
card data acquisition connected to PC, to control the value of pressure in a selected point in the TL. It has been found
that the relation between the voltage value and the output flow rate from the servovalve in most of the path is a linear
relation. The MATLAB m-File program is used to create a representation state of the mathematical model to find a
good simulation for the experimental open loop control test.

Keywords: Electro-hydraulic servovalve, Twin flapper nozzle, Transmission Line effect, Open loop Pressure control,
Voltage linearization Equation, DAP-view Data Acquisition.


1. Introduction

Servovalves were developed to facilitate the
adjustment of fluid flow based on changes in load
motion. The twin nozzle flapper servo valve is a
high quality parts combined from mechanical,
electrical and hydraulic technology and has the
advantages of large power ratio, fast response, and
high level of control precision. It is widely used in
industrial applications such as manufacturing
systems, robotics, materials test machines, active
suspension systems, flight simulation, injection
moulding machinesetc. Hydraulic systems are
also common in aircraft, where their high power-
to-weight ratio and precise control make them an
ideal choice for actuation of flight surfaces
[1]
.
Although they are commonly placed as close
as possible to the device to which they are
supplying fluid in some applications, it is not
possible to place servo valves close to the actuator
due to the plant conditions. This is commonly
seen in the steel rolling industry [2].
In the last decades, some researchers have
published some works in the field of servo valves
modeling.
Watton and Hawkley, (1996) an approach is
developed by utilizing measurements of transient
pressure and flow rate at the inlet and outlet of the
line. A time series analysis technique is used in
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


81

such a way that the number of unknown
coefficients to be estimated is minimized. For
three different line configurations and a range of
operating conditions there is an accurate
prediction which is shown for three different line
configurations and a range of operating
conditions. The evaluation of just two
transmission line functions then allows a simple
model structure to be used for the simulation of
fluid power circuits incorporating long lines [3].
Krus and Nyman (2000) have demonstrated
how the actuation system control surfaces with
transmission line can be simulated using a flight
dynamics model of the aircraft coupled to a model
of the actuation system. In this way, the system
can then be optimized for certain flight condition
by "test flying! the system. The distributed
modeling approach used makes it possible to
simulate this system faster than real time on a 650
MHz PC. This means that even system
optimization can be performed in reasonable time
[4].
Dong Zhu and Lu (2010) prove that the long
pipeline in hydraulic system has some influence
on system performances and causes the system to
become unstable. They target a hydraulic servo
system with long transmission line between
hydraulic power supply and servovalve, a
mathematical model considering pipeline effect
established by means of the theories of
transmission line dynamics and hydraulic control
systems in which pipeline characteristics were
depicted by lumped-parameter model. Dong used
AMESim (a software for modeling, simulation
and dynamic analysis of hydraulic and mechanical
system based on bond graph and which is a
production of imagine corporation of France) to
simulate the impact on system dynamic behaviors
which were investigated theoretically and the
influences of pipeline structural parameters on
hydraulic system dynamic characteristics were
analyzed [5].
Yang and Moan (2011) studied a heaving-buoy
wave energy converter equipped with hydraulic
power take off. This wave energy converter
system is divided into five subsystems: a heaving
buoy, hydraulic pump, pipelines, non-return check
valves and a hydraulic motor combined with an
electric generator. A dynamic model was
developed by considering the interactions between
the subsystems in a state space form. The
simulation results show that transmission line
dynamics play a dominant role in the studied
wave energy converter system. The length of the
pipeline will not only affect the amplitude of the
transient pressures but also affect the converted
power transformed in the generator [6].
The purpose of this work is to study the control
of pressure losses in the transmission lines of
hydraulic system using a servovalve. A theoretical
analyses and experimental test are used to perform
the work by applying direct voltage to control the
pressure in a specific point in TL.
The system's dynamic characteristics have
been tested by using a PC (Personal Computer)
equipped with a data acquisition processor (DAP-
card). This will allow data based modeling to be
carried out, allowing prediction of the system's
response to a given control output.


2. System Description

2.1.1. System Hardware Description

As shown in Fig. 1, the pressure supply line
delivers hydraulic fluid from big power unit
supply to the test rig at a pressure up to 150bar. A
variable pressure relief valve is installed in the rig
so the desired pressure can be achieved on the rig
as the researcher needs. There is a temperature
sensor and flow meter on the supply line to the
servovalve. The valve to be used is an Ultra
servovalve from Moog, of type 4658-249-810,
shown in Fig. 2; the valve consists of two-stage,
nozzle/flapper, and dray torque motor unit.
Service port B (see Fig. 1) is blocked rather
than feeding to the annulus side of the actuator as
might be expected. The service port A is the exit
to the servovalve, where the second flow meter
and pressure transducer located. The servo and
actuator are connected via a long transmission
line. This line is expected to have an important
input to the dynamic response of the system due
to its considerable length. The actual actuator is
fixed into a position - it cannot move. This is
allowable because the system is used to provide
adequate force to counteract roll bending under
load (work roll bending system). The actual
displacement of these actuators are small, and
would ideally be zero. Hence, when modeling this
system it considered reasonable to ignore the
small actuator movements. The PC records the
Data Acquisition Processor (DAP card), which is
connected to the transducer display and amplifier
units as shown in Fig. 1.



J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


82















































Fig. 1. Schematic of System Set Up Cardiff University Laboratory.


2.1.2. System Control Software Description

The DAP-card is connected to the PC and has
its own operating system and it is provided with a
program called DAP-view through which the
control of the DAP-card is built. This program
starts and stops collecting data, as well as
outputting signals and logging every event. This
project requires the use of custom written control
commands, which will collect input signals to the
card, process them in accordance with the desired
control method, and pass them back to the DAP-
view program to be sent to the equipment. Custom
commands are written in C
++
language and have
to be compiled and downloaded into the DAP-
card. C
++
programs can only be changed by the
PC, so any adjustments to the custom commands
require the removal of previous custom
commands and recompilation and installation on
the DAP-card
[7]
.


2.2. Servovalve Construction

The servovalve is an interface between low
energy electrical signals and high-level hydraulic
power. Servo valves are electrically operated
proportional directional control valves. They are
usually four port units which control the quantity
of fluid they pass, as well as the direction. Most
common servovalves are made in the form of a
two-stage device [8].

2.2.1. First stage

The first stage contains a torque motor which
operates an armature and this armature pivots a
'flapper' which is situated between two fixed
nozzles. By applying a current to the torque
motor, the armature is rotated, and this moves the
flapper toward one nozzle, and away from the
other. The flapper is located within the valve and
hence is surrounded by hydraulic fluid. To keep
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


83

the torque motor free from oil, the flapper is encased within a flexible 'flexure tube'[8].
2.2.2. Second Stage

Second stage is typically a four-way spool
valve that controls the fluid flow to two service
ports. There is commonly a mechanical feedback
system in the form of a feedback spring attached
to the spool which acts to oppose the action of the
torque motor on the flapper. See Fig. 2[8].






Fig. 2 . Ultra /Moog servovalve, Type 4658 and its
Cross Sectional View
[9]
.

2.3. Servovalve Modeling

2.3.1 Steady State Modeling of Servovalve

When an electrical current is applied to the
coils of the torque motor, a torque is generated on
the armature. The armature and flapper are
supported on the flexure tube or sleeve which
separates the electro-magnetic and hydraulic parts
of the valve which also provides a low friction
pivot see Fig. 3. Four force components are
considered in the torque motor. These are a
positive function of the applied current and the
rotation. These are a positive function of the
applied current and the rotation. These forces are
opposed by a torque from the stiffness of the
flexure tube, and the net hydraulic force acting on
the flapper element [9].
=
!
"
#$ +%
&
' ()
*
+ (
(,
-
( .
/
)0
1
2 (1)

The mechanical feedback element is one of the
types of servovalve being considered. The torque
of this feedback spring can be considered as
follows for small values of #:

3
45
6
7
8(9 +:),
; =<
=
+(> +?)@ & A =BC (2)

Atonement equation (2) in (1) gives the total
torque on the torque motor Fig. 3, flapper and
spring combination, noting that the torque is zero
at steady state; the angle # can be deduced [9]:

D =
E
F
GH (I
J
K(L
M
(N
O
) (P
Q
(R +S)T
U
V
W
( X
Y
+Z
[
(\ +])
^

(3)

As can be seen from Fig. 4, there is a hydraulic
'bridge circuit' which is supplied with system
pressure. A small amount of fluid can flow out
through the fixed orifice and onward to the two
variable orifices created by the nozzle/flapper
interface, ultimately returning to the tank.
Flapper/nozzles in conjunction with a pair of
orifices used to generate a pressure difference by
small movements of the flapper positioned
midway between the nozzles, as shown in Fig. 4
[10].
The spool area and velocity are A, U
respectively. Typically the nozzle diameter is dn=
0.5mm, the flapper clearance in the mid-position
xnm= 0.03mm, and the orifice diameter do =
0.2mm. It is common for such a device to be used
in servovalves as a mechanical feedback, the
pressure difference generated being used to move
the spool. It will be immediately clear from Fig. 4
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


84

that at the flapper mid position, often called the
null position, the maximum leakage flow back to
tank will exist, hence producing a small inherent
power loss. As illustrating example, the flapper is
moved to the left, by electromagnetic means then
pressure P1 will increase and pressure P2 will
decrease, thus providing a pressure difference
across the spool which will then move unless
restrained in some way. The flow loss and power
loss will decrease as the flapper position is
changed [10]. To analyze the flapper-nozzle
bridge, the conventional restrictor flow equations
are appropriate and given by:

_
`
=a
b
+c
d
e,
f
gh
=ij
k
(l
mn
(o), (4)

p
q
=r
s
(t
u
v,
w
xy
=z{
|
(}
~
+); (5)
!
"
=#
$%
&
'
(
2()
*
( ,
-
)
.
,
/
0
=1
23
4
5
6
7(8
9
:;
<
)
=
; (6)
>
?
=@
AB
C
DE
F
2G
H
I
,
J
K
=L
MN
O
PQ
R
ST
U
V
; (7)

At condition in which the spool motion is
negligible, the steady-state performance of the
double flapper-nozzle amplifier may be derived
from equating Qa = Qx and Qb = Qy. This gives:

W
X
Y
=
Z
[\](^_`a )
b
=
c
d
e
f
, g
h
i
=
j
klm(nopa )
q
=
r
s
t
u
,
va =
w
x
yz
; (8)
{ =16(
|
}~

!
)
"
(
#
$
%
&
)
'
(
(
)*
+
,
)
-
, (9)

Then the differential pressure is given by:

.
/
0
( 2
3
4
=
456a
[1+7(1+8a )
9
][1+:(1(<a )
=
]
,
(10)
Considering the null condition where >a =0 and
?
@
A
( C
D
E
=0; that leads to:
F
G
H
=I
J
K
=
L
(MNO)
(11)
and the null gain condition is:
P(Q
R
S
TU
V
W
)
XYa
=
4Z
(1+[)
2
(12)

Because flapper operation is designed to be
around the central (null) position, the pressure
difference generated may be simply written as:
(\
]
( _
`
) =a
b
c
de
f g
h
(13)
Spool displacement is then determined from
the force balance across the spool which is
dominated by the feedback wire force and the
spool flow reaction force:

(P
i
( P
j
)a
k
=ky+2C
l
m
a
n
cos#[P
o
(P
pqrs
],
and a
t
=wx
u
(14)
And: P
vwxy
= P
z
(P
{
& P
|
= P
}
+P
~


Where, as will the spool end cross section area,
the spool orifice area ao for rectangular ports have
an area gradient w. by combining these equations
then give the relationship between spool
displacement and input current is as follows:
If:
$ =
(+b)x
nm
P
s
u
s
; (15)
[ =k
u
(k
m
+k(r +b)
2
+
o
n
r
2
P
s
x
nm

(16)
k
]
=2C
q
2
w cos0 (P
s
(P
Ioud
) (17)

Then:
x
s
=
(1-u)k
t

[
[k-k
]r
x
nm
rP
s
c
s
+k(+b)(1-u)
(18)

The spool displacement will be proportional to
input current provided that the denominator of Eq.
(18) is positive. The flow reaction equivalent
stiffness k
fr
will probably be much smaller than
the wire stiffness k, so that the effect of load
pressure difference P
load
may not present a
problem. In practice, << 1 and can be neglected.
Notice also that the destabilizing magnetic
constant -k
m
, the magnitude of which can be
varied during manufacture, the process known as
detuning. In particular, [ can be detuned to a very
small value by magnetically increasing k
m
and Eq.
(18) then becomes:
x
s
=
k
t

k(+b)
(19)
The input electrical torque is balanced by the
wire feedback torque because of spool position,
and the flapper will return to its central position
between the nozzles
[10]
. Clearly, for the same
servovalve, the value of k, r and b constants the
spool displacement which can be proportional to
current is dominated.

J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


85





Fig. 3. Feedback Spring Free Body Diagram
[9]
.


Fig. 4. Schematic of a Double Flapper-Nozzle Amplifier
used to Move A Spool
[10]
.




2.3.2 Dynamic Modeling of Servovalve

The steady state performance of the valve will
need to be augmented by a model for the transient
response, and also for the transient response of the
fluid transmission line. This will allow predictive
control to be used. For the servovalve that again
requires a finite time to change its spool position
in response to a change in applied current. The
combination of these issues means that the design
of both open-loop and closed-loop control
systems should take into account these dynamic
issues.
For the force-feedback type shown in Fig. 5, it
is clear what components which contribute toward
the overall dynamic performance. The use of a
current-feedback servo-amplifier means that the
current-buildup characteristic is extremely fast
when compared with other elements of the
servovalve.
The dynamic effect caused by the time
required to generate the drive current can be
ignored. However, there are effects produced from
the flapper inertia and fluid viscosity
[10]
.
The current-buildup and the dynamic torque
equations will have the following types:

k
t
i =(

( "
#
)$ +(%
&
((
)
)*
+
, +
-[.
/
+(0 +1)2](3 +4) +5
6
78
9:
+
;
<
=
>
?@
A
(20)
Apply the continuity equation on each side gives:
B
CD
E
F
G
H(I
J
KL
M
)
N
(P
QR
S
TU
V
WX
Y
Z
=+[
\
]^
_
`a
+
b
c
d
ef
g
hi
(21)

j
kl
m
n
o
p(q
r
st
u
)
v
( x
yz
{
|}
~
!
"
#
=(%
&
'(
)
*+
+
,
-
.
/0
1
23
(22)

Where:
4
56
=78
9
(:
;<
(>), ?
@A
=BC
D
(E
FG
+H) & I =JK

V
a
and V
b
are the internal, small volumes on
either side of and within the flow resistance
bridge.
The static force balance at the spool, including
the flow reaction force, is now modified to
include the dynamic flow reaction force, the spool
viscous damping, and acceleration effects:

(L
M
( O
P
)Q
R
=
S[T
U
+(V +W)X] +2Y
Z
[
\]
^
cos_[`
a
( c
defg
] +
h i j
kl
m
no
(
pq
r
st
u +v
w
xy
z
{|
+}
~

!
"#
(23)
Where:
$
%
=&
'
()
*
+
,(-
.
/0
1
)
2
,
3
4
=5
6
78
9
:
;<
=
>
(24a)
Obviously, the defining equations of
servovalve are nonlinear, and the solution also
requires the load specification so that the load
pressure difference (P
1
- P
2
) can be derived.
Considering the equations presented, it will be
seen that a valve dynamic performance depends
not only on electrical-electromagnetic-geometry
parameters but also on the load it supplied P
load

and, hence, on the load flow rate, the supply
pressure P
s
, and the magnitude of the input
current.
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


86

The tank (return line) pressure is usually
neglected in comparison to the line pressures. The
port opening area (wx
s
) is proportional to spool
displacement which is also proportional to the
current applied to the electromagnetic first stage.
Servovalve manufacturers also quote the rated
flow at the valve rated current and with a valve
pressure drop of 70bar, that is, the total pressure
drop across both ports. Consequently the
servovalve equations could be rewritten in the
following form
[10]
.
?
@
=A
B
CD(E
F
(H
I
) & J
K
=L
M
NOP
Q

(24b)
From the previous equations and the
contribution of dynamics behavior shown in Fig.
5, the amount of hydraulic fluid flow from the
servovalve depends on the pressure and the
current value coming to servovalve amplifier. In
the steady state condition, both of the
electromagnetic and mechanical properties are
considered constant, as well as the effect of the
amplifier, which converted the voltage value
coming from the DAP-view programs to the
servovalve amplifier. In other words, it can be
considered that the hydraulic flow rate (Q) is a
function of the voltage and the pressure as
follows:
R =S ( T ,U ) (25)
At steady state operating condition ( v
ss
, P
1ss
,
P
2ss
, Q
1ss
, Q
2ss
), the first linear term of the Taylor
series expansion for a nonlinear function will be
employed. Consequently small changes in each
parameter lead to:
VW
X
=Y
Z[
\]
^_
(a
bc
de
f
&
gh
i
=j
kl
mn
op
+q
rs
tu
v
(25a)
w
xy
=
z{
|
}~
!
="
#
$%
&
((
)*+
=
,
-./
0
12
3456 789: (25b)
;
<=
=
>?
@
AB
CD
=E
F
GH
IJK
=
L
MNO
P
QR
STUV WXYZ (25c)
[
\]
=
^_
`
ab
c
=
d
e
f
gh
2ij
k
(m
nop
=
q
rst
2(u
v
(x
yz{
)
|}~!"#$ %&'()*+,-./
(25d)
0
12
=
34
5
67
8
=
9
:
;
<=
2>?
@AB
=
C
DEF
2(G
HIJ
)

KLMNOPQR STUVWXYZ[\] (25e)
^
_`
=
ab
cd
ef
=(
gh
ij
kl
mn
op
=
q
r
s
t
=
2(u
v
( x
yz{
)
|
}~
=
2
!"
#
$%
,&'()*+,- ./0123456
(25f)
Also the servovalve equation could be written as:
7
8
=9
:
;<(=
>
( @
A
) & B
C
=
D
E
FGH
I
(25g)
In this application, the singe action operation
will be considered, so the port B (number 2 in
previous equations) has been canceled, see Fig .1,
thus, it is needed to consider the segment of the
first port to find the value of flow gain and
pressure coefficient equations (25b & 25d) at
steady state condition.
Practically, the dynamic characteristic is often
specified by the manufacturer as a frequency-
response diagram for the spool position (input)
and the flow rate (output) for a typical
performance range.






J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


87




Fig. 5. Contribution to Servovalve Dynamic Behavior
[1] [10]
.



Fig. 6. The Line Dynamics Approximation using Lumped Elements For Laminar Mean Flow
[4]
.

2.4. Transmission Line Modeling

Hydraulic pipelines, when they are of
significant length, they have an important effect
upon the performance of many systems. The
electrical analogy is a useful mechanism for
understanding this approach
[10]
. Consider linear
characteristics and a slug of fluid, (a) is cross
sectional area and the length is (l ).

a. Fluid resistance: The pressure drop %P, along
the fluid element for laminar flow , is given
by:
#K =
LMNOP
QR
S
T U #W
X
=YZ (26a)
Hydraulic resistance Electrical resistance
b. Fluid Compressibility:
#\ =
]
^
_`
ab
U #d =e
fg
h
ij
(26b)
Fluid compressibility Electrical
capacitance
c. Fluid Inertia: The pressure drop that is due to
fluid acceleration is given by:
#lm =nop
qr
st
, #v =
wx
y
z{
|}
U #

=
!
"
#$
%&
(26c)
Fluid mechanical mass Electrical
inductance

'()*+,-./0 U 2 =
34567
89
:
;
, <=>?@ABCDE U G
H
=
IJ
K
& LMNOPQRSTUV U X =
Y
Z
...(26d)

If a system is expected to have a frequency
component that is comparable with this frequency,
line dynamics must be modeled with some
accuracy. The issue is how to distribute R, L and
C in the line and how many "lumps" should be
used
[10]
. This work shows a two-lump
approximation using a pair of networks as
shown in Fig. 6.
The set of equations using this approximation,
and working from left to right, may then be
written as follows:
[
\
( ^
_
=
`
a
bc
d
ef
, g
h
(j
k
=
l
m
no
p
qr
, s
t
(
u
v
=
w
x
yz
{
|}
(27a)

~

( !
"
=
#
$
%
&
+
'
(
)*
+
,-
, .
/
( 1
2
=
3
4
5
6
+
7
8
9:
;
<=
(26b)

These equations can be resolved when the
input and the output pressure flow relationships
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


88

have been included to close the solution. So the
pressure and flow meter sensors are needed in the
test rig to record the overall pressure difference
between the input and output for the transmission
line, as shown in Fig. 1 and Fig. 7. This pressure
difference can be used to calculate the losses that
occur in the TL and the fittings which include
such as bends, elbows, restricted valves and
sudden expansion or contraction and other minor
losses.
To solve the previous equations, it is necessary
to make some practical experiences to achieve
steady state condition, through collecting this kind
of data. The values of unknown's constants
mentioned in equations (25-25g) such as that flow
and pressure gains can be calculated, as well as
the losses that occur in transmission line.






(a) Top view for TL Test Rig. (b) Side view for TL Test Rig & Measurement Unite.
Fig. 7. Test Rig Transmission Line, Fluid Power Laboratory, W20/ School of Engineering / CARDIFF
UNIVERSITY/ UK.




Fig.8 Voltage Function Generato
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


89

3. Experimental Approach

3.1. Modeling Assumptions

Fluid properties: As an initial condition, fluid
properties will be; density of the hydraulic oil
860kg/m
3
, absolute viscosity will be taken as
0.0258kg/m*s. Effective bulk modulus will be
1.5GN/m
2
. Temperature effects are considered to
be negligible, because the hydraulic system is
supplied with a heat exchanger which maintains
the hydraulic fluid temperature does not exceed
50
o
C. Pressure variation effects are expected to be
negligible, but will be investigated with respect to
bulk modulus to confirm the system time-delay, a
function of bulk modulus which is not affected.
The bulk modulus is expected to be independent
of pressure variation if there is considerable air in
the system (either dissolved or as bubbles) so in
this case there may be an effect.
Transmission line dimensions: TL consist of
one 16m length of steel pipe the internal diameter
is 12.5mm, wall thickness 2.3mm and with
Young's modulus of 210GN/m
2
see Fig.7 (a).
Pressure transducer: (Parker- Sensor, type:
SCP-150-1-06, 0-20 mA), shown in Fig.7 (b). The
sensor response will be considered to be ideal, i.e.
pure gain with no time delay. As investigated
from (Hawkley), the transducer response has been
considered an ideal sensor
[11]
.
Flow meter: (Parker- Flow- sensor, type;
SCQ-150-0-02, Range -150_+150 L/min, -3_+3
VDC), as illustrated in Fig. 7(b). The response
will be considered accepted after subtracting the
offset values at zero system pressure.
The system has a continuous nature The
sample time must be kept to a minimum to avoid
loss of dynamic characteristics. The DAP view
cannot take more than 1Gbytes and the maximum
voltage with applied to the servovalve is 10Volt.
Transmission line termination is assumed to
be a simple (capacitor, resistance and inductance);
the actuator cannot move (constant volume) as
well the zero internal leakage is expected.
Calibration of pressure transducer and
DAP-card, in all figures presented in this work, it
presents a y-axis value of 'digital output' supplied
from the pressure transducer to the DAP-card and
saved in Excel-file. These values should be
translated from digital voltage value to pressure
value in bar. The digital sensor values are
multiplied by (9.1*10
-3
).
Calibration of flow-meter and DAP-card, to
calibrate the flow rate in (L/min), the digital
output from flow meters which collected by the
DAP-card is needed to multiply by 15.26*10
-3
.

3.2. Open Loop Pressure Control Results

Relying on the reference signal generator
(Function generator) Fig. 8, The generated
voltages are sent to the servovalve to change the
system pressure in the chosen control point, and
by using two types of waveforms, sinusoidal and
square-wave. The voltages value was used in
function generator &4Volt. The delivery pressure
line is set to be 50bar, noting that the highest
values can DAP-card that gives it to the
servovalve is 10Volt. The results are as shown in
figures (Fig. 9, Fig.10 & Fig. 11). In these figures,
the frequency values have been seized by using
the voltages function generator as1Hz.
3.3. Servovalve Transient Response

To find servovalve response and the amount of
flow rate resulting from change voltage value, it is
necessary to set up a special program using C
++

language in the PC and use the DAP-card ports to
send the voltage signal to the servovalve amplifier
card. This voltage signal energizes the servovalve
torque motor to move the spool of the valve to
build the pressure in transmission line. The C
++

program needs to change the voltage value
gradually step by step at time interval of 2seconds
to get a dynamic stability. This program starts
from zero value to the highest value given by
DAP-card (10Volt) and then back to a zero value.
The stability can be observed in the flow rate in
each step of the program as shown in Fig.12.
After obtaining the flow values at each step,
calculate the average flow rates in each step inside
transmission line to find the relationship between
voltage change and flow rate values changing.
The flow rate measuring at point 2 (see Fig.1),
represents the servovalve transient response
output by the effect of the voltage input as
illustrated in Fig.13.

3.4. The Transmission Line Losses and the
Servovalve Gains

To find the properties of flow inside the
transmission line, a steady state conditions has
been create to record the values of pressure and
flow rate inside the TL. By using these data, the
servovalve gain could be found from the Excel-
file that has been recorded by employing the
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


90

equations (25b, 25d & 25f). At 1.5 degree opening
restrictor valve, the pressure inside the TL is
accumulated. 50bar as a pressure line system is
supply with a direct control using the voltages
function generator. The results are coming from
the square-wave voltage illustrated in Fig. 14 &
Fig. 15.
Through the act of the experimental test, the
nature of the flow rate has been checked as a
laminar flow by calculating the Reynolds number
and the velocity from recorded flow. The relation
between the voltage and the spool position inside
the servovalve has been found from Bode diagram
supported by manufacturer data sheet, see
{Appendix-A}.
The fluid power unit of the laboratory supplies
many test rigs. So it was difficult to find out the
miner losses through the line supply the test bench
which have been used as well as the fluctuation
effect. The minor losses mean the pressure drop
due to elbows, junction, reducers, valves and
hosesetc. The major losses occurring in TL can
be calculated from the equation below
[12]
:
#?
@ABCD
=4E(F G H
I
J
) K
L
MN
O
PQ
R , STUVW: X
Y
=
Z
[
\
]
(28a)
An experimental test have been used to find the
total losses ( !P
Total
) between point 2 & 3 as
located in Fig. 1, where:
#_
`abcd
=#f
ghijk
+#m
nopqr
(28b)
Then:
#t
uvwxy
=4z({ G |
}
~
)

!"#$%
&
'(
) (28c)
So the equivalent length L
equiv
= 18.46m, and the
corrected total TL length become:
*
+
=,
-.
+/
01234
(28d)
The total corrected length (L
c
=l) has been used in
equations (26d) to solve the mathematical model.

3.5. MATLAB Simulation and Results

To find a mathematical model for the TL
equations (27a and 27b) and the servovalve flow
rate equations (24) supported by the voltage
linearization equations (25b, 25d & 25f) that
mentioned before, MATLAB m-file program can
be used to represent the transmission line effect
after converting to electric analogy.
There has been a great complexity of the
hydraulics system in the fluid power laboratories
in Cardiff University and the large size of the
power unit system as well as the length of the
pipes connected with TL test rig. Besides, the
hydraulic pressure provided for two laboratories
and several hydraulic apparatus are connected to
them. For this reasons, some assumptions have
been taken in to account such as: the pressure
provided by the supplied pressure line which
delivers the fluid to the servovalve was considered
as constant pressure in MATLAB program. This
will neglect the fluctuation of the pressure value
caused by system complexity.
To find a clear comparison between the
selected practical experiences and the program
built by using MATLAB m-file, it is essential to
find the values for the gain needed. The sample
was taken from the recorded data, which represent
a step input supplied by the function generator and
compared to the output values resulting from the
MATLAB m-file designed program. As shown in
figures (Fig. 16a & Fig. 16b) and (Fig. 16c & Fig.
16d), the consequence output values are
compatible in the behavior as well as the value.
The exceptions of that compatibility seen in the
overshot behavior can be explained by the
pressure input values in the instant of the step
input. The pressure input value to the servovalve
is decreased in the instant of the step input while
the pressure input value in MATLAB program
maintains in constant value. The result predicted
from MATLAB m-file program shows the
pressure values variation in the effect of
servovalve opining and the TL delay effect as
illustrated in Fig.16b.
The noise and oscillations shown in Fig 15 and
Fig.16c, for the values of flow rate result from the
nature of flow meter design. As mentioned
previously there is an error should be dealt with
and reduced by subtracting the offset value and
depending on the average values to solve the
mathematical equations.
The MATLAB m-file program can create a
various shape of input voltage value (sine-wave or
square-wave). The pressure and flow rate output
could be seen for square wave in Fig. 17a &
Fig17b. The behaviors of the mathematical model
have a good approach comparing to the
experimental test seen in Fig. 14 & Fig.15. The
sine-wave input voltages generated in m-file
program are supplied to the servovalve, the
pressure and flow rate output shown in Fig.18a &
Fig. 18b. It is clear that there is a delay effect of
the TL especially in pressure output as seen in
Fig. 18a.




J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


91



Fig. 9. Open Loop Controlled by Voltage Function Generator Ps=52bar with sin-wave voltage Amplitude
Fg=4volt.



Fig. 10. Open Loop Controlled by Voltage Function Generator Ps=50bar with Square-Wave Voltage Amplitude
Fg=4volt.



Fig. 11. Open Loop Controlled by Voltage Function Generator Ps=50bar with Saw-Wave Voltage Amplitude
Fg=4volt.
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


92



Fig. 12. Flow Rate and Servovalve Voltage Data Variation In Transmission Line at Ps=70 bar.


Fig. 13. Transient Responses of Servovalve.




Fig. 14. Open Loop Controlled by C++ Program (1.5 Degree Opening from the Restrictor Valve), Square-Wave,
P1=50bar, Fr=0.25Hz.
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


93



Fig. 15. Open Loop Controlled by C++ Program (1.5 Degree Opening from the Restrictor Valve), Square-Wave,
P1=50bar, Fr=0.25Hz.



Fig. 16a. Open Loop Controlled by Voltage Function Generator P1=50bar, Time Sampling = 8ms.



Fig. 16b. Open Loop Controlled by MATLAB m-file P1=50bar, Time Sampling = 1ms.
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


94



Fig. 16c. The Flow Rate in Open Loop Controlled by Voltage Function Generator P1=50bar, Time= 2.5 s.



Fig. 16d. The Flow Rate in Open Loop Controlled by MATLAB m-file P1=50bar, Time= 2.5 s.



Fig. 17a. The Pressure Values in Open Loop Servovalve Controlled, by MATLAB m-file P1=50bar, Square
Wave, Time= 4.0s.

J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


95



Fig. 17b. The Flow Rate Values in Open Loop Servovalve Controlled, by MATLAB m-file P1=50bar, Square
Wave, Time= 4.0s



Fig. 18a. The Pressure Values in Open Loop Servovalve Controlled, by MATLAB m-file P1=50bar, Sine Wave,
Time= 4.0s.



Fig. 18b. The Flow Rate Values in Open Loop Servovalve Controlled, by MATLAB m-file P1=50bar, Sine Wave,
Time= 4.0s.

J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


96

4. Conclusions

Foregoing became clear that the use of the
open loop control study is very important to know
the properties of the system and its capabilities.
The open loop concept is the first step to entrance
and design the efficient closed loop control to
attenuate the delay TL effect.
The servovalve (mechanical feedback) is one
of the efficient valves that can be used to control
the specific pressure value on the actuator. To
reduce the impact of losses resulting from the
hydraulic flow rate inside a long TL, the closed
loop concept should be used.
The lumped element electrical analogy
method for laminar flow is an efficient method
which simulates the TL module. The relation
between the flow rate and the supplied voltage of
the servovalve amplifier is almost a linear relation
one and the effect of design complexity could be
neglected. This concept is so clear in the transient
response of the electro-hydraulic servovalve.
Found through experimental tests that the
value of the time sampling recorded through a
DAP-view program could be taken as 1
millisecond and that provided a good accurate
data that can be used to calculate the simulation
system in MATLAB m-file program. The DAP-
view program is used to control the servovalve
and it shows a precise action and follows the
program designed in C
++
formula with different
shape of functions.

Nomenclature
Latin Characters
Character Description Units
a Hydraulic pipeline cross
section area
m
2

a
n
, a
nx
, a
ny
Nozzle cross section area m
2

a
n
The nozzle area m
2

a
s
The spool cross section
area
m
2

a
o
The spool orifice area m
2

B
v
The fluid viscous
coefficient
m
2

C
q
, C
qn
,
C
qo

The flow coefficients of
the orifices and the
nozzles
-
C Electrical capacitance Farad
d
i
Pipe ( Transmission line)
internal diameter
m
d
n
The nozzle diameter m
d
o
The orifice diameter m

E
i
Electrical inductance
F Frequency Hz
Fg Voltage input from the
Voltage Function
Generator
Volt
f Darcy friction factor -
Q
1
, Q
2
Flow rate from servovalve m
3
/s,
l/min
Q
a
, Q
b
The orifice discharge m
3
/s
Q
x
, Q
y
The nozzle discharge m
3
/s
i The input torque motor
current
mA
I Current A
!i Input differential current mA
J The flapper inertia kgm/s
P
a
, P
b
Pressure at nozzle a and
nozzle b
N/m
2

P
Load
The pressure difference
on the load
N/m
2

P
s
System pressure N/m
2
;
bar
k
a
Flexure tube rotational
stiffness
Nm/ra
d
k
c
Servovalve constant
k
f
The cantilever springs
stiffness
Nm/ra
d
k
fr
The flow reaction
equivalent stiffness

k
m
Electromagnetic spring
constant of torque motor
Nm/ra
d
k
t
Electromagnetic constant
of torque motor
Nm/A
mp
l Hydraulic pipeline length m
L Electrical inductance Henry
L
c
Corrected length m
L
equiv
Equivalent losses length m
L
TL
Transmission line length m
R Electrical resistance '
r The distance between the
nozzle center line and
flexure joint
m
5 The torque on the flexure
tube
N.m
T
f
The resisting torque N.m
U The spool velocity m/s
U
3
The output mean flow rate
velocity from the
servovalve
m/s
u Control signal volt
V Volume m
3

V
e
Electrical Voltage volt
v input voltage volt
w The rectangular port area
gradient
m
J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


97

x The displacement of the
flapper at the nozzles
m
x
s
The spool displacement m
x
nm
Flapper clearance in the
mid position

y Total spring deflection m
Z Constant for the
servovalve design
parameter (valve
characteristic)



Greek Symbols

Character Description Units
" Effective Bulk modulus N/m
2

Fluid absolute viscosity kg/m.s
# Fluid density kg/m
3

$ The rotation of the
armature and flapper
rad


subscripts
ss Steady state operation
condition
-
-


5. References

[1] DSP Control of Electro-Hydraulic Servo
Actuators (January 2005), Application
Report, Texas Instruments SPRAA76.
[2] Le Bon. A. & Hug, P. (1996), "Book of steel,
"The manufacture of plain carbon sheet steels
- reheating and hot rolling", Lavoisier
publishing.
[3] Watton, J., & Hawkley, C. J. (1996). An
approach for the synthesis of oil hydraulic
transmission line dynamics utilizing in situ
measurements. Proceedings of the Institution
of Mechanical Engineers, Journal of Systems
and Control Engineering, Part I, 77-93.
[4] Krus, P. &. (2000). COMPLETE
AIRCRAFT SYSTEM SIMULATION FOR
AIRCRAFT DESIGN PARADIGMS FOR
MODELLING OF COMPLEX SYSTEM.
ICAS 2000 CONGRESS (pp. 613.1-613.9).
Sweden: Department of Mechanical
Engineering, Link(ping University.
[5] Dong, C., Zhu, Y., & Lu, J. (2010). Modeling
and dynamic behavior analysis of a hydraulic
servo system with consideration of pipeline
effect. International Conference of Logistics
Engineering and Management: Logistics for
Sustained Economic Development -
Infrastructure, Information, Integration (pp.
3153-3160). Harbin, China: College of
Engineering and Technology, Northeast
Forestry University.
[6] Yang, L. & Moan, T. (2011). Dynamic
analysis of wave energy converter by
incorporating the effect of hydraulic
transmission lines. Ocean Engineering,
Elsevier, Vol. 38, pp. 1849-1860.
[7] DAP-view user manual, Microstar labs,
2001.
[8] Cundiff, J. S. (2002). Fluid Power Circuits
and Controls: fundamentals and applications.
Library of Congress Cataloging-in-
Publication: (Mechanical engineering series),
TJ840.C85 2001, 621.2+dc21.
[9] Watton, J. (1989). Fluid power systems,
Modeling, simulation, analog and
microcomputer control. UK: Prentice Hall.
[10] Watton, J. (2009). Fundamentals of Fluid
Power Conntrol. CAMBRIDGE/UK:
CAMBRIDGE UNIVERSITY PRESS.
[11] Hawkley, C. J. (1996). The modelling and
simulation of an electrohydraulic pressure
control system.PhD, thesis. UK: School of
Engineering, UWCC, Cardiff University.
[12] Frank M. White (2005), Fluid Mechanics
Fourth Edition, University of Rhode Island
McGraw-Hill Series in Mechanical
Engineering.











J afar Mehdi Hassan Al-Khwarizmi Engineering J ournal, Vol. 9, No. 3, P.P. 80-99 (2013)


98



!"$ %&() *,-. &/1203 45#&7(03 9):+3;=03 4/1) 9 '&-03 ? 3 4@,A ? 99 -80 ) 2013 (


99



!#$%'( )+'( -/1$23 45/'3.7/93-:5'( .;(<='( ?$=@'( ?(7+BC$D )EF'( HIJ K-L/M'(
N/'3.7/:I'

P$MQ R7:> -STU * <MV( W50 ** R</IJ 7=Q( 7U$> ***
R7:> 8(XZ 7=Q( ****
* %&(&*+(&,./ %0134./ 578 / %&9:.:3(<./ %>-+?./
** %0134./ %&@A / B2C+A %>-+9 - +&*+D'FG
*** 5H3./; JFD&7./ %013K 578 / %&9:.:3(<./ %>-+?./
**** L,=N,./ P&3Q<./ %013K 578 / S-TC/:U./ %0134./ %&@A
JafarMahdi1951@yahoo.com * ,-./13567 8:/;57 :
xue@cf.ac.uk ** ,-./13567 8:/;57 :
moleiwi@yahoo.com :,-./13567 8:/;57***
afmkridi@hotmail.com **** ,-./13567 8:/;57 :



[A\+'(

=?@1AC57 EG< =G5.08G@?5 HJKL57 MO57 /;1Q: R6KS57 UQ< ,W ) >/SC57 ( /XYD. [@D ]/^GA57 \KC_ . . HJKL57 MO57 `abC- d?f gi/2 `B70857 jbl
H_7m57 =G5.08G@?5 =?@1AC57. ,3G5.08Gl./@357 0h7mC57 ]/^GA57 \KC_ EG< . d?f ]/^GA57 InP ED HJKL57 MO57 HP79 Mop57 `CGJ d?f ]/^GA57 rC2
0h7mC57 \KCs< tKO57 [OpC57 d57 `?_7m57 `G15mu57 `CGJ . `v:/w \78O1BK< HJKL57 MO57 HP79 ,JK;^57 FK:/x57 `abC- rC2 57 HJKL57 MO57 H:mS15 [3?5 d
`GXK</@i `DmyLD . I7 {D7/< \78O1B7 InP ED /|K;D /G} H3~<. `G15mu57 `579 85mD hK@a InP ED `G15mu5K< [3S157 ,W ]/|K;C57 `v:/^57 \78O1B7 [2 ) C
++
(
{DK-/< D M;2/: b57. ) DAP-view ( `GXK</@357 ]/X7857 ,W n_7 !sLC57 ) DAP-card ( 2mG;Dm357 d57 K@1WKq# E3C: ,157 d?f ]/^GA?5 `GsO~57 R7/
]0K1OD `LGQD `^v- 8Lf HJKL57 MO57 HP79 Mop57 `CGJ . `Jnf FKGT67 !?}7 ,W 0h7mC57 \KC_ ED $0KO57 FK:/x57. `G15mu57 EG< `JnQ57 F7 d57 H_m157 [2
`G^P . "n2KC57 {DK-/< \78O1B7 [2 ) m-file ( r-Ki {XK1L57 F%< 8a. . ,qK:0 HG&C2 `5KT 9Kx:6 D H'KC1D. `Tm1uC57 ]/X7857 d?f ]/^GA57 `5KS5 ]8Ga
`G?CQ57 "0Kx157 .