Original Article

Thermal and efficiency characterization

of a low-backlash planetary gearbox:
An integrated numerical-analytical
prediction model and
its experimental validation

Proc IMechE Part J:

J Engineering Tribology
0(0) 110
! IMechE 2015
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1350650115622363
pij.sagepub.com

Franco Concli

Abstract
In the automation filed, low-backlash gearboxes are required to guarantee precise positioning. For such kind of applications, planetary speed reducers represent one of the most attractive solutions. This type of gearing ensures at the
same time high power density and reduction ratios. On the other side, the compactness of the solutions leads to high
operating temperatures. For this reason, it is important to be able to quantify the power dissipation and the operating
temperatures already in the design stage, therefore to be able to find the best compromise between the load carrying
capacity and the maximum transmittable power due to thermal limitations. For this reason, an innovative calculation
method capable to quantify the efficiency under different operating conditions and the related operating temperatures
was developed. Experimental tests were performed under different operating conditions to validate the predictions. The
comparison shows good agreement.
Keywords
Gears, efficiency, temperature, computational fluid dynamics, experiments
Date received: 16 June 2015; accepted: 18 November 2015

Introduction
The increasing demand of power transmission capability in more and more reduced spaces represent a
big challenge for the gearbox manufacturers. This fact
is even more severe in the eld of automation where
the miniaturization of the robotic systems is unrestrainable. In the eld of packaging, one of the most
appreciated solution for the torque and speed conversion is the planetary gearing. In the most widely used
conguration, such kind of kinematic consists in two
gears (one external gear called sun and an internal one
called ring-gear) mounted concentric. Additional
gears called planets engage with both the sun and
the ring-gear. The planets have, unlike the sun gear
that has a pure rotation, a rototranslating motion
because the ring-gear does not rotate. The planets
are mounted on a rigid structure called planet-carrier
that is able to transform their rototranslation into a
pure rotation of the output shaft. Depending on the
number of teeth of the gear, the total reduction ratio
can vary a lot making this solution protable for various applications. In planetary gearboxes, unlike in
xed axis gear systems, additional power losses are
induced by the roto-translation of the planets.

Furthermore, due to the high power density, the

heat-exchange area is reduced and high temperatures
can arise limiting the possible operating eld.
Being able to quantify the power dissipation and to
predict the operating temperature allows to nd out
the best design compromise between the load carrying
capacity and the maximum transmittable power due
to thermal limitations.
For this purpose, Bonglioli Mechatronic Research
(BMR) has developed a new hybrid calculation method
based on both analytical relations and numerical
results. The model was rst validated with the data
obtained with dedicated measurements on a real gearbox and successively systematically applied for the
complete characterization of the operating eld of the
gearbox.

Reseach & Development Dept, Bonfiglioli Mechatronic Research, Italy

Corresponding author:
Franco Concli, Bonfiglioli Mechatronic Research, via Fortunato Zeni, 8,
36068 Rovereto (Tn), Italy.
Email: [email protected]

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

Proc IMechE Part J: J Engineering Tribology 0(0)

Power loss sharing

causes power dissipation.

The mostly wide accepted power loss classication1

subdivides the losses (PL ) according to the mechanical
components such as gears, bearings, seals, and
other components like clutches or synchronizers, if
present (subscripts G, B, S, and X, respectively).
Furthermore, the losses are subdivided into loaddependent and load-independent (subscript 0).
It should be pointed out that also the so-called loadindependent losses that are basically related to the
interaction with the lubricant and the sliding of
the seals are indirectly related to the transmitted
torque that induces a change in the operating temperature that, in turn, produces a change in the lubricant
properties. Despite this little discrepancy, in the following, this nomenclature will be accepted and used.
PL PLG PLG0 PLB PLB0 PLS0 PLX

PLG FR 
wi FR  g

The actual force FR tangent to the teeth ank can be

expressed as the product of a function fL and the force
tangential to the pitch circle Fbt . fL takes into account
the force repartition due to the teeth stiness variation
during the contact and the actual number of engaged
gears. For 14" 42, fL can be approximated as
reported in Figure 1(b). Under this hypothesis, the
mean gear meshing power losses can be expressed as,

PLG

Fbt

pb

Z

Di

i fL,i g,i dx
Ai

Ft m t
pb cosw

Ei1

i1 fL,i1 g,i1 dx

Di1

Ei

fL,i 
Ai

g,i
dx P  m  Hv
t

Load dependent power loss of gears

(gear meshing losses) PLG
The gear meshing losses are generated in the contact
between the gear anks due to relative sliding and
rolling. Except when the contact point P coincide
with the pitch point C (Figure 1(a) and (b)), in fact,
the gear anks velocities have dierent tangential
components.
This produce sliding between the teeth. In a
loaded contact and in presence of friction, this

It seems that the power loss due to the sliding between

the teeth is directly proportional to the transmitted
power. This is true as long as the frictional coecient
is assumed as constant. If the frictional coecient is a
function of the temperature, the relation between the
load dependent power losses of gears and the transmitted power is no more linear. In the present study,
for the estimation of the frictional coecient, the relation proposed by ISO standard2 was used (4) even if
many other formulations are available like those

Figure 1. (a) Speed in different positions along the line of action; (b) velocities, friction coefficient, and force repartition trends along
the line of contacts; and (c) instantaneous gear mesh power loss.

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

Concli

proposed by ISO6336,3 Michaelis et al.,4 and Hohn

and Michaelis.5


Ft = cost =b
 0:048
2  t  sinwt  c

0:2

0:05
 Ra0:25
oil

in which Ft is the force at pitch circle, t the transverse

pressure angle, wt the working pressure angle, oil the
dynamic viscosity of the oil at operating temperature,
 the equivalent radius of curvature at the pitch point
of contact, and Ra the arithmetic averaged roughness.
The above-described model can be applied also to
planetary gears providing the relative velocities. The
mechanical power losses due to rolling are related to
the shear stress in the EHL (elastohydrodynamic lubrication) lm. Even if literature provides detailed models
for the prediction of such losses like those proposed
by Talbot and Kahraman6 and Li and Kahraman,7
according to the ndings of Thirumurugan and
Muthuveerappan8 and Andersson9 the rolling power
losses are of an order of magnitude lower than the
sliding power losses. For this reason, this contribution
was neglected in this analysis.

Load independent power losses

of gears PLG0
Small planetary gearboxes used in the automatic
industry are almost always dip lubricated. The interaction of the gears with the oil involves in any case
losses, but in planetary reducers, the phenomenon is
more severe due the presence of the planet-carrier and
the rototranslation of the planets. Literature provides
many dierent empirical models for the prediction of
this kind of losses for simple rotating gears such as
those proposed by Ohlendorf,10 Dawson,11 and
Mauz.12 Dierent authors studied also the lubrication
of a gear pair13 observing that the trend of the losses is
signicantly aected by many dierent operating parameters. A huge amount of dierent equations were
provided also by other authors,1417 but since these
models are derived from experiments, they are applicable or give reliable results only as far as the actual
geometry and operating conditions are similar to that
used in the experiments. To overcome this problem,
Seetharaman and Kahraman18 proposed a physicsbased uid mechanics model able to predict spin
power losses of gear pairs due to oil churning and
windage.
More recently, dierent authors have applied single
phase computational uid dynamics (CFD) simulations to overcome this limitation.1928 The author
has already experience with both multiphase gear
simulations29 and with the modeling of planetary systems.3032 CFD simulations seems to provide extremely accurate results not only for ordinary gears but
also for the complex motion of planetary gears.
On this basis, the author has adapted an open
source CFD code released under the GNU license

for the simulation of planetary gears.33 The code

relies on the numerical solution of some equilibrium
equations representing conservation laws of physics
such as mass (5) and momentum (6) conservation.
In addition, the presence of two phases implies to
include an additional transport equation of a scalar
quantity  that represents the volume fraction of one
of the two phases.
@hui i
0
@hxi i
@hui i @hui ihui i
@hpi


@t
@xj
@t
 


@
@hui i @huj i
@ij



@xj
@xj @xi
@xj

The above equations have been solved numerically

adopting Gauss scheme. The domain for the simulation is represented by the internal volume of the planetary gearbox. From the author experience, the main
contribution to the PLG0 losses is given by the planetcarrier rotation and the related motion of the planets.
For this reason, to speed up the calculation and to
reduce the computational eort, the sun-gear and the
ring-gear have been modeled without teeth. The eectiveness of such assumption was already validated in
previous works.2527 Even if in such assumption
implies that the oil squeezing losses (oil pocketing)
are neglected, it is known that this phenomena assumes
importance only in case of injection lubrication while
becomes negligible on case of splash lubrication.
With this simplication, it is possible to reproduce
the motion of the planet-carrier without topological
changes in the mesh and, therefore, run transient
simulations without the need of remeshing. All the
boundaries have been modeled as no-slip boundaries
(rp 0, v vwall ). Figure 2 shows the analyzed
planetary speed reducer and example of the mesh
adopted for the calculations.
Parametric analysis have been performed to nd
out the best compromise between the results accuracy
and the computational eort. The nal meshes
adopted have approximately 1 M cells for the smallest
gearbox size and 3 M cells for the biggest size. In both
cases, the domain was meshed using the blockMesh
and the snappyHexMesh utilities of OpenFOAM .33
To achieve temporal accuracy and numerical stability
when running the simulations, a Courant number of 1
was adopted. The Courant number is dened as,
Co

tjUj
x

where t is the time step, jUj is the magnitude of the

velocity through the cell, and x is the cell size in
the direction of the velocity. The time step is therefore
determined considering the smaller cell size in the
mesh and the maximum value in the velocity eld.

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

Proc IMechE Part J: J Engineering Tribology 0(0)

Figure 2. (a) Section view of the analyzed planetary gearbox; (b) boundary corresponding to the housing, the bearing, and the sun
gear; and (c) boundaries corresponding to the planet carrier and to the planets.

The Reynolds unresolved terms have been neglected

and the model considered laminar.
To post-process the results, a dedicated utility
(written in C) was developed. This utility, starting
from the velocity and pressure elds and from the
mixture properties (calculated for each cell i with the
CFD code), calculates the resistant torque (viscous 
and inertial (p) contribution) on the gear/planet carrier due to the presence of the lubricant mixture. The
basic relation on which the utility is based are

TLG0


X
Ui

mixi mixi
 Ai  ri
xi
i

TLG0p

pi  A i  r i

where the subscript i refers to the ith cell, mixi and

mixi are the viscosity and the density of the mixture in
the cell calculated as an averaged mean value of the
properties of the dierent phases, Ai is the area of the
cell corresponding to the boundary, and ri is the radial
distance of the cell from the axis.
Simulations were performed for all the sizes of
gearbox and for dierent rotational speeds, static
lubricant levels and temperatures so to have a complete characterization of the gearbox behavior in any
operating conditions. Figure 3(a) represents an example of load independent power loss map calculated
with the CFD approach. Figure 3(b) represents an
example of lubricant distribution and velocities
inside the gearbox. This capability of the code ensures
the possibility not only to calculate the power dissipation but also to get additional information about
the lubrication inside the gearbox.
The simulations were performed on a 108GFLOP
workstation. The characterization of the whole
domain has taken approximately 6 weeks that is a
signicant amount of time but less than the standard

time required to manufacture and test physical prototypes. As described in General eciency model section, once the operating eld was characterized, the
results have been interpolated to have analytical equations directly applicable in the design practice to simulate the behavior of the gearbox under specic
working conditions.

Load power loss of bearings PLB and PLB0

As for the gears, also the bearing power losses can
be subdivided into load-dependent and loadindependent. The load dependent power loss depends
on the sliding between the bearing elements such as
rings and rolling elements or between the journal and
the bearing in journal bearings. The load independent
power losses, in turn, depends from the interaction
with the lubricant (the oil that lubricates the gears
or grease applied directly on the bearing) and from
the sliding between the rings and the bearing seals if
present. Bearing manufacturers have many decades of
experience in this led and have already presented
reliable prediction models.34,35 For this reason, no
further research and validation test were made on
bearings and the SKF model was applied for the calculation of the bearing power losses.

Load independent power loss of seals PLS0

The seals losses arise due to the relative sliding
between the shafts and the seals themselves. The magnitude is dependent from the rotational speed, the
friction coecient, and the contact pressures. One of
the most accepted model to predict such contribution
is that proposed by Niemann and Winter.1 The power
dissipation can be calculated according to (10).
PLS0 lL  d2 

n
1000

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

10

Concli

Figure 3. (a) Example of load independent gear power loss map obtained with 12 CFD simulations; (b) lubricant-air interface inside
the gearbox and velocity contours on different planes; and (c) seal power losses.

where lL is a coecient that describes the operating conditions and varies with temperature and wear of the seal,
d is the seal diameter, and n the rotational speed. This
equation gives the results shown in Figure 3(c).

General efficiency model

The above-described models allow the calculation of
the power losses only if the lubricant temperature is a
priori known. It is not possible to calculate this temperature directly, but it is possible to rate its iteration.
The total power loss PL is in fact dissipated in the
surrounding environment by heat exchange. This phenomenon is governed by the general heat transfer
equation (14).
Q A   
T PL

11

in which Q is the total transferred heat, A the exchange

Kg
surface,  the heat transfer coecient Ks
3 , and
T the
temperature dierence between the surface and the surrounding environment. The heat transfer coecient
was established experimentally with dierent tests in
which both the power losses and the operating temKg
peratures were measured and results in 24.7 Ks
3 . This
value is consistent with the suggested values provided

by literature. It appears that by increasing the operating temperature, the heat ux increases linearly while
the power losses decrease monotonically. This consideration ensures the possibility to nd the equilibrium between the dissipated power PL and the
removed heat Q with an iterative procedure.
By integrating the prediction models together and
by implementing an iterative algorithm for the calculation of the operating temperature (that implies a new
calculation of the power loss, etc.), it is possible to completely characterize the gearbox in terms of mean operating temperature and eciency starting only from
the geometrical data and the operating conditions.
For this purpose Scilab,36 an open source software
for numerical computation, was used. All the above
presented equations were implemented in the code.
Regarding the CFD results, all the simulated conditions were collected in a database and an interpolating
equations for each gearbox geometry was derived and
included in the code. In this way, simulating the operating temperature and eciency takes less than 20 s.
For new geometries for which CFD simulations
should be performed, the computational time required
increases but remains in any case less than 1 week for
the characterization of the whole operating eld (1 h
for each simulation), ensuring results in a shorter way
with respect to experimental tests in which the test rig

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

Proc IMechE Part J: J Engineering Tribology 0(0)

setup and the tests take approximately 1 month (1 day

for each measurement).

Model validation
Once the model was implemented, dedicated experimental tests were performed to validate the method. The
measurements were performed on an energy-closed-loop

test rig, and its conguration is schematically described

in Figure 4. A permanent magnet servomotor (9.) controlled by a servoinverter is connected to the tested
gearbox (6.). The housing of the gearbox is mounted
on a ange. The output shaft of the gearbox is connected to a second gearbox used as multiplier (5.).
Another permanent magnet servo motor (10.) controlled by a second servoinverter is connected to the

Figure 4. (a) Schematic layout of the test rig and (b) test rig.

Table 1. Test summary.

Test #

!in (rpm)

T1 (Nm)

L (ml)

Test #

!in (rpm)

T1 (Nm)

L (ml)

1
2
3
4
5
6

3000
3000
3000
3000
3000
3000

16
0
16
0
16
0

130
130
93
93
60
60

7
8
9
10
11
12

1000
1000
1000
1000
1000
1000

16
0
16
0
16
0

130
130
93
93
60
60

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

Concli

slave gearbox and operate as generator. Torque measurement shafts (4. and 7.) at the input and output of
the tested gearbox measure the transmitted torque
(0.05% F.S.) and rotational speed to determine the
transmitted power. The output torque is feedback controlled with the measurement of the torque meter (7.),
the input speed is feedback controlled with the motor

encoder. The shafts are connected with metal bellows

couplings.
The gearbox surface and ambient temperatures are
measured by ve sensors. To validate the calculations
for the whole operating eld, tests were performed at
dierent speeds, with dierent static oil levels, with
and without load (Table 1). For the tests without

Figure 5. Comparison between the measured (Exp) and the calculated (Calc) power losses; the values above the bars represent the
temperatures.

Figure 6. (a) Measured and (b) calculated power loss maps (T1 16 Nm); (c) measured and (d) calculated temperatures maps
(T1 16 Nm).

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

Proc IMechE Part J: J Engineering Tribology 0(0)

Figure 7. (a) Measured and (b) calculated load independent power loss maps (T2 0 Nm); (c) measured and (d) calculated temperatures maps (T2 0 Nm).

load, the coupling between the gearbox (9.) and the

torque-meter (7.) was disengaged, so that the power
that enters the gearbox is totally dissipated: even if the
output shaft rotates, it does not transmit torque and,
therefore, power.

Results
Figure 5 shows a comparison between the calculated
values and the experimental results. The stripes represent the no-load working condition, whereas the solid
color represents the loaded conditions. The columns
with the error bars represent the experimental measured values. The bars corresponding to the calculated
values are subdivided to show the shear between the
losses. Concerning the calculated values, the colors
represent the losses generated by the bearings, the
seals, the meshing losses, and the load independent
losses of the gears starting from the bottom of the
column. Above the bars, the measured and predicted
temperatures are reported. It appears that the predicted values are compatible with the measurements.
The dierence between the experimental and the computational data is averaged 1% (max 17%) both in
terms of power losses and operating temperature
(max 10%).
Figures 6 and 7 show a comparison in terms of
power loss (a and b) and temperature maps (c and
d), obtained by interpolating the measured (a and c)

and the calculated values (b and d) for the loaded

(Figure 6) and the unloaded (Figure 7) condition.
The maps show that a reduction of the amount of
lubricant has a positive impact on the power loss especially for high rotational speeds, where the PLG0 contribution is more signicant.
While there is a direct relation between the calculated temperature and power loss maps, the measurements show some uctuation of the results that
should be attributed to the measurement uncertainty
and to the fact that even if the testing room was airconditioned, some minor variation in the environment
temperature was present and this has aected the
operating temperatures of the gearbox.
Despite that, the maps (measured and calculated)
are very similar both for the loaded and unloaded
case, conrming the accuracy of the prediction
model that will be included in the company standard
design practice allowing to characterize the so-called
thermal limit already in the design stage (Figure 8).

Conclusions
The goal of this research was to nd a model capable
to predict the eciency and the operating temperature
of small planetary gearboxes. Some models available
in literature for the gear meshing, the bearing, and
the seal losses and some additional equations for the
load independent power losses of gears were used and

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

Concli

Figure 8. GUI developed to handle with the iterative global efficiency model: top-left: input data; top-right: convergence monitor;
bottom: results.

iteratively solved to estimate the operating temperature. The additional equations were obtained by interpolating the results of CFD numerical simulations
generated with a specically developed tool. The
results of the global eciency model were validated
with experimental measurements showing good
agreement.
This approach was successively used to characterize the complete operating eld of the gearbox. It has
allowed to reduce signicantly the eort needed to
generate the eciency maps and the thermal characterization of the whole gearbox gamma avoiding the
need of testing each size and each gear ratio in several
operating conditions. Because of the GNU license, in
fact, the calculation was signicantly parallelized on
many CPUs allowing to perform the whole analysis
on a standard 108GFLOPS workstation in less than 1
week for each size. Moreover, the adoption of an
open-source tool gives the possibility to customize
the code for the specic industrial needs: a further
step will be the overcoming of the geometrical simplications introduced in the CFD model and the adoption of such tool to optimize the internal shape and
the lubricant circulation in the gearbox.

Declaration of conflicting interests

The author(s) declared no potential conicts of interest with
respect to the research, authorship, and/or publication of
this article.

Funding
The author(s) received no nancial support for the research,
authorship, and/or publication of this article.

References
1. Niemann G and Winter H. Maschinenelemente Band 2:
Getriebe Allgemein, Zahnradgetriebe Grundlagen,
Stirnradgetriebe 2. Springer, Berlin: Auflage, 2003.
2. ISO/TR 14179-2:2001 Gears Thermal capacity, Figure 4.
3. ISO 6336-4:1996 Calculation of the load capacity of spur
and helical gears.
4. Michaelis K, Hohn B-R and Vollmer T. Thermal rating
of gear drivesBalance between power loss and heat
dissipation. AGMA, 1996.
5. Hohn B-R and Michaelis K. Optimization of gearbox
efficiency. Goriva i Maziva 2009; 48: 441480.
6. Talbot D and Kahraman A. Prediction of mechanical
power loss of planet gear roller bearings under combined
radial and moment loading. ASME J Mech Des 2013;
135: 121007.

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015

10

Proc IMechE Part J: J Engineering Tribology 0(0)

7. Li S and Kahraman A. Prediction of spur gear

mechanical power losses using a transient
elastohydrodynamic lubrication model. Tribol Trans
2010; 53: 554563.
8. Thirumurugan R and Muthuveerappan G. Study on
mesh power losses in high contact ratio (HCR) gear
drives. In: 16th National Conference on Machines and
Mechanisms (iNaCoMM2013), Roorkee, India,
pp.169176.
9. Andersson M. Churning losses and efficiency in gearboxes. Stockholm: KTH, 2014.
10. Ohlendorf H. Verlustleistung und Erwarmung von
Stirnradern. Munich: SAGE Journals, 1962.
11. Dawson PH. Windage loss in larger high-speed gears.
Proc IMechE 1984; 198: 5159.
12. Mauz W. Hydraulische Verluste von Stirnradgetrieben
bei Umfangsgeschwindigkeiten bis 60 m/s. Stuttgart,
Germany: Universitat Stuttgart, 1987.
13. Ariura Y and Ueno T. Lubricant churning loss and its
behavior in gear box in cylindrical gear systems. J Jpn
Soc Lubr Eng 1975; 20: 67175.
14. Terekhov AS. Hydraulic losses in gearboxes with oil
immersion. Russian Eng J 1975; 55: 711.
15. Boness RJ. Churning losses of discs and gears running
partially submerged in oil. In: Proceedings of the
International Power Transmission and Gearing
Conference: New Technologies for Power Transmissions
of the 90s. Design Engineering Division, ASME,
Chicago, Ill, USA, 1989, pp.355359.
16. Hohn RB, Michaelis K and Vollmer T. Thermal rating
of gear drives balance between power losses and heat
dissipation. Alexandria, VA, USA: AGMA FTM, 1996.
17. Luke P and Olver A. A study of churning losses
in dip-lubricated spur gears. Proc IMechE, Part G:
J Aerospace Engineering 1999; 213: 337346.
18. Seetharaman S and Kahraman A. Load-independent
spin power losses of a spur gear pair: Model formulation. J Tribol 2009; 131: 022201 (11 pp.).
19. Marchesse Y, Changenet C, Ville F, et al. Investigation
on CFD simulations for predicting windage power
losses on spur gears. J Mech Des 2011; 133. Article
number 024501.
20. Chaari F, Romdhane MB, Baccar W, et al. Windage
power loss in spur gear sets. WSEAS Trans Appl Theor
Mech 2012; 2: 159168.
21. Hill MJ, Kunz RF, Medvitz RB, et al. CFD analysis of
gear windage losses: Validation and parametric aerodynamic studies. J Fluid Eng 2011; 133.
22. Gorla C, Concli F, Stahl K, et al. Hydraulic losses of
gearbox: CFD analysis and experiments. Tribol Int
2013; 66: 337344.

23. Gorla C, Concli F, Stahl K, et al. CFD simulation of

splash losses of a gearbox. Adv Tribol 2012. Article ID
616923, 10 pp. http://dx.doi.org/10.1155/2012/616923.
24. Concli F, Gorla C, Della Torre A, et al. Windage power
losses of ordinary gears: Different CFD approaches
aimed to the reduction of the computational effort.
Lubricants 2014; 2: 162176.
25. Concli F and Gorla C. Analysis of the oil squeezing
power losses of a spur gear pair by mean of CFD simulations. In: ASME 2012 11th Biennial Conference on
Engineering Systems Design and Analysis, ESDA2012,
Nantes, France, 2012, pp.177184.
26. Concli F and Gorla C. A CFD analysis of the oil
squeezing power losses of a gear pair. Computational
and experimental study. Int J Comput Methods Exp
Meas 2014; 2: 157167.
27. Concli F and Gorla C. Oil squeezing power losses in
gears: A CFD analysis. WIT Trans Eng Sci 2012; 74:
3748.
28. Gorla C, Concli F, Stahl K, et al. Load independent
power losses of ordinary gears: Numerical and experimental analysis. In: 5th World Tribology Congress,
WTC 2013, Torino, Italy, 2013, pp.12431246.
29. Concli F, Gorla C, Della Torre A, et al. Churning
power losses of ordinary gears: A new approach based
on the internal fluid dynamics simulations. Lubr Sci
2014.
30. Concli F, Gorla C and Conrado E. Analysis of power
losses in an industrial planetary speed reducer:
Measurements and computational fluid dynamics calculations. Proc IMechE, Part J: J Engineering Tribology
2014; 228: 1121.
31. Concli F and Gorla C. Computational and experimental analysis of the churning power losses in an industrial
planetary speed reducer. WIT Trans Eng Sci 2012; 74:
287298.
32. Concli F and Gorla C. Influence of lubricant temperature, lubricant level and rotational speed on the churning
power loss in an industrial planetary speed reducer:
Computational and experimental study. Int J Comput
Methods Exp Meas 2013; 1: 353366.
33. OpenFOAM, http://www.openfoam.com (accessed 1
December 2015).
34. INA/FAG, Walzlagerkatalog, Schaeffler Technologies
GmbH & Co. KG, 2010.
35. SKF, Catalogo Generale, 6000 IT, SKF, 2006.
36. Scilab, http://www.scilab.org (accessed 1 December 2015).

Downloaded from pij.sagepub.com by Ramiro Martins on December 18, 2015