XML Template (2012) [14.12.

2012–10:47am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

Original Article

Proc IMechE Part C:

J Mechanical Engineering Science
Load distribution in spline coupling 0(0) 1–11
! IMechE 2012
teeth with parallel offset misalignment Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0954406212471916
pic.sagepub.com

Francesca Curà1, Andrea Mura1 and Michele Gravina2

Abstract
In spline couplings, torque is theoretically transmitted by all teeth, supporting the same loading level. In practice, due to
manufacturing and mounting errors, not all teeth transmit the same amount of torque and as a result an overloading
condition may occur. In traditional design practice, this uneven load sharing between teeth is often neglected or taken
into account by means of a simplified approach. In this article, a theoretical method, non-finite element method based, is
developed in order to determine both exact number of engaging teeth and shared forces in involute spline couplings with
parallel offset errors. The described process consists of an iterative procedure, in which algorithm is divided into three
main modules: the first one aims to calculate geometrical parameters, the second one analytically determines both
deformation and stiffness of teeth and the third one calculates the actual number of engaging teeth and the shared load.
The algorithm has been benchmarked against finite element method results, showing a very good agreement.

Keywords
Spline coupling, load sharing, transmission design, manufacturing errors, parallel offset

Date received: 16 August 2012; accepted: 30 November 2012

pressure distribution, and fretting wear7 are available

Introduction in literature, but the problem of load sharing between
Spline couplings are mechanical devices used to con- teeth is not investigated in detail.
nect two rotating shafts. They are composed of two In two recent works, Silver et al.8 and Chase et al.9
parts: one is the hub and the other is the shaft. Hub developed a model based on statistics for predicting
and shaft are connected and the torque is transmitted the average number of engaging teeth and the corres-
by means of a number of teeth and grooves. ponding loading level on each engaged pair.
Theoretically all teeth equally contribute to the Medina et al.10 investigated the evolution of slip
torque transfer, but in practice, because of both and pressure between teeth of involute side-fit spline
dimension tolerances and manufacturing errors, not couplings subjected to a steady torque and rotating
all teeth are engaged during the normal functioning.1 misalignment using boundary integral element method.
As a consequence, the transmitted torque may not Elkholy et al.11 studied the effect of angular mis-
be equally shared among the teeth and as a result each alignments related to the generation of frictional and
tooth may be loaded by a different force value. tilting moments transferred through the shaft to the
Traditional design methods consider a number of bearings.
teeth less than the total number. For example, Tjernberg12 considered the pitch error in spline
Giovannozzi1 considers 50% of engaging teeth, couplings and developed a finite element method
Niemann2 considers 75%, French standard3 about (FEM) model of a single tooth and a simplified the-
50% and Dudley4 considers 50% of engaging teeth. oretical model.
These assumptions are generally conservative, bring-
ing to overestimate the engaging forces, and so to 1
Politecnico di Torino Department of Mechanical and Aerospace
overdimension the component. Engineering – C.so Duca degli Abruzzi 24, Torino, Italy
2
Above all in particular applications, as in aerospace Avio s.p.a. Via I Maggio 99, Rivalta di Torino, Italy
field, spline couplings need an accurate design phase. In
Corresponding author:
order to limit the weight, the determination of the exact
Andrea Mura, Politecnico di Torino Department of Mechanical and
number of engaging teeth becomes very important. Aerospace Engineering – C.so Duca degli Abruzzi 24, 10129 Torino,
Many studies about spline couplings concerning Italy.
both theoretical5 and experimental6 analyses of Email: [email protected]

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:47am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

2 Proc IMechE Part C: J Mechanical Engineering Science 0(0)

Adey et al.13 developed a software based on the

boundary element method (BEM) to predict the con-
tact stresses and load transfer in spline couplings.
The above mentioned papers consider the problem
of the non-equal load sharing of spline couplings, but
a general theoretical method to determinate the exact
number of engaging teeth is lacking.
In this article, a numerical iterative method, alter-
native to commercial FEM-based codes, is presented
in order to determine both exact number of engaging
teeth and shared forces in spline couplings.
In particular, the exact load value in each tooth
may also be obtained in real working conditions
when manufacturing errors (in term of tolerances)
may occur.
Only involute teeth are considered in this research.
The algorithm consists of an iterative process,
which starts with only one tooth in contact (one
tooth of the shaft spline in contact with one tooth
of the hub). At each step an additional pair of Figure 1. Spline coupling with parallel offset error (in green
mating teeth is introduced in the model until the the gap between teeth is highlighted).
total external torque is balanced and no further
teeth come in contact due to loaded teeth deflections. As described previously, only one couple of teeth is
The deformation is calculated considering the initially in contact under external loads. Then, thanks
tooth as a cantilever beam,13,14 subjected to bending to teeth deflections, the algorithm allows other couple
and shear strains and considering the deformation of of teeth to get in contact. For the efficient working of
the tooth root as proposed by O’Donnell.15 this procedure for the calculation of the actual
The algorithm is divided into three main modulus: number of engaging teeth, a previous evaluation of
the first is the geometrical, the second refers to the the gap between all teeth is required.
deformation and stiffness of the teeth and the third To determine this starting configuration, a geomet-
determines the engaging conditions. rical model of the splined coupling has been imple-
In particular, in this work the parallel offset error mented in Matlab.
has been considered, but the developed model may be The involute profiles of shaft and hub teeth active
easily extended to consider other kind of errors (as flanks are defined by several points. The points of
spacing errors, profile errors, etc.) by simply modify- shaft teeth active flanks are rigidly rotated around
ing the first module of the algorithm. the geometrical centre of the shaft, following a circu-
This algorithm has been validated by means of lar path, until they become coincident with the cor-
FEM models. responding points on the involute hub active flanks
(Figure 2). The circumferential gap at each radial
location of the flank is calculated as the length of
Calculation process the arc representing the previously described trajec-
The load sharing calculation program consists of a tory. As an example, Figure 2 shows a point of the
modular code. The first module (geometrical analysis) shaft involute profile before (P) and after (P0 ) the
is a routine calculation of the gap between teeth and rotation. The circumferential distance between P
the radial position of the geometrical contact point. and P0 represents the local gap between teeth.
This subprogram also allows a graphical representa- The main output of this first routine is a matrix of
tion of the splined coupling. The second part (teeth circular gaps along each shaft active flank profile. A
deflections analysis) calculates both stiffness and simple observation of this matrix can easily allow the
deformation of teeth, while the third one (solver) is identification of the first engaging tooth and the
the core of the program, which iteratively determines sequence of the following mating couple of teeth.
the load sharing and the actual number of teeth in This result is graphically shown in Figure 1.
contact by using the information obtained in both
first and the second part.
Calculation of teeth deflections
Teeth deformation has been calculated considering
Calculation of the gap between teeth
the tooth as a cantilever beam.14,16 Generally, in
The introduction of a parallel offset error between spline couplings the load is transferred by the contact
shaft and hub causes a non-equal clearance distribu- of the two sides of the engaging teeth so that the load
tion between the teeth, as shown in Figure 1. is distributed, in nominal conditions, along the

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:48am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

Curà et al. 3

Figure 2. Gap between hub and shaft teeth.

Figure 4. Model of the tooth spited into slices.

and the total deformation (B þ S) results as the sum

of the deformation of each slice.
Instead, the deflection due to tooth root deform-
ation R is calculated considering the entire tooth as a
cantilever beam.13
The displacement due to bending deflection results
by the application of the theory of elasticity.18 Each
integral gives the rotation ’ of the generic slice and
the sum of these contributions gives the total rotation
at the end of the beam (equation (2))
Z x1 Z xiþ1
MðxÞ MðxÞ
¼ dx þ Að1Þ þ    þ dx
0 E  Ið1Þ xi E  IðiÞ
Z xn
MðxÞ
þ AðiÞ þ    þ dx þ AðnÞ ð2Þ
xn1 E  IðnÞ

where M(x) is the bending moment acting at the i-th

slice, E is the Young’s modulus of the material, I(i) is
Figure 3. Load applied on the tooth.
the second moment of area of the i-th slice and A(i) is
the i-th integration constant.
contact surface. For the purpose of this study, the The integral of the rotation gives the deformation
load (P) is applied on a single point at the pitch diam- of each slice and the sum of each slice gives the total
eter (Dp), as shown in Figure 3. This assumption may deformation at the point where the force is applied
be justified on the basis of a 2D FEM model17 calcu- (equation (3))
lating the pressure distribution along the involute
Z x1
profile.
The total deformation TOT is obtained as the sum B ¼ ð1Þ  Að1Þ  x  dx þ Bð1Þ þ   
0
of three elastic contributions: bending B, shear S and Z xiþ1
tooth root R deformations:14 þ ðiÞ  AðiÞ  x  dx þ BðiÞ þ   
Zxixn
TOT ¼ B þ S þ R ð1Þ þ ðnÞ  AðnÞ  x  dx þ An ð3Þ
xn1
Considering the bending and shear deflection, the
tooth is divided into n slices, as shown in Figure 4, where B(i) is the i-th integration constant.

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:48am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

4 Proc IMechE Part C: J Mechanical Engineering Science 0(0)

Integration constants are obtained by imposing no The tooth root deformation R is calculated by the
rotation and displacements at the fully constrained O’Donnell formulation15
end of the beam (x ¼ 0, constants A(1) and B(1), equa- "    
tions (4) and (5)). The other constants A(i) and B(i) cos2  16:67 L 2 L
are obtained imposing the necessary consistency rela- R ¼ P    þ2  ð1  Þ 
WE  h h
tions between rotations and deformations at each slice  
2
interface (respectively equations (6) and (7)). tan 
þ 1:534  1 þ ð11Þ
2:4  ð1 þ Þ
PðLÞ2
Að1Þ ¼ ð4Þ where  is the load inclination angle, L is the distance
2  E  Ið1Þ between the tooth root and the load, h is the tooth
width at L, and W is the tooth width.
PðLÞ3 Once the total deformation TOT has been
Bð1Þ ¼ ð5Þ obtained, the tooth stiffness KT is calculated as the
6  E  Ið1Þ
ratio between load and deformation:

PðL  xði  1ÞÞ2 PðL  xði  1ÞÞ2 P

AðiÞ ¼  þ Aði  1Þ KT ¼ ð12Þ
2  E  IðiÞ 2  E  Iði  1Þ TOT
ð6Þ
Calculation of the actual number of engaging teeth
PðL  xði  1ÞÞ 3 and the corresponding load sharing
BðiÞ ¼ þ Aði  1Þ  xði  1Þ þ Bði  1Þ
6  E  IðiÞ The actual number of engaging teeth and the corres-
3
PðL  xði  1ÞÞ ponding load sharing is calculated by means of an
 þ AðiÞ  xði  1Þ ð7Þ iterative process. This recursive logic can be easily
6  E  Iði  1Þ
explained with reference to the flow chart shown in
where P is the applied load and L is the distance Figure 5, where Z is the total number of teeth.
between load application point and tooth root, At the beginning of the iterative process only one
according to Figure 5. pair of teeth is in contact, the corresponding load
The shear deformation can be obtained by the value is applied and the teeth deformation is calcu-
integral calculation of the shear displacement18 lated; if the deformation is higher than the circular
(equation (8)) gap required for a new pair of teeth to get in contact
(gap between teeth and engagement order in unloaded
Z x1
P condition have been previously calculated with the
S ¼  xð1Þ þ Cð1Þ þ    dedicated subroutine described in par. 2.1), the new
0 G  Cð1Þ
Z xiþ1
P
þ xðiÞ þ CðiÞþ   

xi G  CðiÞ
Z xn
P
þ  xðnÞ þ CðnÞ ð8Þ
xn1 G  CðnÞ

where  is the shear deformation factor, G is the shear

elastic modulus and C(i) is the i-th integration
constant.
The first integration constant C(1) is obtained by
imposing no displacements at the fully constrained
end of the beam (x ¼ 0), hence

Cð1Þ ¼ 0 ð9Þ

The other constants are obtained by imposing the

necessary consistency relations between rotations and
deformations at each slice interface, hence

P
CðiÞ ¼  xði  1Þ
G  Cði  1Þ
P
 xði  1Þ þ Cði  1Þ ð10Þ
G  CðiÞ Figure 5. Flow chart of the iterative algorithm.

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:48am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

Curà et al. 5

deformation is again calculated, together with the The applied force Fi on each teeth pair may be
number of engaged teeth. calculated as the teeth deformation di multiplied by
The algorithm proceeds until the total external the teeth stiffness Ki, hence
torque is balanced and no further teeth get in contact
due to loaded teeth deflections. T ¼ K1  d1  r1 þ K2  d2  r2 þ    þ Kn  dn  rn
At the end of the recursive process, the actual ð14Þ
number of engaging teeth and the corresponding
applied load is obtained. Ki stiffness is calculated by the subroutine
Mathematically, the algorithm calculates the equi- described in the previous section considering the
librium of the splined coupling. teeth of hub and shaft as two springs in series.
If a generic step of the process with n teeth pairs The deformation of each tooth at the n-th step may
engaged, the external torque T has to be balanced by be considered as the sum of the deformation necessary
the force applied on each teeth pair Fi (acting at the to close the gap between the n-th tooth pair n and the
radius ri calculated by the geometrical subprogram) additional deformation calculated at the n-th calcula-
according to equation (13) tion step fn (being fn the unknown of the problem), as
follows
T ¼ F1  r1 þ F2  r2 þ    þ Fn  rn ð13Þ
d1 ¼ fn þ ð1  n Þ
d2 ¼ fn þ ð2  n Þ
Table 1. Parameters of the spline coupling.
di ¼ fn þ ði  n Þ
Modulus (mm) 1.27
dn ¼ fn
Number of teeth 26
Outside diameter (mm) 35.2
Replacing the above relationships in equation (14),
Root diameter (mm) 31.3 the following is obtained
Pressure angle ( ) 30
Material Steel T ¼ K1  ðfn þ ð1  n ÞÞ  r1 þ K2  ðfn þ ð2  n ÞÞ  r2
Young’s modulus (MPa) 206,000
þ    þ Kn  fn ð15Þ
Poisson’s ratio 0.3

Figure 6. FEM model of the spline coupling.

FEM: finite element method.

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:49am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

6 Proc IMechE Part C: J Mechanical Engineering Science 0(0)

Finally it is possible to write the expression giving Analytical models refer to three different values of
the deformation at the n-th calculation step fn the parallel offset misalignment, respectively 0.02, 0.05
  and 0.08 mm, obtained by moving the shaft with
T  ½K1  ð1  n Þ  r1 þ K2  ð2  n Þ  r2 respect to the hub; three torque values (200, 1000,
þ þ Ki  ði  n Þ  ri  5000 Nm) have been considered and then 9 test cases
fn ¼ have been run.
ðK1  r1 þ K2  r2 þ  þ Kn  rn Þ
ð16Þ Analytical results have been compared to those
obtained by FEM models. In particular, two 2D
Writing equation (16) in a more compact form, it is FEM models, constituted by plain strain solid
possible to obtain elements (Figure 6), have been considered: these
models refer two different values of the parallel
nP
1 offset (PO), respectively of 0.02 and 0.08 mm.
T ðKi  ri  ði  n ÞÞ Each FEM model has been run with three load
i¼1
fn ¼ ð17Þ values: 200 Nm, 1000 Nm and 5000 Nm; totally 6
P
n
ðKi  ri Þ FEM cases have been considered, resumed in Table 2.
i¼1
Boundary conditions have been imposed on
the external diameter of the hub. Only tangential
constraints have been applied to allow radial
Test cases deformation (like in real working conditions). The
The calculation process developed in the present torque has been applied to the nodes at the internal
work has been applied to a spline coupling, whose diameter of the shaft.
parameters are resumed in Table 1. The contact between teeth has been modelled using
contact elements along to the involute profiles of hub
and shaft, as shown in Figure 7.
Table 2. Test cases.

Parallel
Test case offset (mm) Load (Nm) Results
1 0.02 200 The first result obtained from the calculation process
2 0.02 1000 is the teeth engaging order that depends on the spline
3 0.02 5000 coupling geometry. The engaging order is obtained
4 0.08 200 from the calculation of gap between the teeth: the
5 0.08 1000 first engaging teeth pair is that with gap equal to
zero, and then they are followed by the pair with
6 0.08 5000
the minimum gap and so on.

Figure 7. Contact elements on teeth profiles.

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:49am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

Curà et al. 7

Table 3 shows the engaging order and the respect- torque by the number of teeth, while Ti is the effective
ive gap between the teeth obtained with parallel offset torque contribution calculated by the analytical
level on 0.02, 0.05 and 0.08 mm (teeth have been num- model.
bered counter clockwise starting from the tooth on the Three torque levels (200, 1000 and 5000 Nm) have
top of the shaft (Figure 1)). been considered for each value of the parallel offset
FEM results have been compared to the corres- (Figure 11, 0.02 mm; Figure 12, 0.05 mm; Figure 13,
ponding analytical ones in terms of load sharing 0.08 mm).
among teeth. Finally, Figure 14 shows the effect of the parallel
As an example, Figure 8 shows the force distribu- offset level on the load sharing.
tion for the test case 6 (FEM simulation, Table 2). From the analysis of Figures 9 and 10 it is possible
Figures 9 and 10 show the comparison between to observe that analytical and FEM results match
FEM and analytical results in terms of torque carried very well referring to the load sharing between teeth.
by each tooth (number 1, 2, . . . , 26), respectively with Furthermore, results in terms of applied torque
0.02 mm and 0.08 mm parallel offset values (test cases values also show a very good agreement. In particu-
3 and 6, Table 2). lar, it may be noted that the maximum difference in
Figures 11 to 13 show the effect of torque level on terms of shared torque between FEM and theoretical
the load increase Ti results is about 4.6% (expressed as the percent dif-
ference between TFEM and Tanalytical with respect
Ti ¼ Ti  Tav ð18Þ to TFEM).
It may be also emphasized that all test cases refer
where Tav is the average torque contribution of to a complete engaging of teeth for the considered
each tooth obtained dividing the external applied loading values.

Table 3. Teeth engaging order and relative gap.

PO ¼ 0.02 mm PO ¼ 0.05 mm PO ¼ 0.08 mm

Engaging Gap between Engaging Gap between Engaging Gap between

order teeth (mm) order teeth (mm) order teeth (mm)

22 0.00E þ 00 22 0.00E þ 00 22 0.00E þ 00

23 4.10E05 23 8.90E05 23 1.22E04
21 1.90E03 21 4.70E03 21 7.60E03
24 3.00E03 24 7.30E03 24 1.15E02
20 8.30E03 20 2.09E02 20 3.35E02
25 1.04E02 25 2.57E02 25 4.08E02
19 1.89E02 19 4.75E02 19 7.63E02
26 2.19E02 26 5.43E02 26 8.64E02
18 3.31E02 18 8.30E02 18 1.33E01
1 3.67E02 1 9.14E02 1 1.46E01
17 5.00E02 17 1.25E01 17 2.01E01
2 5.41E02 2 1.35E01 2 2.15E01
16 6.86E02 16 1.72E01 16 2.76E01
3 7.31E02 3 1.82E01 3 2.91E01
15 8.80E02 15 2.20E01 15 3.53E01
4 9.24E02 4 2.31E01 4 3.69E01
14 1.07E01 14 2.68E01 14 4.29E01
5 1.11E01 5 2.78E01 5 4.44E01
13 1.24E01 13 3.11E01 13 4.98E01
6 1.28E01 6 3.20E01 6 5.11E01
12 1.39E01 12 3.47E01 12 5.56E01
7 1.42E01 7 3.55E01 7 5.68E01
11 1.50E01 11 3.76E01 11 6.01E01
8 1.52E01 8 3.81E01 8 6.10E01
10 1.58E01 10 3.94E01 10 6.30E01
9 1.59E01 9 3.97E01 9 6.35E01
PO: parallel offset.

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:49am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

8 Proc IMechE Part C: J Mechanical Engineering Science 0(0)

Figure 8. Force distribution, parallel offset error 0.08 mm and 5000 Nm torque.

Figure 9. Load sharing in test case 3 (PO ¼ 0.02 mm, Figure 10. Load sharing in test case 6 (PO ¼ 0.08 mm,
Torque ¼ 5000 Nm). Torque ¼ 5000 Nm).
PO: parallel offset. PO: parallel offset.

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:50am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

Curà et al. 9

Figure 11. Load sharing theoretical model (PO ¼ 0.02 mm).

PO: parallel offset.

Figure 12. Load sharing theoretical model (PO ¼ 0.05 mm).

PO: parallel offset.

From the analysis of Figures 11 to 13, it is possible both exact number of engaging teeth and shared
to point out that the torque level increasing makes the forces in involute spline couplings with parallel
load distribution less uniform between teeth. offset errors.
The influence of the parallel offset (Figure 14) does The described process consists of an iterative
not appear so meaningful, even if, in the case of the procedure, in which the algorithm is divided into
maximum parallel offset value (0.08 mm), an higher three parts: the first part aims to calculate geometrical
number of teeth may transmit the maximum delta parameters and errors, the second calculates
torque (Figure 13). both deformation and stiffness of teeth and the third
determines the number of engaging teeth and the
shared load.
Conclusions As mentioned above, in this work only a parallel
In this study, a numerical iterative method, non- offset error has been considered, but the developed
FEM-based, has been developed in order to determine model may be easily extended to consider other kind

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:51am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

10 Proc IMechE Part C: J Mechanical Engineering Science 0(0)

Figure 13. Load sharing theoretical model (PO ¼ 0.08 mm).

PO: parallel offset.

Figure 14. Load sharing influence of the parallel offset value (Torque ¼ 5000 Nm).

of errors (as pitch errors, profile errors, and so on) by basis to preview the behavior of spline couplings in
simply modifying the geometrical part of the terms of load sharing.
algorithm. As a matter of fact, the above described method
The implemented procedure has been validated by may be eventually extended to different geometries of
means of FEM models. teeth and to more complex operating conditions.
It may be observed that analytical and FEM results
match very well referring to the load sharing between
teeth and also in terms of applied torque values show Funding
a very good agreement. This work was supported by the Regione Piemonte.
Furthermore, it may be pointed out that the influ-
ence of the parallel offset has been easily emphasized
by the theoretical model. Acknowledgment
Finally, it may be concluded that the developed The authors thank Exemplar s.r.l. for the aid in creation of
procedure represents an useful tool and a powerful FEM models.

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

XML Template (2012) [14.12.2012–10:51am] [1–11]
K:/PIC/PIC 471916.3d (PIC) [PREPRINTER stage]

Curà et al. 11

References convegno nazionale, 7–10 September 2011, Universitá

1. Giovannozzi R. Costruzione di macchine. vol. 1, Degli Studi Di Palermo.
Bologna, Terza Edizione: Pàtron Editore, 1980. 17. Rondano M. Studio della distribuzione di carico sui denti
2. Niemann W. Elementi di macchine. vol. 1, Milano: di accoppiamenti scanalati. Master Thesis, Politecnico di
Edizioni Scienza e Tecnica, 1983. Torino, 2011 (in italian).
3. NF E22-144-3. Cannelures cylindriques droites à flancs 18. Curti G and Curà F. Comportamento meccanico dei
en développante. Module métrique, à centrage sur materiali. C.L.U.T. Editrice, 1999.
flancs, 1986.
4. Dudley DW. How to design involute splines. Product
Appendix
engineering. Ottobre, 1957, pp.75-80.
5. Barrot A, Paredes M and Sartor M. Extended equa- Notation
tions of load distribution in the axial direction in a
spline coupling. Eng Failure Anal 2009; 16: 200–211. A(i) i-th integration constant
6. Cuffaro V, Curà F and Mura A. Analysis of the pres- B(i) i-th integration constant
sure distribution in spline couplings. Proc IMechE, Part C(i) i-th integration constant
C: J Mechanical Engineering Science 2012. DOI: di i-th teeth deformation;
10.1177/0954406212440670. Dp pitch diameter
7. Leen SB, Hyde TH, Ratsimba CHH, et al. An investi- E Young’s modulus
gation of the fatigue and fretting performance of a rep- Fi force applied on each teeth pair
resentative aero-engine spline coupling. J Strain Anal G shear elastic modulus
Eng Des 2002; 37(6): 565–583.
I(i) second moment of area of the i-th slice
8. Silvers J, Sorensen CD and Chase KW. A new statis-
Ki i-th teeth stiffness
tical model for predicting tooth engagement and load
sharing in involute splines. AGMA Technical KT tooth stiffness
Resources, 2010, ISBN 978-1-55589-982-0. L distance between load application point
9. Chase KW, Sorensen CD and DeCaires BJ. Variation and tooth root
analysis of tooth engagement and loads in involute M(x) bending moment
splines. IEEE Trans Autom Sci Eng 2010; 7(4): 746–754. n engaged teeth pairs
10. Medina S and Olver AV. An analysis of misaligned P load
spline couplings. Proc IMechE, Part J: J Engineering T external torque
Tribology 2002; 216: 269–279. Tav average torque contribution of each
11. Elkholy AH and Alfares MA. Misalignment loads in tooth
splined gear coupling. Int J Comput Appl Technol
Ti effective torque contribution
2002; 15(1/2/3): 128–137.
12. Tjernberg A. Load distribution and pitch errors in a
W tooth width
spline coupling. Mater Design 2001; 22: 259–266. Z total number of teeth
13. Adey RA, Baynham J and Taylor JW. Development of
analysis tool for spline couplings. Proc IMechE, Part G: B bending deflection
J Aerospace Engineering 2000; 214: 347–357. R tooth root deformation
14. Cornell RW. Compliance and stress sensitivity of spur S shear deflection
gear teeth. ASME J Mech Des 1981; 103: 447–459. TOT total tooth deformation
15. O’Donnell WJ. The additional deflection of a cantilever ’ tooth slice rotation
due to the elasticity of the support. ASME J Appl Mech  shear deformation factor
1960; 27: 461. Ti load increase
16. Cuffaro V, Curá F and Mura A. Calcolo della defor-
mazione dei denti di alberi scanalati. In: AIAS 40

Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016