IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

LOAD SHARING ANALYSIS OF PLANETARY GEAR BOX

Sagar B Malkapure1, Wadkar S B2

1
M.E Student, Mechanical, Sinhgad College of Engineering Pune, Maharashtra, India
2
Professor, Mechanical, Sinhgad College of Engineering Pune, Maharashtra, India

Abstract
One of the advantages of epicyclic transmissions is that the input torque is divided in a number of parallel paths. For n planet
epicyclic system, each sun–planet–ring path is to transmit 1/n of the input torque. However, this is only true in the ideal case
when there is equal load sharing between all the planets in the epicyclic system. Because of manufacturing errors, equal load
sharing is not possible and the degree of inequality in load sharing has major role for gear system sizing, tolerancing schemes,
and torque ratings. Therefore it is important to understand the fundamental cause of the unequal load sharing behavior in
epicyclic gear sets. Load sharing behavior is associated with positional errors causing one or more planets to lead or lag the
other planets. When the error is positive the planet with error lead the other planet while when the error is negative planet with
error lags the other planet. Several manufacturing errors can introduce positional errors. Some of the common contributors are
carrier pinhole position error, planet size variation, and run out of the gears. A set of precision planetary gear set is selected with
the objectives of experimental and theoretical investigation of the load sharing behavior by introducing the position errors for the
planets, the load shared by each planet is estimated by maximum stresses induced in each pin. The results of the experimentation
are then validated with the FEM (Finite Element method) results.

Keywords: Epicyclic gear train, Pinhole position error, Pin stresses, Load sharing
--------------------------------------------------------------------***----------------------------------------------------------------------

1. INTRODUCTION behavior are observed. The experimental data are compared

to the predictions of a state-of-the-art multibody contact
The epicyclic gear train (EGT) is known as planetary gear analysis model - Gear System Analysis Modules (GSAM)
train (PGT). Epicyclic gear train is a gearing system [1]. Ajit Bodas et. al. presented a state of the art contact
consisting of one or more ring gears, or planet gears, mechanics model of planetary gear set to study the effect of
revolving about a central, or sun gear. The planet gears are a number of manufacturing and assembly related carrier and
mounted on a movable arm or carrier which itself may rotate gear errors on the load sharing amongst the planets .Three
relative to the sun gear. Epicyclic gearing systems also different group of errors are considered: (i) time-invariant,
incorporate the use of ring gear or annulus, which meshes assembly independent errors such as carrier planet pinhole
with the planet gears. Planetary gears are classified as position errors, (ii) time variant, assembly dependent errors
simple and compound planetary gears. Compared to simple such as planet tooth thickness errors, and (iii) time variant,
planetary gears, compound planetary gears have the assembly-dependent errors such as gear run out errors[9].
advantages of larger reduction ratio, higher torque-to-weight B.Boguski et. al. proposed a new method of measuring
ratio. In many epicyclic gearing systems, one of these three planet load sharing of planetary gear sets. The method uses
basic components is held stationary; one of the two strain gauges mounted directly on the planet pins to measure
remaining components is an input, providing power to the continuously the loads carried by the planets assembled in a
system, while the last component is an output, receiving fixed carrier [13].This paper aimed at experimental and
power from the system. The ratio of input rotation to output theoretical investigation of the load sharing behavior by
rotation is dependent upon the number of teeth in each gear, introducing the position errors for the planets and the load
and which component is held stationary. Epicyclic gearing is shared by each planet is estimated by maximum stresses
less expensive, when tooled properly. Epicyclic gear sets are induced in each pin. The results of the experimentation are
smaller than offset gear sets, having high reduction ratio, then validated with the Ansys results.
high radial loads on output shaft. This makes gear lighter,
more compact, more efficient and less noisy. The need of
1.1 Force Distribution in Epicyclic Gear Train
light weight construction and resources results into gearbox
designs with high load capacity and power density with high The force distribution in complete epicyclic gear train is as
expectations for reliability of the gear. Additional there is a shown in Fig-1 in which the input torque is applied on the
diversity of planetary gears for different application cases. sun gear in clockwise direction causes the sun gear to rotate
H.Ligata et. al. presented results of a comprehensive about its own centre, while planet gears are revolving about
experimental and theoretical study to determine the the sun gear in anticlockwise direction or vice versa when
influence of certain key factors in planetary transmissions the ring gear is fixed. The carrier is incorporated to hold the
on gear stresses and planetary load sharing. Pinion position planet gear through pins and produce output of the gear box
errors are introduced as a representative key manufacturing in anticlockwise direction.
tolerance, and the resultant changes in the planetary

_______________________________________________________________________________________
Volume: 03 Issue: 08 | Aug-2014, Available @ http://www.ijret.org 150
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

applying tensile force on it. The complete frame is attached

to the concrete foundation with the help of foundation bolts
to give rigidity and damp the vibration produced by the gear
box. This set up was intended to measure the input and
output torque. In the present experiment this is not the aim,
therefore the system is modified for the static load
condition.

Fig -1: Simple epicyclic gear train

1.2 Free Body Diagram of Complete Epicyclic Gear

Train
Fig- 2 shows the free body diagram of complete epicyclic
gear train in which the input torque is applied to the sun gear
and this torque is resolve into two component namely
tangential force and radial force. In which the tangential Fig -3: Experimental setup
force causes bending stresses on the gear tooth and radial
force produces the compressive stresses on the gear. The sun 2.1 Mounting of Strain Gauge on Pin
gear is engaged to the three planet gears which are at 120°
apart from each other. The planet gear meshes with the sun A strain gauge (Micro Measurements BF-350 3AA(11)No-
gear and ring gear which produces two tangential forces, F) is mounted on pins as shown in Fig- 4 to measure strains
therefore twice the tangential force is require to hold the due to the bending of the pins as a result of the radial loads
planet gear. carried by the planet bearings and planets. Conditioned
strain signals from each gauge are converted to stress values
and summed to find the total of pin bending stresses such
that the percentage represented by a gauge of this total stress
value represented the load carried by that planet.
Mathematically, the load sharing factor
𝜎𝑛
𝐿𝑆𝐹 = 𝑛 𝜎 ...... 1
𝑖=1 𝑛

Where, 𝜎 = individual pin stress

Fig -2: F.B.D of epicyclic gear train

2. EXPERIMENTAL SETUP
Epicyclic gear box used for experimentation contains sun
gear, three planet gears and a ring gear. The sun gear is
input and the carrier is output. The epicyclic gear box is
driven by the 2.2 kw motor having maximum 3000 rpm. The
flexible coupling is used in between output shaft of motor
and input shaft of epicyclic gear box. Brake drum
dynamometer is attached on the output shaft of epicyclic
gear box and two S type load cells are used to measure Fig -4: Mounting of strain gauge on pin
torque with a two decimal accuracy. In this one load cell is
fixed to the frame and other is attached to the screw for

_______________________________________________________________________________________
Volume: 03 Issue: 08 | Aug-2014, Available @ http://www.ijret.org 151
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

2.2 Strain Gauge Calibration Setup

Strain indicators are used to display the voltage difference
but in this case strain indicator needs to calibrate to obtain
desired output. Strain indicators are calibrated to display the
stresses induced in the test specimen. In this case, the strain
gauges are mounted on the cantilever beam having
rectangular cross section. This cantilever beam is fixed at
the one end and the other end is free. The pan is attached on
the free end to apply the weight as shown in Fig.6.1.The full
bridge circuit is used to increase the accuracy of the
experimental results. The calibration is carried out with
reference to the theoretical results. As the 1 kg of load
applied on the free end, the theoretical value comes out to be
30.62 N/mm2, this value is inserted in to strain indicator to
replace the voltage difference value. After calibration the Fig -6: Loading arrangement
strain indicators are used to display stress value irrespective
of the test specimen and the applied weight. 3. PIN STRESS ANALYSIS
3.1 Pin Stresses with No Error
The values of the pin stresses are calculated by using FEM
Software as shown in Fig -7. Table -1,Table -2 ,Table -3 and
Table -4 gives the comparison of theoretical values of the
pin stresses with the experimental values of the pin stresses
and the FEM pin stresses for no error on the pins and
percentage load sharing by the pins respectively. The
distance between application of force on the lever and center
of the shaft is 595 mm.(i.e. a leverage is used to apply the
torque to the gears).

Load case 1
Applied torque = 2918.47 N-mm

Fig -5: Strain gauge calibration setup

2.3 Loading Arrangement

For static analysis, loading arrangement is incorporated to
apply the input torque to the sun gear by changing the
weight, instead of providing the input torque from the
motor. The loading of gear box is made with the help of a
lever of 595mm in length on one side of which a pan is
attached for the purpose of addition of weights, while the
other end is having a key slot which can be fitted in the
input shaft of the gear box. This results in providing the
input torque to the sun gear. Three strain indicators are used
to display the strains at three positions of strain gauges on
three different pins.
Fig -7: Software results for the pin stress

Table -1: Comparison of Theoretical and Experimental pin

stresses
Applied Theoretical Experimental Stress (N/mm2)
Torque Stress
(N-mm) (N/mm2 ) Pin 1 Pin 2 Pin 3
2918.47 1.439 1.38 1.23 1.44
5836.95 2.878 2.90 2.73 3.02
8755.42 4.317 4.55 4.17 4.70

_______________________________________________________________________________________
Volume: 03 Issue: 08 | Aug-2014, Available @ http://www.ijret.org 152
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

11673.90 5.756 5.44 5.13 5.62

14592.37 7.196 7.34 7.18 7.80
17510.85 8.653 8.56 8.51 8.88

Table -2: Comparison of Theoretical and FEM pin stresses

Applied Theoretical FEM Stress (N/mm2)
Torque Stress
(N-mm) (N/mm2 ) Pin 1 Pin 2 Pin 3
2918.47 1.439 1.43 1.33 1.55
5836.95 2.878 2.86 2.67 3.11
8755.42 4.317 4.29 4.29 4.66
Chart -1: Percentage load sharing by F.E.M Versus Torque
11673.90 5.756 5.72 5.72 6.22
Chart -1 shows the percentage load sharing which is
14592.37 7.196 7.15 7.15 7.77 calculated from the results of the pin stresses by finite
element method for different values of the torque. Pin 3
17510.85 8.653 8.88 8.58 9.33 shows more load than pin 1 and pin 2. Pin 2 shows
minimum load.
Table -3: Percentage Load Sharing (Experimental) for
different values of Torque 36.00
% of Load Sharing (EXPERIMENTAL)

Applied % Load Sharing (Experimental) 35.00

Torque (N- 34.00
mm) Pin 1 Pin 2 Pin 3 33.00
32.00
2918.47 34.07 30.37 35.56
31.00
5836.95 33.53 31.56 34.91 30.00 Pin1
29.00 Pin2
8755.42 33.90 31.07 35.02 28.00
27.00 Pin3
11673.90 33.60 31.69 34.71

14592.37 32.89 32.17 34.95

17510.85 32.99 32.79 34.22

Torque (Nmm)

Table -4: Percentage Load Sharing (F.E.M) for different Chart -2: Experimental percentage load sharing Versus
values of Torque Torque
Applied % Load Sharing (F.E.M)
Torque (N- Chart -2 shows the percentage load sharing which is
mm) Pin 1 Pin 2 Pin 3 calculated from the results of the pin stresses from
2918.47 33.18 30.86 35.96 experimentation for different values of the torque. Load
sharing (L.S) for pin 3 decreases with increasing torque. Pin
5836.95 33.10 30.90 36.00 2 shows tendency of more load sharing at high torque. Load
sharing of pin 1 decrease with high torque.
8755.42 33.10 30.94 35.96
From Chart -1 and Chart -2 it can be concluded that actual
11673.90 33.08 30.94 35.97 load sharing is different than the theoretical because of
manufacturing errors and other parameters.
14592.37 33.09 30.96 35.96
17510.85 33.08 30.96 35.97

_______________________________________________________________________________________
Volume: 03 Issue: 08 | Aug-2014, Available @ http://www.ijret.org 153
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

10
9
8
Stress (N/mm2)

7 F.E.M.Pin1
6
5 F.E.M.Pin2
4
F.E.M.Pin3
3
2 Theoretical
1
0 Exp. pin1
11673.9
2918.47
5836.95
8755.42

14592.37
17510.85

Exp. pin2
Exp. pin3
Fig -9: Software results for planet pin 2 for error of 100µm
Torque (Nmm)

Chart -3: Comparison of the theoretical pin stresses with

F.E.M and Experimental pin stresses Versus Torque

Chart -3 shows the comparison of the theoretical pin stresses

with F.E.M and experimental pin stresses for the different
values of torque. From the figure it is clear that with
increase in the torque the pin stress increases. The
theoretical and experimental result match closely with the
finite element method results

3.2 Pin Stresses with Error

Fig -8 shows the maximum value of the von – mises stresses
generated in the planet pins when an error of 100 µm is
provided on the pin 1 for the load case 1 of the Table -5
Fig -10: Software results for planet pin 3 for error of 100µm
Load case 1
Applied torque = 2918.47 N-mm Table -5: Comparison of Theoretical and Experimental pin
stresses (with error on pin 1)
Applied Theoretical Experimental Stress (N/mm2)
Torque Stress
(N-mm) (N/mm2 ) Pin 1 Pin 2 Pin 3
2918.47 1.461 1.54 1.33 1.50

5836.95 2.922 3.03 2.72 2.88

8755.42 4.382 4.58 4.06 4.42

11673.90 5.844 6.06 5.30 5.70

14592.37 7.305 7.58 6.76 7.20

17510.85 8.766 9.02 8.06 8.56

Fig -8: Software results for planet pin 1 for error of 100µm

Fig -9 and Fig -10 shows the stresses generated in planet pin Table -6: Comparison of Theoretical and FEM pin stresses
2 and pin 3. Applied Theoretical FEM Stress (N/mm2)
Torque Stress
(N-mm) (N/mm2 ) Pin 1 Pin 2 Pin 3
2918.47 1.461 1.50 1.34 1.43

5836.95 2.922 3.01 2.68 2.86

_______________________________________________________________________________________
Volume: 03 Issue: 08 | Aug-2014, Available @ http://www.ijret.org 154
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

8755.42 4.382 4.52 4.02 4.30 Chart 4 shows the comparison of the percentage load
sharing by experimental results with that of F.E.M results
11673.90 5.844 6.03 5.36 5.73 for different values for torque. From the figure it is clear that
the pin no.1 (with an error of 100 µm) is highly stressed
14592.37 7.305 7.54 6.71 7.17 with the higher percentage of load sharing.
17510.85 8.766 9.05 8.05 8.60
4. CONCLUSIONS
Table -7: Percentage Load Sharing (Experimental) for
 The stresses in the planet pins increases with the
different values of Torque (with error on pin 1)
increase in the torque.
Applied % Load Sharing (Experimental)  Manufacturing errors and planet pin-hole position
Torque (N- errors prevents the equal load sharing in the planetary
mm) Pin 1 Pin 2 Pin 3 gear box.
2918.47 35.24 30.43 34.32  When an error of 100µm is provided on the pin of the
planet, the planet pin with error takes the more load
5836.95 35.11 31.51 33.47 as compared to the remaining planets.
 The pin stresses calculated theoretically are in good
8755.42 35.06 31.08 33.84 agreement with experimental results and FEA results.
11673.90 35.52 31.06 33.41 REFERENCES
14592.37 35.19 31.38 33.42 [1]. H.Ligata“Internal Gear Strains and Load Sharing in
Planetary Transmissions: Model and Experiments”, SAE
17510.85 35.17 31.43 33.38 Paper No. 700178, (1970) pp. 655 - 656.
[2]. A. Singh, “Load sharing behavior in epicyclic gears -
Table -8: Percentage Load Sharing (F.E.M) for different physical explanation and generalized formulation”,
values of Torque (with error on pin1) Mechanism and Machine Theory 45 (No. 3) (2010) pp. 511
– 530.
Applied % Load Sharing (F.E.M) [3]. A. Kahraman, “Static Load Sharing Characteristics of
Torque (N- Transmission Planetary Gear Sets: Model and
mm) Pin 1 Pin 2 Pin 3 Experiment”,SAE Paper No. 1999-01 pp. 1050 - 1060.
2918.47 35.20 31.32 33.47 [4]. A. Bodas, A. Kahraman, “Influence of carrier and gear
manufacturing errors on the static load sharing behavior of
5836.95 35.20 31.32 33.47 planetary gear sets”, JSME InternationalJournal Series C 47
(2004) pp 908–915.
8755.42 35.20 31.32 33.47
BIOGRAPHIES
11673.90 35.20 31.32 33.47
Malkapure Sagar .B Student, M.E
14592.37 35.20 31.32 33.47 Mechanical Design Engineering,
Sinhgad College of Engineering, Pune
17510.85 35.20 31.32 33.47

36
% of load Sharing

35 Prof.(Dr.) S. B. Wadkar
34
33 Experiment Professor , Mechanical Engineering
32 al Pin 1 Department, Sinhgad College of
31 Experiment Engineering ,Pune
30 al Pin 2
29 Experiment
al Pin 3
2918.47
5836.95
8755.42
11673.9
14592.37
17510.85

Torque (Nmm)

Fig -14: Comparison of the percentage load sharing by

experimental results and F.E.M results versus Torque

_______________________________________________________________________________________
Volume: 03 Issue: 08 | Aug-2014, Available @ http://www.ijret.org 155