Modeling and Simulation of Gear Pumps

based on Modelica/MWorks

Chen Liping
1
Zhao Yan
1
Zhou Fanli
1
Zhao Jianjun
1
Tian Xianzhao
2

1
CAD Center, Huazhong Univ. of Sci. & Tech., Wuhan, China, 430074
1
Suzhou Tongyuan Software & Control Tech. Co., Suzhou, China, 215123
[email protected] [email protected]
{fanli.zhou,jjzhao168}@gmail.com, [email protected]
Abstract
In this paper, we present a new method of modeling
the hybrid external gear pumps based on Modelica
and assess it experimentally. We model the whole
working process of an external gear pump. The
chamber of the pump is divided into a set of Control
Volumes (CVs), whose effective volumes change
along with the rotation of the gears. The CVs take in
fluid from the inlet port and squeeze fluid out at the
outlet port. The whole model of the pump also takes
into account flow ripple, pressure distribution, leak-
ages, meshing conditions, etc. Details of each com-
ponent of the whole pump are provided. From the
pressure distribution in space of the gear tooth, we
can calculate radial force on shaft, based on which
the shaft motion can be simulated.
We carry out a set of predictions in MWorks and
report some results on the post processing. These
results are consistent with those from the experi-
mental data.

Keywords: Hybrid Gear-Pump Modeling; Control
Volumes; Pressure Distribution; Radial Force on
shaft; Modelica

Nomenclature
B the gear thickness
h
c1
the radial clearance of the bearing
h
c2
the axial clearance of the bearing
n the tooth number
1 Introduction
The gear pumps are among the oldest and most
commonly used pumps in the industry. It has become
the main choice for fuel system designers due to long
life, minimum maintenance, high reliability, capabil-
ity to operate with low lubricating fuel, low heat in-
put to fuel, small size, and low weight.
External gear pumps use a simple mechanism
(two gears) to generate flow and therefore have a
minimum number of parts associated with the design.
However, many factors ignored in the design, for
example the volumetric efficiency and the pressure
peaks, may greatly influence the performance of the
pump.
In order to improve the design of the pumps, a
good first step is to develop a mathematic model for
the simulation. Some work focusing on different as-
pects of this aim has been done in the past few years.
Manring and Kasaragadda [1] presented an ap-
proach to finding a solution for the instantaneous
length of action for the two contacting teeth in order
to study the flow ripple. In theoretical research,
pumps are considered in an ideal case: (1) the fluid is
incompressible, (2) fluid leakage is neglected, and (3)
the pump parts are inflexible.
Heisler et al [2] used the computational fluid dy-
namics (CFD) to better analyze the effect of modify-
ing the design of an existing external gear pump.
This paper also developed a new approach to simu-
lating the helical gears. Although the computational
time and model complexity have been reduced,
simulations can be done only under some restrictive
assumptions, including restricted flow domain, defi-
nite values of the boundary conditions, and fixed
position of the gear axes.
Wahab [3] presented a simple analytical and ex-
plicit approach to predicting the leakage flow rate
under different inlet/outlet pressure differences.
Castilla et al [4, 5] studied experimentally the
movement of the shaft of a driven gear in a gear
pump, in particular the dynamics of the shaft in the
journal bearing of a gear pump.
Elia [6] developed a mathematical model that
simulates the running in of external gear pumps, tak-
ing into account the factors mentioned above. Be-
Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
421
sides, the model also estimates the quantity of mate-
rial taken away.
Following a different way, Vacca et al [7] provid-
ed a global description of the pump or motor. They
implemented a model, named HYGESim, in the
AMESim platform. The model uses both standard
sub-models and C++ sub-models developed by the
authors.
Though valuable in some aspects, these studies
are unable to develop a global hybrid model, which
is the goal of our paper. We construct a hybrid mod-
el of the external gear pumps by using a multi-
domain modeling language, Modelica on MWorks

[8]. The model takes into account the leakages, com-
pressibility of the oil, flow ripple, pressure distribu-
tion, etc. It can also calculate forces and torques,
which are the base of modeling the shaft motion.
The paper is outlined as follows. Section 2 de-
scribes the mechanical system. Section 3 gives an
overview of the theory on which the pumps are mod-
eled based and details about each component of the
whole pump. Section 4 reports the simulation results
and campares them with those in [7]. Section 5 con-
cludes the paper.
2 Pump Description
Figure 1 shows a cross-sectional view taken through
the gears of a typical gear pump.

Figure1 Control Volumes Defined in the Fluid Dynamic
Model
Note: like most actual gear pumps, this pump has
two identical gears to displace fluid. The superscripts
g1 and g2 denote the driving and driven gears re-
spectively. The tooth number of gear 1 is the same as
that of gear 2. The Control Volumes (CVs) within
the tooth gaps are variables dependent on the gear
angle. The subscripts i is the index of these Control
Volumes. The Outlet and Inlet volumes are two fixed
volumes at the outlet and inlet ports respectively.
To produce a flow within a gear pump, fluid is
carried by the CVs from the intake side of the pump
to the discharge side of the pump through the transi-
tion zone. As the gears rotate, these CVs increase
their pressure to when reach the high-pressure cham-
ber. As the gear teeth mesh in the meshing zone, flu-
id is squeezed out of each tooth gap by the mating
tooth. When two tooth pairs contact, a trapped vol-
ume is generated. This may cause positive pressure
peaks and the onset of cavitation. (To avoid this, the
trapped volume must be connected to the high or low
pressure chambers, which is the role of the relief
grooves in the lateral bushes.) On the intake side, the
gear teeth are coming out of the mesh. The volumes
of CVs increase so that fluid is inhaled into the tooth
gaps. This process repeats itself for each revolution
of the pump and therefore displaces fluid at a rate
proportional to the pump speed.
3 Model Description
3.1 Overview
In this section, we first give an overview of the
pump model. Some components come from a free
library HyLibLight [9] based on Modelica.
The aim of this paper is to construct a hybrid
model of the external gear pumps. It considers the
leakages, compressibility of the oil, flow ripple,
pressure distribution, etc. Specifically, the pump is
modeled under the following assumptions:
1) The gears and the housing case are rigid; only
the oil is compressible.
2) The position of the shaft is known before the
simulation and is fixed during the simulation.
3) The pressure at every single isolated region
and fixed volume is well-proportioned.
4) The temperatures in all CVs are the same and
constant.
5) The tooth numbers of the two gears are the
same.

In Figure1, a pump is divided into 2n+2 control
volumes, where n is the tooth number of a gear. Fig-
ure 2 shows the flows between those control vol-
umes. This method is similar to but not the same as
that in [6]. There is no variation in the number of the
control volumes which causes variation in the num-
ber of differential equations.
Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
422

Figure 2 The Flows Between the Control Volumes
Note: arrows do not mean the flow directions; the
direction of the flow depends on the pressure differ-
ential and the rotation direction. As the gears rotating,
the channels of some flows are shutting down and
some of others (not shown in this figure) are turning
on.
Dark blue lines in the meshing zone represent the
leakages from a trapped volume on one gear to a CV
on the other gear.
The flow between CV
i
g1
and CV
i+1
g2
is the leakage
through the clearance between two meshing surfaces;
The flow between CV
i
g1
and CV
i
g2
is the leakage
through the clearance between two un-meshing sur-
faces.
Relief Grooves are chambers to which the CVs
connect in turn. Flows from the CVs to the Relief
Grooves in the meshing zone are shown in Figure 3.


Figure 3 Flows from the CVs to the Relief Grooves (RG)
RG1 on the outlet side is used to decrease the
peak of the pressure trapped in the CVs; RG2 on the
inlet side is used to restrain the cavitation.
As shown in Figure 3, the dashed line 1 means
the flow between CV
8
g2
and RG1 is forming. Solid
line 2 and 4 mean that there is a fluid flow between
CV
8
g1
and CV
7
g2
through RG1. Dashed line 3 means
that the channel between CV
7
g1
and RG1 is being
turned off.

Figure 4 Calculated Forces & Torques Acting on Gear 1
Figure 4 shows the calculated Forces and Tor-
ques acting on the driving gear, where F
p
g1
is the re-
sultant force of the pressure distributed along the
circle, F
c
g1
is the contact force, F
b
g1
is the force load
on the bearing of gear1 and T
i
is the torque supplied
by the pressure in CVi.
Note: most of CVs, except those in the meshing
zone, do not have any effect on the gear rotation,
because the pressures in these tooth gaps counteract
and no torque is acting (for example, T
i+1
=0). De-
tails will be given in the next few sections.
The coordination system is shown in Figure 5.

Figure 5 Coordination System
3.2 Control Volumes
3.2.1 Volume
A Control Volume is the space between two adja-
cent teeth. The maximum volume (see Figure 5, the
gray region) of this gap is related to the gear dimen-
sions (e.g. tooth number, modulus, thickness, etc).
As the gears rotating, gear pairs mesh together. A
tooth on the other gear squeezes into the gap, so the
effective volume of the gap decreases (see Figure 5,
the blue region). The effective volume of CVi is a
function of the angle
i


Figure 6 Effective Volume as a Function of Angle
Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
423
This relationship can be expressed in an approxi-
mate equation:
max 1 2
3 2 1 max min
max min
, < - or > +
[ ( ) / 2
,
cos(( ) ) ( ) / 2)]
V
x x x V V V
otherwise
f V V
t u t u
t

+ +

where

3 31 32
2 21 22
1 11 12
cos(( ) )
cos(( ) )
cos(( ) 2 )
x c f c
x c f c
x c f c
t
t
t
= +
= +
= +


1 2
2 / ( ) f t u u = +


11 12 21
22 31 32
0.25 0.75 0.25
0.75 0.4 0.6
c c c
c c c
= = =
= = =

2 / ( 1) ,if 1

2 / ( 0.5) ,if not
n Index gear
phase
n Index
t
t

=

1
and
2
are the critical angles (see Figure 7).

Figure 7 Critical Angles
3.2.2 Pressure distribution
The pressure in each CVi is given by the equation:
( )
cham
p q q
V
|
= +

where
oil bulk modulus
q
cham
the rate of the net flow that flows into the
chamber
p pressure of the volume

The pressure distribution is shown in Figure 4.
3.2.3 Radial Force Produced by Pressure
As shown in Figure 4, the load on a bear is the
sum of the loads produced by the distributing pres-
sure and the contact force. Ideally, the loads due to
the pressure of most CVs have only one radial com-
ponent (F
r
in Figure 4), since we have assumed (in
section !) that the pressure in
any isolated region is well-proportioned.
When a CV turns into the meshing zone, the
chamber will be divided into 2 or 3 parts by the cou-
pling tooth on the other gear (see Figure 8). The
load due to the pressure in red zone 1 is equal to the
sum of F
t
and F
r
. And

2
1
2
1 1
( )
( * )
g
t i m a
g
r i m
F B P r r
F B P r
=
=

Similarly, for zone 2 and zone 3 we have:

1
2
1
2 2
( )
( * )
g
t i n m
g
r i m
F B P r r
F B P r
=
=


2
3 1
2
3 1 1
( )
( * )
g
t i a n
g
r i n
F B P r r
F B P r

=
=

Then:

3 3
1 1
i tj rj
j i
F F F
= =
= +



Figure 8 The chamber of CV
i
g1
divided into 3 parts
3.2.4 Torque
Because only the component Ft has an effect on
the gear rotation, the torque produced by CV
i
g1
can
be calculated as follows:
1 2 1
( ( ) ( ) ( )) / 2
t a m t m n t n a
T F r r F r r F r r = + + + + +

Interfaces and Icon
The interfaces and icon are shown in Figure 9.

Figure 9 Icon of a Control Volume
3.3 Internal Leakages
Both gears rotate within the housing. The inter-
space is made because of lubrication between two
relative moving objects. The gap between two gear
teeth and casing is assumed to maintain a known
value. Leakage occurs through these internal clear-
ances due to both the pressure differences and the
relative motion.
Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
424
There are four types of sources of internal leak-
ages due to the pressure differences:
Q
p
: Flows from one CV to an adjacent one
through the tips of the gear teeth and the case.
Q
z
: Flows from the outlet volume to the inlet vol-
ume through the meshing zone.
Q
m
: Flows through the gap between two mating
surfaces.
Q
d
: Flows from the CVs to the drainage circle
through the interspace between the gear and the case.
There is another kind of leakage that is caused by
the relative motion and oil viscosity (see Figure 10,
Q
w
). It results in a little increase in the flow rate be-
cause the rotation brings oil to the outlet volume.

Figure 10 Internal Leakages
In general, for the first type of leakage it is as-
sumed that the pressure around the case rises linearly
across each tooth tip. Hence, the volumetric flow
from CV
i
g1
to CV
i-1
g1
due to the pressure drop is giv-
en by ([2]):

3
/ (6 )
pi
Q dp G dp h b r v = =
where dp is the pressure differential.
The diagram of leakage QP is shown in Figure
11.

Figure 11 Diagram of Leakage Q
p

where n is the tooth number, LR is a laminar resistor
model [9].
Other leakages due to pressure are similar to Q
pi
,
see [2][4].
The volumetric flow produced by the relative mo-
tion has a uniform distribution from zero to v, where
v is the relative velocity of the tip with respect to the
case. The volumetric flow is defined as [2]:
1
/ 2
wi c
Q h B v =
The interfaces and icons of five types of leakages
are shown in Figure 12.

Figure 12 Icons of the Five Types of Leakages
3.4 Mechanical Parts
Mechanical parts contain the mounting, the bears
and the mesh models. Mounting is used to fix the
bears that support the gear pair. This paper does not
cover the position of the shaft. So we choose a com-
ponent FixedTranslation from Modelica.Mutibody to
fix the two gears. Model IdealGear is used to trans-
mit torque, which does not consider elasticity, damp-
ing or backlash. If we want to consider these effects,
the gears have to be connected to other elements in
an appropriate way.
The diagram of mechanical parts is shown in
Figure 13.

Figure 13 Diagram of the Mechanical Parts
1-Driving gear 2-Driven gear 3-IdealMesh 4-
FixedTranslation 5-flange that transmits a driving
torque
6-flange that transmits the resistance moment 7-
frame, forces on shaft1
Interfaces and icon of the mechanical parts are
shown as follows:

Figure 14 Interfaces and Icon of the Mechanical Parts
Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
425
3.5 Radial Forces
Radial Forces produced by the CVs has been cal-
culated (see section 3.2.3). In the model of CV, radi-
al force is decomposed into x- and y- components,
which are then summed up in the model RadialForce.
Figure 15 shows the diagram of the RadialForce as
well as its interfaces and icon.

Figure 15 Interfaces, Icon and Diagram of Model Radi-
alForce
3.6 Channels from CVs to Inlet/Outlet Volumes
Along with the rotation of the gears, when a CV
comes to the position of CV
i+2
g1
, the channel between
it and the outlet volume opens; when it comes to the
position of CV
i
g1
, this channel becomes to close.
Critical angles, shown in Figure 16, are calculated in
respective models.

Figure 16 Critical Angles of Channel Model Connecting
CV and Outlet Volume
Equations of the critical angles are given as fol-
lows:

1 gap
a u = +

2
u t q =
where is a parameter about the pump dimension.
can be calculated from Figure 16.
The hydraulic diameter of the channel connecting
a CV to the outlet volume is given by:

1 2
,if < <

,otherwise
max
min
d
hd
d
u u
=

Figure 17 Hydraulic-Diameter as a Function of Angle

Interfaces and icons of a channel are shown in
Figure 18.

Figure 18 Interfaces and Icon of a Model connected to
Outlet Volume (High Pressure Chamber)
The channel between a CV and the inlet volume
is similar.
3.7 Viscosity Torque
In the ideal case, Viscosity models can be treated
as dampers. There are two types of viscosity torques
acting on a gear, the radial torque and the axial
torque (see Figure 19). The equations of these tor-
ques are given by:

2
1
/
r c
A r h t v e =

0
2
3
2
2
2 d
4 / (3 ) ,if
a
r
a
r
c
c 0
r r r
h
r h r r
v
t e t
te v
=
=
}


Figure 19 Radial Torques and Axial Torques
Interfaces and icons of the torque models are
shown in Figure 20.
Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
426

Figure 20 Interfaces and Icons of Torque Models
Other resistance moments, for example those
caused by the bears, can be added as a component.
3.8 System Model
Combining all the components above, we obtain a
whole pump as follows:

Figure 21 Diagram of a Whole Pump
The interfaces and icon are shown in Figure 22.

Figure 22 Interfaces and Icon of a Pump Model
4 Simulation Results
We implement the simulation based on the model
shown in Figure 23 to enable the comparison be-
tween our results and the experimental results in [7].

Figure 23 Simulation system diagram
The simulation curves of the volumetric flow rip-
ple as well as the average flow rate (in L/min) is
shown in Figure 24.

Figure 24 Volumetric flow ripple and average flow
rate
The pressure distribution in a tooth space within a
complete rotation cycle is shown in Figure 25.
Note: the pressure distribution in Figure 25 is
very similar to the pressure distribution along the
circle of a gear. But actually they are not the same.

Figure 25 Pressure Distribution in a Tooth Space for a
Complete Rotation Cycle
Order:

( 1e6) cos( )
( 1e6) sin( )
x
y
p P
p P

= +
= +

The pressure distribution above is shown in Fig-
ure 26.

Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
427
Figure 26 (Left) Pressure distribution in a Tooth
Space for a Complete Rotation Cycle
The radius of the unit circle at the bottom is 10
bars.
Figure 27 (Right) Radial Force on Shaft 1 for a Com-
plete Rotation Cycle and its Components in the Coordi-
nation in Figure 5.
By changing the parameters (e.g. G
0
, G
1
shown in
Figure 23), we obtain a set of predictions. Both the
simulation results and the experimental results from
[7] are listed in Figure 28Figure 29. It is clear that
the simulation and the experimental results are very
close.

Figure 28 Dimensionless Characteristics of the
Pump in Steady States

Figure 29 Dimensionless Values of the Volumetric
Efficiency of the pump in Steady Conditions
5 Conclusions and Future Work
We present a new method of modeling the hybrid
external gear pumps based on Modelica and assess it
via simulation. The model takes into account the
flow ripple, pressure distribution, leakages, meshing
conditions, etc. Details of the components that make
up the whole pump are introduced. Comparisons be-
tween the result and the experimental data are given
in the end. Good agreement between them is founded.
From the pressure distribution in the space of a
gear tooth, radial force on shaft can be easily calcu-
lated, on which the motion of the shaft can be simu-
lated based. This is left for future work.
Another weakness of the study is that there is lit-
tle comparison with real applications and measure-
ment data. It is a drawback as some minor malfunc-
tions and parameters probably could not be identified.
Such comparison is also left for future work.
Acknowledgments
The paper is supported by National Nature Science
Foundation of China (No.60874064, No.60704019,
No.60736019), Key Project of National High
Technology Research and Development Program
(2009AA044501).
References
[1] Noah D.Manring, Suresh B.Kasaragadda.
The Theoretical Flow Ripple of an External
Gear Pump. Journal of Dynamic Systems,
Measurement, and Control. Transaction of
the ASME 2003, September: 396-404
[2] Aaron S.Heisler, John J.Moskwa, Frank
J.Fronzak. SIMULATED HELICAL GEAR
PUMP ANALYSIS USING A NEW CFD AP-
PROACH. Fluids Engineering Division Summer
Meeting 2009-78472
[3] Abdul Wahab. ANALYTICAL PREDICTION
TECHNIQUE FOR INTERNAL LEAKAGE IN
AN EXTERNAL GEAR PUMP. Proceedings of
ASME Turbo Expo,GT 2009-59287
[4] R.Castilla, M.Gutes, P.J.Gamez-Monter,
E.Codina. Experimental Study of the Shaft Mo-
tion in the Journal Bearing of a Gear Pump. Jour-
nal of Engineering for Gas Turbines and Power,
2009-052502
[5] R.Castilla, M.Gutes, P.J.Gamez-Monter,
E.Codina. Numerical Analysis of the Shaft Mo-
tion in the Journal Bearing of a Gear Pump. Jour-
nal of Engineering for Gas Turbines and Power,
2010-012504
[6] G.Dalpiaz, G.DElia, E.Mucchi, A.Fernandez der
Rincon. MODELING RUN IN PROCESS IN
EXTERNAL GEAR PUMPS. Proceedings of
ESDA2006-95466
[7] Andrea Vacca, Germano Franzoni and Paolo
Casoli. ON THE ANALYSIS OF EXPERI-
MENTAL DATA FOR EXTERNAL GEAR
MACHINES AND THEIR COMPARISON
Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
428
WITH SIMULATION RESULTS. International
Mechanical Engineering Congress and Exposition,
IMECE2007-422664
[8] Fan-Li Zhou, Li-Ping Chen, Yi-Zhong Wu,
Jian-Wan Ding, Jian-Jun Zhao, Yun-Qing
Zhang. MWorks: a Modern IDE for Model-
ing and Simulation of Multidomain Physical
Systems Based on Modelica. Modelica 2006,
September 4th 5th: 725-732
[9] P.Beater. Modeling and Digital Simulation of
Hydraulic Systems in Design and Engineer-
ing Education using Modelica and HyLib.
Modelica Workshop 2000 Proceedings:33-40
Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011
429