A CFD ANALYSIS OF THE OIL SQUEEZING POWER LOSSES

OF A GEAR PAIR
FRANCO CONCLI, CARLO GORLA
Politecnico di Milano, Mechanical Department, Milan, Italy

ABSTRACT
Efficiency improvement is the new challenge in all fields of design. In this scenario the reduction of power
losses is becoming more and more a main concern also in the design of power transmissions. For this reason
it is important to have some models available in order to quantify the power losses since the design stage. The
power losses of a gear transmission can be subdivided primarily into bearing losses, seal losses, meshing
losses and churning losses [1]. Literature provides some models for the prediction of most of this losses.
Some models, obtained on the basis of experimental tests, can be found in literature which describe the
influence of gear geometric and kinematic parameters on hydraulic losses like for instance those proposed by
Mauz [2], who has concentrated on hydraulic losses, or by Dowson [3], who has concentrated on windage
losses. Nevertheless the authors maintain that a deeper understanding of the physical phenomena responsible
of gear losses is still needed in order to improve existing models and CFD simulation can be an effective
approach for such investigation. Marchesse et al. [4], on the basis of a state of the art on the application of
CFD to gear power losses, applied CFD models to study air-windage losses of gears and have validated their
results by means of experimental tests. Simulations of the churning losses of planetary gearboxes and
squeezing losses have been investigated by the authors in previous works [5, 6]. The authors have also
investigated the oil-windage losses [9] (100% immersion in the lubricant) by mean of CFD simulations and
the results have been compared with the experimental ones available at FZG [10] giving good agreement [9].
In order to improve the understanding of the mechanisms involved in churning losses (dip lubrication with
the presence of a free surface), CFD simulations have been run in order to investigate the effect of the same
parameters, like helix angle, tip diameter, face width, rotational speed and lubricant properties on the
churning power losses generated by a single rotating gear. The numerical results have been compared with
the measurements.
This report is focused on the oil squeezing power losses. This kind of losses is associated with the
compression-expansion process by the meshing teeth. The contraction of the volume at the gear mesh implies
an overpressure that induces a fluid flow primarily in the axial direction and this, for viscous fluids, means
additional power losses and a decrease of the efficiency.
In this work this phenomena has been studied by means of some CFD simulations. The influence of some
operating conditions like the lubricant properties, the rotational speed and the temperature have been studied.
Keywords: Efficiency, gear, power losses, oil squeezing, squeezing, lubrication, CFD

1 INTRODUCTION
Efficiency is becoming more and more a main concern in the design of power transmissions for all
the fields of application and increasing the efficiency of power transmissions can give a significant
contribution to the reduction of pollutant emissions and energy savings. Moreover the reduction of
power losses, which are dissipated in the form of heat, can imply improvements in other aspects
like reliability, related to different thermal regimes, downsizing, weight reduction and system
architecture.
The first step to improve efficiency is to have appropriate models to compare different design
solutions. Even if power transmissions are typically a part of a complex system in which other
parts very often have much bigger power losses, like combustion engines for instance,
nevertheless improvements in their efficiency can determine big savings on a global scale.
Power losses in gear transmissions are given by several contributions like gears, bearings, seals
and other transmission components. As far as gears are concerned, while losses related to gear
meshing have been deeply investigated and reliable models are already available [1], some other
losses as those determined by the interaction with lubricant still need to be studied and modelled.

In particular effects related to oil churning, oil squeezing and windage are still to be investigated in
detail. These losses, in fact, depending on the gearbox configuration and operating conditions, can
amount up to approximately the 4060% of the total losses. Moreover, this source of losses is the
only one in which discrete margins of improvement are still possible, while for other sources, like,
for instance, frictional losses, thanks to the results of many years of studies, the remaining
possibilities of improvement are comparatively lower.
In order to predict these kind of losses, literature provides many publications about the load
dependent power losses, but only few works about load independent power losses. These works
are primarily concerning the churning losses of ordinary gears, the spin power losses and the
losses of the bearings. What is still wanting are appropriate studies to predict the oil
pumping/squeezing power losses. The works on this topic are few and simplified. Strasser [11]
proposed a simplified fluid-dynamic model to predict the oil squeezing power losses in a spur gear
pair. This model takes into account the variation of the gap between the teeth but it approximates
the geometry. Seetharaman and Kahraman [12] proposed a physic based fluid mechanics model to
predict the spin power losses of a gear pair due to oil churning and windage. This model calculates
the losses like the sum of the losses associated to the interactions of individual gears and the fluid
and the power losses due to pumping of the oil at the gear mesh. To predict the oil pumping losses,
Seetharaman and Kahraman developed an analytical model that take into account the variation of
the volume of the gap between the teeth without simplifying the geometry like in the model
proposed by Strasser. These models are able to give results in a very short time but they give only
approximate results. For this reason the authors have investigated this kind of losses by mean of
CFD simulations that are able to correctly solve the velocity and the pressure fields without
simplifications and, therefore, give more information about the phenomena. In this study the
influence on this kind of losses of the oil temperature and of the rotational speed of the gears have
been investigated.
PROBLEM DESCRIPTION
When a gear pair is mating, the volume of the cavity between the teeth is continuously expanding
and contracting. In the first steps of the engagement, the volume of the gap is reducing. This
contraction of the volume implies an overpressure in the gap between the teeth. The lubricant is
therefore squeezed out of the cavity. The oil flux takes place primarily in the axial direction thanks
to the big lateral opened areas. While the gears are rotating, the minimum value for the volume is
reached. From this point, the volume of the gap between the teeth starts a sudden expansion. This
increase of the volume causes a negative pressure in the gap. Due to the negative pressure, the
lubricant flows from the oil bath into the gap. This process is cyclic and, in general, there are
multiple cavities that are squeezed together.

Figure 1: change of the volume of the cavity during the engaging of the gears
Due to the viscous effects of the lubricant, this additional flows caused by the changes in the
volume of the gap between the teeth implies additional losses. The value of this losses is strongly
dependent from the lubrication type. In this first step of the research, a dip lubrication with 100%
oil has been considered. This solution has two main advantages. The first one is that this condition
is the simplest one: a partial immersion, in fact, complicates a lot the problem because it involves
multiphase flows. The second one is that with a 100% lubricant immersion, the losses increase
significantly and it is simpler to calculate them and to appreciate the influence of the different
operating conditions. The aim of this initial study was to understand the effect of the mating gears
on the axial flows, so the full immersion lubrication was considered appropriate at this point of the
research.
2 GEOMETRY
A standard spur gear pair has been used. Pinion and gear have the same dimensions. Table 1
summarized the main parameter of the gears.
Table 1: geometrical parameters of the gear pair

0
[rad]
[rad] 0,349066
a [mm]
28,8
ra [mm]
15,175
ri [mm] 13,4375

r [mm]
z [/]
rb [mm]
m [mm]
b [mm]

14,4
36
13,5315
0,8
15,5

From table 1 it is possible to observe that the adopted gears are small gears, with a working
diameter of 28.8 mm and a tooth high of about 1.74 mm.

Figure 2: Geometrical model

The geometry adopted for this study is shown in figure 2. Only five teeth pro gear have been
modelled. The aim of this work, in fact, was to study in detail the behaviour of the lubricant in the
field of action and not the windage (caused by the interaction of the rotating gears and the oil
bath). Furthermore, in order to reduce the amount of elements for the discretisation and thanks to
the symmetry, only the half of the domain has been modelled.
The domain has been discretized with a swept mesh. This meshing technique consists in creating a
mesh on one side of the region, known as the source side, and then copying the nodes of that mesh,
one element layer at a time, until the final side, known as the target side, is reached. The whole
model has been therefore discretized with triangular prisms. This kind of elements allows a larger
aspect ratio compared with the tetrahedral cells in which it will invariably affects the skewness of
the cell, which is undesirable as it may impede accuracy and convergence. Some mesh refinements
have been performed in order to find the best compromise between the accuracy of the solution
and the computational time. A big problem of the meshing operation is that its necessary to model
the thin oil film that rises between the mating teeth. The thickness of this film has a big influence
on the results in terms of power losses. For this reason some geometries with different film
thickness have been modelled. After a convergence study, the final model had a film thickness of
about 0.5 m. The required mesh is therefore extra fine. The smallest element has a length of
about 0.2 m.

a)
Figure 3: a) Mesh in the mating zone; b)Details of the mesh

b)

3 NUMERICAL MODEL
To simulate the problem, a Finite Volume approach has been used. CFD is based on 5 governing
equations: the conservation equations for mass, momentum (3 equations) and energy. In this
simulations, the energy equation is not activated.

0
[

(1)

T ]

where represent the density, the time, the velocity vector,

gravitational body force and external body forces respectively.

(2)

the pressure, and the

For compressible flows, the unknown are the velocity components and the density while the
pressure is evaluated with a constitutive equation. For incompressible flows, the variables are the
pressure and the velocity components. The solution of the system of equations for incompressible
flows is accomplished by the fact that there are not equations where the pressure is explicity
defined. To calcluate it, the continuity equation is substituted with an equation for the pressure;
with some manipulation the pressure appears like unknown term in the momentum equation.
For this reasons a SIMPLE (Semi Implicit Method for Pressure-Linked Equations) scheme has
been adopted as suggested for flows in closed domains to solve the pressure-velocity-couplig [13].
As boundary conditions for the gear surfaces, a no-slip-wall condition has been selected. For the
lateral sides of the domain the same no-slip-wall conditions has been selected in order to describe
the gearbox case. A zero pressure outlet and a symmetry boundary conditions have been chosen in
order to describe the sides of the gears.
The operating conditions have been reproduced imposing a rigid motion of the boundaries
corresponding to the pinion and the gear. With this boundary motion it is possible to reproduce the
real operating conditions in which gears are meshing. The time step for the transient analysis has
been chosen as dynamic on the base of the Courant number. This allow to increase the calculations
without the risk to lose the convergence.
This boundary motion implies a deformation of the fluid domain. For this reason it is necessary do
update the mesh every time step. In order to do that, a dynamic mesh model has been adopted:
spring-based smoothing + remeshing. In this method, the edges between any two mesh nodes are

idealized as a network of interconnected springs. The initial spacing of the edges before any
boundary motion constitute the equilibrium state of the mesh. A displacement at a given boundary
node will generate a force proportional to the displacement along all the springs connected to the
node. Using Hook's Law, the force on a mesh node can be written as

( )

(6)

where and are the displacements of node and its neighbour ,

is the number of the
neighboring nodes connected to the node and
is the spring constant between node and its
neighbor . The spring constant for the edge connecting nodes and is defined as

(7)

At equilibrium, the net force on a node due to all the springs connected to the node must be zero.
This condition results in an iterative equation such that

(8)

Since displacements are known at the boundaries, Equation (8) is solved using a Jacobi sweep on
all interior nodes. At convergence, the positions are updated such that

where
and
respectively.

(9)

re used to denote the positions at the next time step and the current time step,

On zones with a triangular or tetrahedral mesh, the spring-based smoothing method is normally
used. When the boundary displacement is large compared to the local cell sizes, the cell quality
can deteriorate or the cells can become degenerate. This will invalidate the mesh and
consequently, will lead to convergence problems when the solution is updated to the next time
step. To circumvent this problem, it is possible to locally remesh some cells agglomerate that did
not satisfy the skewness criteria or the maximum and minimum length allowed for the cells. In
order to do that a 2.5D remeshing method has been adopted.

OPERATING CONDITIONS

The influence of the rotational speed and the temperature on the power losses has been
investigated.
The power losses have been calculated like the product between the resistant torque on the driving
shaft and its rotational speed. The calculation of the resistant torque is based on a surface integral
and it is composed of two parts: the first given by the pressure effects and the second by the
viscous effects. In all the simulations the energy equation has not been activated. This choice was
taken in order to speed up the calculation times. The operating temperature was defined trough the
lubricant properties (density and viscosity). Combining three different rotational speeds and three
different temperatures, a 3x3 full factorial design has been followed.
Table 3 shows ranges for the parameters adopted in the simulations. is the operating temperature
in [C], the rotational speed of the planet carrier in [rad/s], the density in [Kg/m] and the
dynamic viscosity in [Pa*s]. A commercial oil with this characteristics can be, for example, the
Klbersynth GH 6-220 Synthetic Gear Oil.
Table 3: parameters adopted for the simulation
(rad/s)
52-105-157

T (C)
40-65-90

5 RESLULTS
Figure 5b shows the contours plot of the pressure on the symmetry plane and figure 5a the
streamlines. It is possible to observe that in the upper cavity, that is decreasing its volume, the
pressure is higher than in the oil bath. In the lower cavity, in turn, the lower pressure depends on
the fact that the volume of the gap is increasing.

a)

b)

Figure 3: a) Velocity streamlines in the symmetry plane; b) Contour plot of the pressure field in
the symmetry plane
Figure 6 to 8 show the results in terms of power loss on the driving wheel for different operating
conditions and for 1 cycle. The mating process, in fact, is cyclic and, therefore, also the resistant
torque necessary to squeeze the oil has a periodic trend.

a)

b)

c)

Figure 6: power losses on the driving shaft; a) 40C 52rad/s; b) 65C 52rad/s; c) 90C
52rad/s

a)

b)

c)

Figure 7: power losses on the driving shaft; a) 40C 105rad/s; b) 65C 105rad/s; c) 90C
105rad/s

a)

b)

c)

Figure 8: power losses on the driving shaft; a) 40C 157rad/s; b) 65C 157rad/s; c) 90C
157rad/s
From the diagrams it can be seen that, as expected, the resistant torque decreases with temperature
while it increases with rotational speed.
Figure 1c shows the case in which the temperature is the highest (90C) while the rotational speed
the lowest (52rad/s). In this conditions also with a 100% dip lubrication the losses are less than 1W
and can be neglected in the computation of the total efficiency of the gearbox (that gears are
designed in order to carry out some kW).
Figure 1a shows the results for the same rotational speed but for the lowest temperature (40C). In
this case the losses increase significantly but remain limited to about 2W.
A more sudden increment is related to an increment of the rotational speed. Figure 2a can be
compared to figure 1a. A doubling of the rotational speed from 52 to 105 rad/s with the same
temperature of 40C induces an increase in the losses up to 4 times. An additional increase of 50%
of the rotational speed up to 157rad/s causes a sudden growth of the loss that arises 21W. Even if
21W can be just a little percentage of the total losses of a geared transmission, this cannot be
neglected.
The previous diagrams shows also that the trend of the losses vs. the angle of rotation in always
the same. Figure 9 shows in detail those trend

Figure 9: power losses versus length of contact

Figure 9 shows the power losses along the contact path. In the position 1 the first teeth pair
(1s+2s), as shown in figure 10a, is already engaged. The second teeth pair (2s+3s) is starting the
contact. In the position 2 of figure 9, the first teeth pair ended the contact (figure 10b).
Between 1 and 2, the mating teeth pairs are therefore 2. In the position 3, another teeth pair
(3s+4d) starts the contact (figure 10c) while the second teeth pair (2s+3s) is already engaged. Like
in position 2, in position 4 a teeth pair (2s+3s) ended the contact (figure 10d).
From this position to position 5, the engaged teeth pair is again only one, like between position 2
and position 3. Finally, in position 5, like in position 3, a teeth pair starts the contact while another
teeth pair is already engaged (figure 10e).

Figure 10: different engaging steps

The trends of the losses in terms of PVZ0,S vs. rotational speed (for different temperatures) and in
terms of PVZ0,S vs. temperature (for different rotational speeds) are shown in figure 10 and figure
11.

Squeezing power losses PVZ0,S [W]

T=40C
T=65C
T=90C

6
5

Klbersynth GH 6-220

4
3
2
1
0
40

60

80

100

120

140

160

Rotational speed [rad/s]

Squeezing power losses PVZ0,S [W]

Figure 11: averaged power losses due to squeezing versus rotational speed for three different
temperatures
7
Omega=52rad/s

Omega=105rad/s

Omega=157rad/s

Klbersynth GH 6-220

3
2
1
0
30

50

70

90

110

Temperature [C]

Figure 11: averaged squeezing power losses versus temperatures for different rotational velocities
From the diagrams appears that the power losses increase with the rotational speed in a more than
linear manner and decrease with temperature approximately linearly.

6 CONCLUSIONS
A preliminary state of the art has put in evidence the weight of the contribution of different
sources of power losses and the availability of reliable prediction model for each of them. Based
on the results of such preliminary investigation, load independent gear losses, mainly determined
by the interaction with the lubricant, has proven to be quantitatively important, to have significant
margins of improvement and to be lacking of reliable prediction models. On the other side, taking
into account the recent developments of CFD (computational fluid dynamics) and the related
hardware and software tools, CFD simulation have been proposed as a possible approach to
investigate such kind of losses.

For this reason a research program dedicated to the investigation of load independent power losses
of gears has conducted and some different aspects of the no-load power losses of gears have been
analyzed. In particular CFD simulations have been performed to analyze the squeezing
phenomena.
The numerical simulations allowed to describe the amount of loss due to pocketing along the
contact path. This can be very helpful for the understanding of the phenomena. Furthermore, the
simulations show the influence on the losses of different parameters: this kind of loss increases, as
expected, with the rotational speed in a more than linear manner and decreases with temperature
approximately linearly. The results show that, for low regimes and high temperature, this kind of
loss are particularly low and can be neglected in the computation of the total efficiency of the
transmission. The decrease of the temperature has less influence on the losses than the increase of
the rotational speed. In order to cut in half the losses an increase of 125% in the temperature (from
40C to 90C) is necessary while a decrease of 50% of the rotational speed (form 157rpm to
105rpm) is enough.
The results shows also that the increase of the losses from the best case (90C and 57rad/s) to the
worst case (40C and 157rad/s) (temperature increase of 125% and rotational speed increase of
150%) can be of about 2000% (from less than 1W up to about 20W). This fact suggests that also
this kind of losses have to be taken into account in order to obtain a good prediction of the
efficiency of a geared transmission.
Future work will be addressed to the study of other geometries and other kinds of lubrications.

References
[1] NIEMANN, G., WINTER, H.: Maschinenelemente Band II Getriebe allgemein,
Zahnradgetriebe Grundlagen, Stirnradgetriebe, Springer-Verlag, Berlin Heidelberg Ney York
Tokio 1983
[2] MAUZ, W.: Hydraulische Verluste von Stirnradgetrieben bei Umfangsgeschwindigkeiten bis
60 m/s, Universitt Stuttgart 1987.
[3] DAWSON, P.H., Windage loss in larger high-speed gears, Proc. Inst. Mech. Eng., 198A (1),
pp.51-59.
[4] MARCHESSE Y., et all. Investigation on CFD simulation for predicting windage power losses
in spur gears, ASME J. Mech. Des.
[5] F. CONCLI, C. GORLA - Computational and experimental analysis of the churning power
losses in an industrial planetary speed reducer - Advances in Fluid Mechanics 2012 2012 - WIT
Transactions on Engineering Sciences - Volume 74, 2012, Pages 287-298
[6] F. CONCLI, C. GORLA - Influence of lubricant temperature, lubricant level and rotational
speed on the churning power losses in an industrial planetary speed reducer: computational and
experimental study - CMEM Vol. 1 No. 2
[7] F. CONCLI, C. GORLA - Oil squeezing power losses of a gear pair: a CFD analysis Advances in Fluid Mechanics 2012 2012 - WIT Transactions on Engineering Sciences - Volume
74, 2012, Pages 37-48
[8] F. CONCLI, C. GORLA - Analysis of the oil squeezing power losses of a spur gear pair by
mean
of
CFD
simulations

ASME
ESDA
Volume
2
(http://www.asme.org/products/proceedings/2012-proceedings-of-the-asme-11th-biennial-con-(1))
[9] C. GORLA, et all. CFD simulations of splash losses of a gearbox, Advanced in tribology,
2012 (http://downloads.hindawi.com/journals/at/2012/616923.pdf)
[10] B.-R. HHN, K.MICHAELIS, H.-P. OTTO: Influence on no-load gear losses, Ecotrib 2011

[11] D. STRASSER, Einfluss des Zahnflanken- und Zahnkopfspieles auf die

Leerlaufverlustleistung von Zahnradgetrieben Dissertation zur Erlangung del Grades DoktorIngenieur Bochum 2005
[12] S. Seetharaman, A. Kahraman, Load-Independent Spin Power Losses of a Spur Gear Pair:
Model Formulation, Journal of Tribology APRIL 2009, Vol. 131
[13] Comini, G., Fondamenti di termofluidodinamica computazionale, SGEditoriali, Padova 2004