Research ArticleDynamic Analysis of Helical Planetary Gear Sets underCombined Force and Moment Loading
Yanfang Liu, Junbin Lai, Peng Dong, and Xiangyang Xu
Beijing Key Laboratory for High Efficient Transmission and System Control of New Energy Resource Vehicle,School of Transportation Science and Engineering, Beihang University, Beijing 100191, China
Correspondence should be addressed to Peng Dong; [email protected]
Received 1 August 2017; Accepted 5 November 2017; Published 22 November 2017
Academic Editor: Tai Thai
Copyright Β© 2017 Yanfang Liu et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The dynamic behavior of a single-stage planetary gear set with helical gears of multishaft automotive automatic transmissions hasbeen studied, in which one component of the planetary gear set is imposed by additional external vertical and axial loading fromcountershaft gear pair in addition to the moment. Under these combined loading conditions, the contributions of the deflections ofthe ring gear and the carrier cannot be neglected. A three-dimensional nonlinear time-variant dynamicmodel considering not onlythe transverse, torsional, axial, and rotational motions of the gears but also the elasticity of the mounted shafts has been developedby combining the lumped parametermethodwith finite elementmethod.The naturalmodes and the forced vibration responses dueto static transmission errors have been obtained.The proposed dynamic model is employed to describe the effects of the combinedexternal loading condition and positioning on the dynamic behavior of a four-planet system.
1. Introduction
Planetary gear trains (PGTs) are the core component of thepowertrain transmission systems used in many aerospace,vessel, automotive, constructionmachinery, and wind-powerapplications. Under dynamic conditions, the forces actingon gears are amplified, which would lead to larger dynamicloads and stresses and potentially reduce the fatigue life ofgears. Furthermore, high-frequency alternating forces dueto the dynamic behavior are transmitted to the radiatingsurfaces as excitation for gear whine. Therefore, dynamicanalysis is essential in describing the vibration and noise andeven durability of any geared power transmission systems.A dynamic model for predicting vibration amplitudes anddynamic forces is necessary to evaluate the noise and durabil-ity of a given gear system,which is helpful to identify potentialproblems and find capable solutions.
Lumped parameter method is widely used to develop thedynamic model of PGTs, in which the PGT is simplifiedas a topology structure by assuming each component asrigid and the connection through kinematical joints, springs,and dampers. Various models have been proposed by manyresearchers over the past decades.With respect to the number
of freedoms, there are pure torsional vibration models [1],two-dimensional models [2], and three-dimensional models[3]. In terms of considering the stiffness changes and backlashor not, there are linear or nonlinear and time-invariant ortime-variant dynamic models [1, 4β6]. Those models cananalyze the vibration characteristics, such as natural modesand forced vibration responses. Another unique behavior,namely, modulation sidebands for PGTs, is studied by a fewresearchers. Inalpolat and Kahraman [4] predicted modu-lation sidebands of a simple-stage planetary gear train byintroducing the influence of manufacturing errors on thegear transmission error excitations into a nonlinear time-variant dynamic model. The inflexibility assumption of thegear system in the lumped parameter method deviates fromthe real elasticmechanical structure. Kubur et al. [7] proposeda dynamic model combining a three-dimensional discretedynamicmodel of counter-shaft helical gear pairs with a finiteelement model of shaft structures. Abousleiman et al. [8]proposed a hybrid model for spur and helical gear planetarydrives inwhich the contributions of the deflections of the ringgear and the carrier are introduced via substructures derivedfrom3Dfinite elementmethodswhile the sun gear is assumedto be rigid.
HindawiShock and VibrationVolume 2017, Article ID 4635204, 13 pageshttps://doi.org/10.1155/2017/4635204
2 Shock and Vibration
A planetary gear set isconnected with a gearpair to transfer power
5
2
1st axis
B
CD
AF
E
2nd axis
Input
Output Output
(a) ZF 9-speed AT [9]
1st axis
2nd axis
Input
Output Output
C1C3B1C4
C2C5
1
2
11
5
8
4
7
6
10
3
9
PGS1
PGS2
PGS3
0
A planetary gear set isconnected with a gearpair to transfer power
(b) ShengRui 8-speed AT [10]
Figure 1: Schemes of two FWD AT examples with parallel-axis layouts.
Furthermore, there are many new developed multispeedautomatic transmission layouts in which planetary gear setsand gear pairs are assembled in parallel axes in recent tenyears. Figure 1 shows two examples of this kind of structuralschemes. This kind of structural layout is well suitablefor front-wheel drive (FWD) vehicles. Only gear pairs aremounted on the counter-shaft as shown in Figure 1(a) whileboth planetary gear sets and gear pairs are mounted on thecountershaft as presented in Figure 1(b). One gear of a gearpair is usually fixedly connected with one component of aplanetary gear set. For example, the driving gear of the gearpair 5-2 in Figure 1(a) is fixedly connected with the carrierof the planetary gear set on the far left through shaft, andthe driven gear of the gear pair 6-10 in Figure 1(b) is fixedlyconnected with the ring gear of the planetary gear set onthe left side of the counter-shaft; the details are shown inthe dotted box in Figure 1. When the power is transmittedthrough gear pairs, due to the gear tooth meshing forces,additional vertical forces and even axial forces are generatedand transmitted to the supporting shafts of the gears. Whenone gear is connected with one component of a planetarygear set, the vertical and/or axial forces will be shared by theplanetary gear set too, whose force environment is differentfrom the traditional concentric automatic transmission basedon multiple planetary gear sets. As for this kind of structure,the shafts of planetary gear sets imposed by vertical and/oraxial forces in addition to torque would have much moredeflection and thus influence the dynamic behavior of thegear system. So it is necessary to investigate the dynamiccharacteristics of this of kind gear system.
Without loss of generality, a single-stage PGT imposedby combined external vertical forces and/or axial forces in
addition to the moment is introduced in the following. Thisstudy aims at developing a dynamic model of this kind ofPGTs considering the flexibility of all components. As thefocus is not only on the gears but also on other components,the model must have the capability of including the mostrelevant design parameters of gears, shafts, and bearings toquantify the direct influence of each of these parameterson the overall dynamic behavior. Based on the model, thedynamic behavior of PGTs under additional external verticalforce and/or axial forces will be predicted and analyzed.Limited parameter studies will be performed to describe theeffect of some of the key system parameters on the dynamicbehavior. The intent of this model is to help designersobtain themost favorable configuration for themost desirabledynamic behavior at any stage of the development of the gearsystem.
2. Dynamic Model
2.1. Physical Model and Assumptions. A single-stage plane-tary gear train (PGT) considered in this study consists of a sungear and a ring gear which are coupled to each other by ππplanets. Figure 2 illustrates a PGT with four planets,ππ = 4,in which the ring gear and carrier are the input and output,respectively, while the sun gear is stationary. The ring gearload includes the vertical πΉπ¦ and axial forces πΉπ§ in additionto the moment ππ = π ππΉπ₯, where π π is base circle radius ofthe ring gear and πΉπ₯ is the circumferential force. The planetsare mounted on a carrier through bearings and pins, and thecarrier is supported flexibly in all directions. The dynamicmodel of the system shown in Figure 2 employs the followingassumptions:
Shock and Vibration 3
L1
FyFxFz
Planet 4
Planet 1
Planet 2
Planet 3
Sun(s)
Carrier (c)
Ring(r)s
rp
c
Bearing(b2)Bearing(b1)
Planet Pin
Figure 2: A typical PGTwith four equally spaced planet gears undercombined loading.
(i) The gear contact is treated as a solid-solid contactproblem, that is, the elastohydrodynamic lubricationproblem is not included in this study.Thenonlinearityis due to the teeth separation, which will simplify theequations of motion.
(ii) The elastic deformations of gear bodies are assumedto be negligible.
(iii) The effects of the inertias connected to the PGT atthe input and output sides are not included here con-sidering that these can only cause resonances at lowfrequencies.
(iv) Frictional forces caused by tooth sliding motions areassumed to be negligible.
The coupled three-dimensional (transverse-rotational-axial) model must be general so that a PGT with any numberof planets can be analyzed, in which the planets can be spacedequally or unequally around the sun gear at any phasingrelationship, and the sun and ring gear and carrier can befloating in the radial direction.
For a single-stage planetary gear system withππ planets,each planet is in mesh with the sun gear (π ) and the ring gear
(π) while it is supported through its axis by a bearing held bythe carrier (π). So any planet branch located at an arbitraryplanet position angle ππ can be considered as a combinationof three subsystems: a sun and planet gear pair, a ring andplanet gear pair and a carrier and planet gear pair, shown inFigure 3.
The dynamicmodel of a PGT allows each element includ-ing gears and the carrier to translate in π₯, π¦, and π§ (axial)directions, denoted by π₯π, π¦π, and π§π, and rotate aboutthese translation axes, denoted by ππ₯π, ππ¦π, and ππ§π (π =π , π, π, π1, . . . , πππ). For the sake of convenience, define threevariables in place of the rotations with π€π₯π = π πππ₯π, π€π¦π =π πππ¦π, and π€π§π = π πππ§π, where π π is the base circle radius of agear or the radius of the circle passing through planet centersof the carrier. Gears aremodeled as rigid diskswithmassesππand mass moments of inertias πΌπ. So a mass matrix mπ anda displacement vector qπ are then defined for each elementπ as
mπ = diag [ππ ππ ππ πΌππ 2ππΌππ 2π
πΌππ 2π ] ,qπ = {π₯π π¦π π§π π€π₯π π€π¦π π€π§π}π .
(1)
Before obtaining the dynamic model of the entire gearsystem, the equations of motion of these three subsystemswill be formulated first based on lumped parameter method.Figure 3(a) illustrates an external helical gear pair, that is,the sun gear π meshing with the planet gear-π located at anarbitrary angle ππ. Figure 3(b) illustrates an internal helicalgear pair, that is, the ring gear πmeshing with the same planetgear-π. Figure 3(c) illustrates the model of the carrier π andthe same planet gear-π pair.2.2. Model Formulation. The equations of motion of a singlecentral gear-π (π = π, π ) and planet gear-π (π = 1, . . . , ππ) pairare given as
[mπ 00 mπ
]{ qπqππ} + βπππππ [[
PπππPππ βPπππPππβPπππPππ PπππPππ
]]{ qπqππ} = { fπ
fπππ} , (2a)
Pπ π = {cosπ½ cosππ π β cosπ½ sinππ π β sinπ½ sinπ½ cosππ π β sinπ½ sinππ π cosπ½} , (2b)
Pππ = {cosπ½ cosπππ β cosπ½ sinπππ sinπ½ sinπ½ cosπππ β sinπ½ sinπππ β cosπ½} , (2c)
fπ =
{{{{{{{{{{{{{{{{{{{{{{{{{{{{{
00000ππ (πππ π )
}}}}}}}}}}}}}}}}}}}}}}}}}}}}}
+ βπ πππ πππ π
{{{{{{{{{{{{{{{{{{{{{{{{{{{
cosπ½ cosππ πβ cosπ½ sinππ π
β sinπ½β sinπ½ cosππ πsinπ½ sinππ πβ cosπ½
}}}}}}}}}}}}}}}}}}}}}}}}}}}
, (2d)
4 Shock and Vibration
SunPlanet-k
zs
zs
ys
ys
zpk zpk
espk
kspk
xsxs
k
xpk
xpk
ypk
ypk
(a)
Planet-kRing
zr
zr
yr
yr
zpkzpk
xrxr
ypk
ypk
k
xpkxpk
krpk
erpk
(b)
Planet-k
Carrier
zpk zpk
ypk
yc
yczc
zc
xc
xc
k
ypk
xpkxpk
(c)
Figure 3: Dynamic models of (a) a sun and planet gear pair, (b) a ring and planet gear pair, and (c) a carrier and planet gear pair.
fπ ππ = βπ πππ πππ π
{{{{{{{{{{{{{{{{{{{{{{{
β cosπ½ cosππ πcosπ½ sinππ π
sinπ½β sinπ½ cosππ πsinπ½ sinππ πβ cosπ½
}}}}}}}}}}}}}}}}}}}}}}}
, (2e)
fπ =
{{{{{{{{{{{{{{{{{{{{{{{{{{{
00000ππ(πππ π)
}}}}}}}}}}}}}}}}}}}}}}}}}}}
+ βπππππππππ π
{{{{{{{{{{{{{{{{{{{{{{{
cosπ½ cosπππβ cosπ½ sinπππsinπ½
sinπ½ cosπππβ sinπ½ sinπππβ cosπ½
}}}}}}}}}}}}}}}}}}}}}}}
, (2f)
fπππ = βfπ, (2g)
where π½ is the helical angle; ππ is the mean torque applied togear-π; πππ = ππ β πΏπππΌπ, where πΌπ is the transverse operating pressure angle and πΏππ = 1 for external gear pair and πΏππ = β1
for internal gear pair; πππ is themeshing stiffness of gear-π and
Shock and Vibration 5
planet gear-π; βππ = 0 if the teeth lose contact and otherwiseβππ = 1 according to the assumption of solid-solid contact; πππis the static transmission error (TE) excitation function. Andthe relative gear mesh displacement along the line of actionπππ is given as
πππ = [Pππ βPππ] [ qπqππ] β πππ. (3)
The planet is attached to the carrier through a pin and abearing. Since the planet gear is free to rotate, the torsionalmotion of the planet gear is uncoupled from the carriermotions. The relative displacement along all axis, denoted by
ππ₯ππ, ππ¦ππ, and ππ§ππ, and rotation around these axis, denotedby πππ₯ππ and πππ¦ππ, are given as
ππ₯ππ = π₯π β π₯π + π€π§π sinππ, (4a)
ππ¦ππ = π¦π β π¦π β π€π§π cosππ, (4b)
ππ§ππ = π§π β π§π + π€π¦π cosππ β π€π₯π sinππ, (4c)
πππ₯ππ = ππ₯π β ππ₯π, (4d)
πππ¦ππ = ππ¦π β ππ¦π. (4e)
The pin-bearing assembly is represented by the stiffnessππ₯, ππ¦, ππ§, πππ₯, and πππ¦. So the equations of motion of thecarrier π and planet gear-π pair are given as
[mπ 00 mππ
]{ qπqππ} + πππ [ Kπ11π βKπ12π
βKππ12π Kπ22π]{ qπ
qππ} = {fπ
0} , (5a)
Kπ11π =
[[[[[[[[[[[[[[[
ππ¦ 0 0 0 0 βππ¦ cosππ0 ππ₯ 0 0 0 ππ₯ sinππ0 0 ππ§ ππ§ cosππ βππ§ sinππ 00 0 ππ§ cosππ πππ¦π 2π + ππ§ cos
2 ππ βππ§ cosππ sinππ 00 0 βππ§ sinππ βππ§ cosππ sinππ πππ₯π 2π + ππ§ sin
2 ππ 0βππ¦ cosππ ππ₯ sinππ 0 0 0 ππ₯ sin2 ππ + ππ¦ cos2 ππ
]]]]]]]]]]]]]]]
, (5b)
Kπ12π =
[[[[[[[[[[[[[[
βππ¦ 0 0 0 0 00 βππ₯ 0 0 0 00 0 βππ§ 0 0 00 0 βππ§ cosππ βπππ¦π πππ 0 00 0 ππ§ sinππ 0 βπππ₯π πππ 0
ππ¦ cosππ βππ₯ sinππ 0 0 0 0
]]]]]]]]]]]]]]
, (5c)
fπ =
{{{{{{{{{{{{{{{{{{{{{{{{{
00000ππ(πππ π)
}}}}}}}}}}}}}}}}}}}}}}}}}
, (5d)
Kπ22π = diag [ππ¦ ππ₯ ππ§ πππ¦π 2π
πππ₯π 2π
0] . (5e)
In addition to the gear pairs, the elastic shafts andbearings also influence the dynamic characteristics of the geartransmission. Here, the housing holding some bearings isassumed rigid. Considering that the system hasππ bearings,
the overall bearing stiffness matrix Kπ can be written interms of individual bearing stiffness Kππ (π = 1, 2, . . . , ππ) asKπ = diag [Kπ1 Kπ2 β β β Kπππ]. Themodel of shafts is devel-oped with the finite element method (FEM). The stiffness
6 Shock and Vibration
and mass matrices of each finite shaft elements π (π =1, 2, . . . , ππ π) of each individual shaft π (π = 1, 2, . . . , ππ ) canbe calculated to be a stiffness Kπ π and mass Mπ π matrices of6Vπ (Vπ = ππ π+1) dimension. Assembling all shaft stiffnessKπ and mass matricesMπ of the system of 6βππ 1 Vπ dimensions:
Kπ = diag [Kπ 1 Kπ 2 β β β Kπ ππ ] ,Mπ = diag [Mπ 1 Mπ 2 β β β Mπ ππ ] . (6)
As a gear pair connects one node on one shaft to the othernode on the next shaft according to (2a) and (5a), overall gearmass and stiffness matrices can be assembled with those ofshafts and bearings. So the mass and stiffness matrices of thesystem are given asM = Mπ +Mπ and K = Kπ + Kπ + Kπ.
The damping matrix C is not derived to this point inthe formulation. That is because the damping values are notknown in most cases even in the final stages of gear design.Therefore, for practical engineering purpose, a set of constantdamping values are used in the forces response calculationsbased on the experience on damping of gear meshes. Here,2.5 percent modal damping is used when solving the forcedresponse.
Based on above equations, the three-dimensional dy-namic model of the overall single-state planetary gear systemwithππ planets can be written in matrix form as
MX (π‘) + CX (π‘) + KX (π‘) = F. (7)
2.3. Gear ContactModel. In order to analyze the forced vibra-tions with above dynamic models, the excitations such asthe gear transmission error and stiffness and manufacturingerrors should be defined up-front through the loaded gearteeth contact analysis. A load distribution model consid-ering the conditions of compatibility and equilibrium wasinitially proposed by Conry and Serieg [11] based on Hertzmechanics and developed by the researchers from the gearand powertrain lab of Ohio State University [12]. Giventhe tooth compliance, applied torque, and the initial toothseparation, elastic deformations at any point of the gearsurface are computed by using amodified Simplex algorithm.Yue et al. [13] analyzed the teeth contact characteristics ofhelical gear pairs with a finite element model and proposeda computational method based on genetic algorithm tooptimize gear profile parameters to reduce the transmissionerror, which is used here to calculate the excitations, whosedetailed modeling and simulation was introduced in [14] andnot discussed in this study.
The mesh stiffness due to tooth bending, base rotation,and shear deformations is computed by using simplified finiteelement formulations as a function of gear rotation. Thetime-varying transmission error and mesh stiffness of eachexternal and internal gear pair are predicted as a periodicfunction at the mesh frequency of the gear mesh, written as
ππ π (π‘) = πΏβπ=1
π(π)π π sin (ππππ‘ + πππ ππ + π(π)π π ) , (8a)
πππ (π‘) = πΏβπ=1
π(π)ππ sin (ππππ‘ + πππππ + ππΎπ π + π(π)ππ) , (8b)
ππ π (π‘) = ππ π + πΏβπ=1
π(π)π π sin (ππππ‘ + πππ ππ + π(π)π π + ππΎπ‘π ) , (9a)
πππ (π‘)= πππ+ πΏβπ=1
π(π)ππ sin (ππππ‘ + πππππ + ππΎπ π + π(π)ππ + ππΎπ‘π ) .(9b)
Here, π(π)ππ
and π(π)ππ
(π = π , π) are the amplitude and phaseangle of the πth harmonic term; πππ and π(π)ππ (π = π , π) arethe average and alternatingmesh stiffness amplitude; the gearmesh frequency ππ is defined as ππ = ππ |ππ β ππ| =ππ|ππ β ππ|, where ππ and ππ are absolute angular velocityand teeth number of gear π and ππ is the absolute angularvelocity of the carrier; πΏ is the total number of harmonic termconsidered; πΎπ π is the phase angle between ππ π(π‘) and πππ(π‘); πΎπ‘π is the phase angle between time-varying transmission errorand mesh stiffness.
3. An Example Analysis
A 4-planet planetary gear set in Figure 2 is used here as anapplication example, where the sun is stationary; the ring gearand the carrier are the input and output, respectively. Theexternal vertical and axial loads in addition to the momenttransmitted from a counter-shaft gear pair are imposed on thering gear, and a design parameterπΏ1 is defined by the distancebetween the middle plane of the counter-shaft gear pair andthe middle plane of the ring bearing, πΏ1 (0 β€ πΏ1 β€ πΏ0)where the maximum πΏ0 is determined by the widths of thegears and the bearing. One part of the imposed load is finallyburdened by the bearing of the ring gear, and the other partis burdened by the ring gear and finally transferred to thesun gear through planet gears.The load shares depend on thevalue of πΏ1 according to Newtonβs equilibrium law; that is,there is the minimum load share on the ring gear at πΏ1 = 0while the maximum is at πΏ1 = πΏ0.
Table 1 lists the basic gear parameters of the example geartrain. The planets are equally spaced but not in phase [15]since ππ/ππ = integer for equally spaced planets. Meanwhile,the dynamic mesh forces are sequentially phased; that is,sum of the phase angles is an integer multiple of π. Thiscase is one of the most common conditions used in productapplications because of its lower vibration and noise levels[16]. Table 2 lists the input moment and forces of the ringgear for gear contact model, which corresponds to about 80%of the related maximum moment of this planetary gear setin an actual automatic transmission. To be convenient forcomparison, the traditional driving way is named as βpuremoment loadingβwhile the drivingway in this study is namedas βcombined loadingβ since the driving source includes notonly the moment but also the vertical and axial forces.
Shock and Vibration 7
Aver
age m
esh
stiffn
ess e
rror
(%)
r-p1 r-p2 r-p3 r-p4Gear pair
L1 = L0
L1 = L0/2
L1 = 0
β80
β60
β40
β20
0
(a)
s-p1 s-p2 s-p3 s-p4Gear pair
L1 = L0
L1 = L0/2
L1 = 0
β3.0
β2.0
β1.0
0.0
1.0
Aver
age m
esh
stiffn
ess e
rror
(%)
(b)
Figure 4: The influence of the design parameter πΏ1 on average mesh stiffness (a) ring-planet gear pairs and (b) sun-planet gear pairs.
Table 1: Basic design parameters of the example gear train.
Parameter Sun Ring PlanetNumber of teeth 23 73 24Module (mm) 1.47Pressure angle (β) 17.5Helical angle (β) 15.5Face width (mm) 18 20 19Root diameter (mm) 31.25 114.61 32.45Outside diameter (mm) 39.59 40.79Minor diameter (mm) β 106.81 βTooth thickness (mm) 2.58 2.67 2.48Backlash (πm) 74
Table 2:The imposed combined forces and moment loading on thering gear.
πΉπ₯ (N) πΉπ¦ (N) πΉπ§ (N) ππ = π πΉπ₯ (Nm)4267 1405 1387 385
Table 3 shows the results of the power flow analysis. Itis observed from these results that imposed vertical andaxial forces will lead to obvious disparity of mesh forcesand stiffness between planets, that is, the uniform loadsharing between planets is deteriorated. With the conditionof combined loading, the additional imposed vertical andaxial forces at different position on the ring gear wouldchange average gear mesh stiffness while they would notchange the gear mesh forces since there is different elasticshaft deflection at different position even with the samevertical and axial forces. Figure 4 shows the relative deviationof calculated average mesh stiffness under the conditionof combined loading by comparing with the pure momentloading condition, which has the following features:
Mes
h sti
ffnes
s (N
/m
)
0
200
400
600
800
1 2 30Mesh cycle
Pure torqueL1 = 0
L1 = L0/2
L1 = L0
Figure 5: Time-varying mesh stiffness of ring-planet 1 underdifferent loading conditions.
(i) The average mesh stiffness of each gear pair is differ-ent when imposing the additional vertical and axialforces, not similar to those under the condition ofpure torque loading.Themaximum change is for gearpairs with planet 1 whose radial position is collinear tothe additional vertical force direction and away fromthe imposed force position.
(ii) The average mesh stiffness of ring-planet gear pairs isreduced seriously when imposing additional verticaland axial forces on the ring gear. The maximumreduction is for the planet 1 at πΏ1 = πΏ0. The detailedtime-varying mesh stiffness of ring-planet 1 is pre-sented in Figure 5.The reduction of the average meshstiffness is tremendous for ring-planet gear pairs andmild for sun-planet gear pairs. Here, the ring gear isthe driving gear and so the kinetic transfer path from
8 Shock and Vibration
Table 3: Static gear mesh forces and average gear mesh stiffness values.
Pure torque loading Combined loadingMesh force
(N)Mesh stiffness
(N/πm)Mesh force
(N)Mesh stiffness (N/πm)
πΏ1 = 0 πΏ1 = πΏ0/2 πΏ1 = πΏ0s-p1 1697.2 396.9 1505.5 395.4 394.1 386.8s-p2 1699.9 396.9 1814.1 397.6 397.0 397.0s-p3 1697.2 396.9 1882.4 398.2 397.6 398.2s-p4 1699.9 396.9 1587.5 396.1 396.7 397.1r-p1 1769.5 581.4 1569.5 313.9 231.8 193.4r-p2 1772.3 581.4 1891.3 287.4 294.7 297.3r-p3 1769.5 581.4 1962.5 385.3 290.7 256.2r-p4 1772.3 581.4 1655.0 254.2 268.4 277.8Note. The symbol s-p1 represents the sun-planet 1 gear pair and the like.
ring gear to sun gear through planet gears can weakenthe changes of average mesh stiffness of sun-planetgear pairs.
(iii) With the increasing of the design parameter πΏ1,starting at 0 to πΏ0, the average mesh stiffness isreduced for all gear pairs.
Figure 6 shows an example of the load distribution ofplanet 3 with different design parameter πΏ1. It is obviousthat there is less distributed load on planet 3 when thereis minimal load share on the ring gear; that is, πΏ1 = 0.Meanwhile, the bias load distribution on meshing gears isfound.The imposed vertical loadwould lead to the bending ofshafts of the planetary gear set and so destroy the uniformityof load distribution.
The undamped natural frequencies ππ predicted by themodel are listed in Table 4. It is observed that the elasticmodelmodifies the original resultsπR of the rigidmodel.Themaximummodification (ππ β πR)/πR is about Β±10% for loworders such as below order 10th, and less than β20% for highorders such as order 20th, shown in Figure 7(a). The relativeerror to the result at πΏ1 = 0, calculated as (ππ β πππΏ1=0)/πππΏ1=0,is shown in Figure 7(b). It is observed that nonzero designparameter πΏ1 will reduce natural frequencies for low orders,such as before 16 orders here, due to the cantilever effect.
The amplitudes, π(π)π π
and π(π)ππ, of the first five mesh har-
monics (π β [1, πΏ]) of the static transmission error excitationsare given in Table 5 for pure moment loading and combinedloading conditions. Here, phase angles, π(π)
π πand π(π)
ππof each
harmonic are assumed to be zero. The steady-state responseX of the planetary gear set to the force vector F defined by(2a) is predicted.Thedynamic force on each gearmesh is thenobtained using the expression
πΉππ (π‘) = πππ (π‘) οΏ½οΏ½ππ (π‘) + βπππππ (π‘) πππ (π‘)(π = π , π π = 1, . . . , ππ) . (10)
The proposed dynamic model is used to analyze theforced vibration responses at pure and combined loadingconditions. Figures 8 and 9 show a set of curves of dynamic
Table 4: Predicted natural frequencies of the example gear trainsystem.
Rigid shaft πR (Hz) Elastic shaft ππ (Hz)πΏ1 = 0 πΏ1 = πΏ0/2 πΏ1 = πΏ02.9 2.8 2.8 2.9468.8 447.6 472.2 459.6622.2 584.9 612.3 592.1715.5 695.5 789.2 772.9978.6 962.9 978.9 983.31112.7 1071.7 1112.6 1112.71168.2 1112.7 1128.3 1142.11210.5 1138.2 1169.9 1206.91463.3 1370.8 1449.9 1517.61698.8 1698.5 1698.6 1698.61976.5 1940.9 1959.7 1958.02205.8 1991.6 2009.3 1996.52496.3 2436.7 2511.2 2562.03397.9 3193.5 3211.0 3220.53566.6 3292.1 3305.9 3310.14474.8 3865.4 3941.4 3938.14517.4 4504.9 4504.5 4504.64997.7 4611.5 4602.8 4672.55539.0 4945.4 4885.9 4886.56816.6 5233.9 5149.3 5156.6
gear mesh forces for this system plotted against the meshfrequency. The following observations merit attention.
(i) For pure moment loading condition, the dynamicgear mesh forces are quite the same for each planet.It is noteworthy that the peak dynamic mesh force forsun-planet pair is smaller wholly than the one at puretorque loading condition, no matter what the designparameter πΏ1 is.
(ii) For combined loading condition, the dynamic gearmesh forces are different for each planet. There is asimilar trend for planets on a diagonal line.Thedesignparameter πΏ1 influences the dynamic gear mesh
Shock and Vibration 9
(N/mm)
Roll
angl
e (m
m)
0
78
156
91011121314151617181920
2 4 6 8 10 12 14 16 18 200Face width (mm)
(a) Ring-planet 3 at πΏ1 = 0
(N/mm)
Roll
angl
e (m
m)
26.5
63.9
101.2
91011121314151617181920
2 4 6 8 10 12 14 16 18 200Face width (mm)
(b) Sun-planet 3 at πΏ1 = 0(N/mm)
Roll
angl
e (m
m)
0
111
223
91011121314151617181920
2 4 6 8 10 12 14 16 18 200Face width (mm)
(c) Ring-planet 3 at πΏ1 = πΏ0/2
(N/mm)
Roll
angl
e (m
m)
8.3
70.2
132.1
91011121314151617181920
2 4 6 8 10 12 14 16 18 200Face width (mm)
(d) Sun-planet 3 at πΏ1 = πΏ0/2(N/mm)
Roll
angl
e (m
m)
0
134
267
91011121314151617181920
2 4 6 8 10 12 14 16 18 200Face width (mm)
(e) Ring-planet 3 at πΏ1 = πΏ0
(N/mm)
Roll
angl
e (m
m)
6.7
74.1
141.4
91011121314151617181920
2 4 6 8 10 12 14 16 18 200Face width (mm)
(f) Sun-planet 3 at πΏ1 = πΏ0
Figure 6: Load distribution of planet 3 along the face width.
Nat
ural
freq
uenc
y m
odifi
catio
n (%
)
L1 = L0/2
β30
β20
β10
0
10
20
3 5 7 9 11 13 15 17 191Orders
L1 = 0
L1 = L0
(a)
L1 = L0/2
β5
0
5
10
15
Nat
ural
freq
uenc
y er
ror (
%)
3 5 7 9 11 13 15 17 191Orders
L1 = L0
(b)Figure 7: The predicted natural frequencies by the proposed dynamic model. (a) Modification to the results by rigid shaft model and (b)relative error to the results at πΏ1 = 0.
10 Shock and Vibration
Table 5: The first five harmonic amplitudes of the static transmission error excitations.
π s-p1 s-p2 s-p3 s-p4 r-p1 r-p2 r-p3 r-p4
Pure torque loading
1 0.115 0.0542 0.046 0.0303 0.022 0.0044 0.008 0.0095 β β
πΏ1 = 01 0.079 0.15 0.15 0.081 0.694 0.96 0.567 1.032 0.034 0.0595 0.0599 0.035 0.161 0.209 0.141 0.273 0.018 0.0287 0.029 0.018 0.023 0.026 0.015 0.0434 0.0065 0.011 0.011 0.0068 0.035 0.034 0.022 0.0595 0.0005 0.0016 0.0016 0.0005 0.036 0.032 0.019 0.057
πΏ1 = πΏ0/21 0.12 0.154 0.262 0.037 1.04 0.887 1.04 0.9762 0.047 0.06 0.098 0.025 0.306 0.191 0.225 0.2463 0.022 0.029 0.044 0.015 0.049 0.023 0.028 0.0374 0.0077 0.011 0.016 0.0062 0.065 0.031 0.037 0.0535 0.0005 0.0016 0.0023 0.0008 0.059 0.029 0.035 0.054
πΏ1 = πΏ01 0.214 0.104 0.283 0.046 1.21 0.851 1.33 0.9542 0.078 0.044 0.105 0.028 0.426 0.182 0.323 0.2343 0.034 0.022 0.048 0.016 0.066 0.022 0.044 0.0354 0.012 0.009 0.018 0.006 0.074 0.03 0.058 0.0495 0.0006 0.0013 0.0025 0.0007 0.081 0.028 0.045 0.053
1160 20002560 3200
3960
5200
6640
9280
0
1000
2000
3000
4000
5000
Dyn
amic
mes
h fo
rce (
N)
2000 4000 6000 8000 100000Frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(a)
0
1000
2000
3000
4000
5000
Dyn
amic
mes
h fo
rce (
N)
2000 4000 6000 8000 100000Frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(b)
0
1000
2000
3000
4000
5000
Dyn
amic
mes
h fo
rce (
N)
2000 4000 6000 8000 100000Frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(c)
0
1000
2000
3000
4000
5000
Dyn
amic
mes
h fo
rce (
N)
2000 4000 6000 8000 100000Frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(d)Figure 8: Dynamic mesh forces at the ring-planet gear pairs with mixed dynamic model; (a) ring-planet 1, (b) ring-planet 2, (c) ring-planet3, and (d) ring-planet 4.
Shock and Vibration 11
0
500
1000
1500
2000
Dyn
amic
mes
h fo
rce (
N)
2000 4000 6000 8000 100000Frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(a)
0
500
1000
1500
2000
Dyn
amic
mes
h fo
rce (
N)
2000 4000 6000 8000 100000Frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(b)
0
500
1000
1500
2000
Dyn
amic
mes
h fo
rce (
N)
2000 4000 6000 8000 100000Frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(c)
0
500
1000
1500
2000
Dyn
amic
mes
h fo
rce (
N)
2000 4000 6000 8000 100000Frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(d)Figure 9: Dynamic mesh forces at the sun-planet gear pairs; (a) sun-planet 1, (b) sun-planet 2, (c) sun-planet 3, and (d) sun-planet 4.
forces of each ring-planet gear pair differently. Thelargest peak mesh force is for ring-planet 1 because ofits special position.The increasingmagnitude is about2 times.
(iii) Peak dynamic mesh force amplitudes are observedat a number of frequencies. Each of these resonancepeaks can be linked to a certain natural mode andcertain transmission error harmonic amplitude excit-ing that particularmode. For instance, with themixeddynamic model at πΏ1 = πΏ0, the peak at ππ = 2560Hzis due to the natural model at ππ = 2562.0Hz excitedby first harmonic of transmission errors, representingthe condition ππ β ππ/1, so as the other peaks.
Figure 10 shows a set of curves of dynamic bearing forcesfor this system plotted against the mesh frequency. Thedynamic bearing force excited by the transmission error fromsun-planet gear pairs or ring-planet gear pairs is almostidentical and so only outstood in the sun-planet 1 and ring-planet 1, shown in Figures 10(a) and 10(b), by the rightsecondary axis because of different magnitudes. In thecondition of pure moment loading, the reached bearing
dynamic force originated from gearmeshing is tiny. However,the magnitude of bearing forces under combine loadingconditions is expanded significantly regardless of which geartooth meshing the transmission error excitation is generatedfrom. These dynamic forces will propagate from there to thevehicle body through the transmission mountings and thuslead to noise and vibration.
4. Conclusions
In this study, by combining the lumped parameter methodwith finite element method, a three-dimensional, nonlinear,time-variant dynamic model has been developed to simulatethe dynamics of a single-stage helical planetary gear setunder combined force and moment loading conditions. Theperiodic excitation functions have been derived in terms offunctional transmission error and mesh stiffness parametersfrom an elastic contact analysis. The proposed model hasbeen employed to investigate the effects of the combinedloading and positioning on the overall dynamic behavior of afour-planet planetary gear set system quantitatively.
12 Shock and Vibration
4880
3920
2560
0
100
200
300
400
Bear
ing
forc
e (N
)
2000 4000 6000 8000 100000Response frequency (Hz)
L1 = L0
L1 = L0/2
L1 = 0 Pure torque
(a)
3920
48802560
0
100
200
300
400
Bear
ing
forc
e (N
)
2000 4000 6000 8000 100000Response frequency (Hz)
L1 = L0
L1 = L0/2
L1 = 0 Pure torque
(b)
0
100
200
300
400
Bear
ing
forc
e (N
)
2000 4000 6000 8000 100000Response frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(c)
0
100
200
300
400
Bear
ing
forc
e (N
)2000 4000 6000 8000 100000
Response frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(d)
0
100
200
300
400
Bear
ing
forc
e (N
)
2000 4000 6000 8000 100000Response frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(e)
0
100
200
300
400
Bear
ing
forc
e (N
)
2000 4000 6000 8000 100000Response frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(f)
0
100
200
300
400
Bear
ing
forc
e (N
)
2000 4000 6000 8000 100000Response frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(g)
0
100
200
300
400
Bear
ing
forc
e (N
)
2000 4000 6000 8000 100000Response frequency (Hz)
Pure torque
L1 = L0/2
L1 = 0
L1 = L0
(h)
Figure 10: Dynamic bearing force on the output side of the planet carrier shaft excited by the transmission error of (a) sun-planet 1, (b)ring-planet 1, (c) sun-planet 2, (d) ring-planet 2, (e) sun-planet 3, (f) ring-planet 3, (g) sun-planet 4, and (h) ring-planet 4.
Shock and Vibration 13
Based on the results presented, the following conclusionscan be derived:
(i) Under combined force and moment loading condi-tion, bias load distribution on meshing gears easilyoccurs, which is determined by the collinearity of thepositioning of external vertical force with the bearingof the driving component significantly.
(ii) The refined natural frequencies have been revealedafter considering the flexibility of all components inthe proposedmodel, comparingwith lumped dynam-ic models with rigid assumptions, which is helpful toexplain the linkage between the resonance peaks.
(iii) Different from the previous results from pure mo-ment loading, assuming that all planets have goodconsistencies, dynamic mesh forces would not beidentical for all planets under combined force andmoment loading conditions. The maximum peakdynamic mesh force amplitude occurs for planetswhose radial position is collinear to the additionalvertical force direction and most away from theimposed force position. An interesting point is thatpeak dynamic mesh force amplitudes are mostlyreduced formeshing gears in addition to a few planetsmeshing with the driving gear.
(iv) Themagnitude of bearing dynamic forces is expandedsignificantly under combined force and momentloading nomatterwhat the positioning is.Thebearingdynamic forces will propagate from there to the vehi-cle body through the transmission mountings andthus lead to potential vibration and noise problem.
Conflicts of Interest
The authors declare that there are no conflicts of interestregarding the publication of this article.
Acknowledgments
The work presented in this paper has been supported finan-cially by the National Natural Science Foundation of China(Grants 51705012 and 51405010), National Aerospace ScienceFoundation of China (Grant no. 2015ZA51003), and BeijingKey Laboratory for High Efficient Power Transmission andSystem Control of New Energy Resource Vehicle. Finally, theauthorswould like to thankDr. AhmetKahraman (Ohio StateUniversity) for his helpful materials.
References
[1] A. Kahraman, βFree torsional vibration characteristics of com-pound planetary gear sets,β Mechanism and Machine Theory,vol. 36, no. 8, pp. 953β971, 2001.
[2] J. Lin and R. G. Parker, βAnalytical characterization of theunique properties of planetary gear free vibration,β Journal ofVibration and Acoustics, vol. 121, no. 3, pp. 316β321, 1999.
[3] P. Velex and L. Flamand, βDynamic response of planetary trainsto mesh parametric excitations,β Journal of Mechanical Design,vol. 118, no. 1, pp. 7β14, 1996.
[4] M. Inalpolat and A. Kahraman, βA dynamic model to predictmodulation sidebands of a planetary gear set having manufac-turing errors,β Journal of Sound and Vibration, vol. 329, no. 4,pp. 371β393, 2010.
[5] Y. Guo and R. G. Parker, βPurely rotational model and vibrationmodes of compound planetary gears,βMechanism and MachineTheory, vol. 45, no. 3, pp. 365β377, 2010.
[6] Z.-H. Liu, S.-J. Wu, X.-S. Wang, and B. Qian, βNonlineardynamics of compound planetary gear sets based on incremen-tal harmonic balance method,β Journal of Vibration and Shock,vol. 31, no. 3, pp. 117β122, 2012.
[7] M. Kubur, A. Kahraman, D. M. Zini, and K. Kienzle, βDynamicanalysis of a multi-shaft helical gear transmission by finite ele-ments: model and experiment,β Journal of Vibration and Acous-tics, vol. 126, no. 3, pp. 398β406, 2004.
[8] V. Abousleiman, P. Velex, and S. Becquerelle, βModeling of spurand helical gear planetary drives with flexible ring gears andplanet carriers,β Journal of Mechanical Design, vol. 129, no. 1, pp.95β106, 2007.
[9] M. Ebenhoch, J. Greiner, and G. Gaertner, β9-speed Transmis-sion for Front Wheel Drive Applications,β in Proceedings ofInternational VDI Congress-Drivetrain for Vehicles, pp. 57β80,Friedrichshafen, Germany, 2013.
[10] X. Xu Key, βKey Engineering Technology of Shengrui 8AT,β inProceedings of the 2011 International Symposium on Transmis-sion Innovation and China Industrialization, Shanghai, China,2011.
[11] T. F. Conry and A. Seireg, βA mathematical programming tech-nique for the evaluation of load distribution and optimal modi-fications for gear systems,β Journal ofManufacturing Science andEngineering, vol. 95, no. 4, pp. 1115β1122, 1973.
[12] D. R. Houser, βResearch in the gear dynamics and gear noiseresearch laboratory,β in Proceedings of 1982 SAE InternationalOff-Highway and Powerplant Congress and Exposition, vol. 87,pp. 53β58.
[13] H. Yue, Y. Liu, X. Xu, and J. Lai, βStudy on vibration characteris-tics and tooth profile modification of a plus planetary gear set,βJournal of Vibroengineering, vol. 16, no. 2, pp. 954β964, 2014.
[14] J. Lai, Research on the Tooth Contact Characteristic and NoiseReduction Method of Helical Gear [Master, thesis], BeihangUniversity, Beijing, China, 2015.
[15] M. Inalpolat and A. Kahraman, βA theoretical and experimentalinvestigation of modulation sidebands of planetary gear sets,βJournal of Sound and Vibration, vol. 323, no. 3β5, pp. 677β696,2009.
[16] A. Kahraman, βPlanetary gear train dynamics,β Journal ofMechanical Design, vol. 116, no. 3, pp. 713β720, 1994.
RoboticsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporation http://www.hindawi.com
Journal of
Volume 201
Submit your manuscripts athttps://www.hindawi.com
VLSI Design
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 201
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Modelling & Simulation in EngineeringHindawi Publishing Corporation http://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
DistributedSensor Networks
International Journal of