Вы находитесь на странице: 1из 14

Measurement 73 (2015) 439–452

Contents lists available at ScienceDirect

Measurement
journal homepage: www.elsevier.com/locate/measurement

Mesh stiffness measurement of cracked spur gear


by photoelasticity technique
Naresh K. Raghuwanshi, Anand Parey ⇑
Mechanical Engineering Department, Indian Institute of Technology Indore, Indore, India

a r t i c l e i n f o a b s t r a c t

Article history: The mesh stiffness plays a vital role in gear dynamics. Mostly, analytical expressions are
Received 29 December 2014 being used to calculate the time varying mesh stiffness. In the present work, photoelastic-
Received in revised form 17 April 2015 ity technique has been used for measuring the stress intensity factor (SIF) for cracked gear
Accepted 29 May 2015
tooth. Subsequently SIF has been used to calculate the gear mesh stiffness. The variations in
Available online 10 June 2015
the SIF and mesh stiffness have been quantified with angular displacements of the gears.
Photoelasticity experiments have been performed for different crack lengths at the tooth
Keywords:
root of the spur gear pair. Experimental results of mesh stiffness variation are compared
Spur gear
Mesh stiffness
with one of the analytical method i.e. potential energy method, which is widely used by
Crack researchers to calculate gear mesh stiffness.
SIF Ó 2015 Elsevier Ltd. All rights reserved.
Photoelasticity

1. Introduction [3,4]. Dolan and Broghamer [5] used photoelasticity for cal-
culating the stresses in the fillets for different types of gear
Gears are used to transmit power from the driving shaft teeth. Jianfeng et al. [6] proposed a new method to mea-
to the driven shaft. The mesh stiffness of gear pair is an sure the tooth dynamic deformation by using dynamic
important parameter which affects the vibration response speckle photography; they achieved tooth dynamic defor-
of the gear box. Gear mesh stiffness varies during engage- mation for full mesh cycle by automatic image processing.
ment to disengagement of gear pairs. The mesh stiffness Ramesh et al. [7] used the photoelasticity for determining
varies due to the geometry of the gear tooth, change in the mixed mode SIFs. Abersek and Flasker [8] used weight
the load position during mesh cycle and number of tooth function method to derive the SIF and shape factor for a
contact pairs. cracked gear tooth. They used narrow gears (two dimen-
The high stress concentration at the tooth root may ini- sional stress states) for the analysis and compared analyt-
tiate crack at the root which propagates during service. ical and experimental results for an edge crack at the root
When crack is present, the SIF plays an important role to of the gear. Shao et al. [9] discussed the variation in SIF
know about crack propagation and can be measured by with change in crack length, gear module, number of teeth
photoelasticity method, which is widely used for experi- and load for crack at the pitch circle and at the tooth root of
mental stress analysis [1,2]. This method is used specially gear. Daniewicz et al. [10] presented the calculation of
for complex geometries like gears. Photoelasticity is used strain energy release rate and mesh stiffness of cracked
to determine the stress distribution in a body at each point. spur gear tooth.
Digital photoelasticity is used for accurate determination Many researchers have investigated the mesh stiffness
of stresses at every point within the interest domain by analytical modelling and numerical simulations. Yang
and Lin [11] used the potential energy method for calculat-
⇑ Corresponding author. ing the effective mesh stiffness. They included three com-
E-mail address: anandp@iiti.ac.in (A. Parey). ponents: Hertzian energy, bending energy, and axial

http://dx.doi.org/10.1016/j.measurement.2015.05.035
0263-2241/Ó 2015 Elsevier Ltd. All rights reserved.
440 N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452

compressive energy. Tian [12] refined the above model by four teeth are meshed simultaneously. In case of single
introducing the shear energy. Wu et al. [13] considered tooth contact pair the total effective mesh stiffness consists
Tian [12] method for evaluating the mesh stiffness and of gear 1 and gear 2 tooth mesh stiffness. The instanta-
studied the effect of crack propagation in the tooth root neous single tooth contact pair mesh stiffness for cracked
for different crack levels and quantified the dynamic pair in the direction of force can be written as [10,13,24],
response of a gear box. Chaari et al. [14] proposed an ana-
1
lytical method to quantify the reduction of gear mesh stiff- kcðspÞ ¼ 1
ð1Þ
þ k1 þ k1c2
ness for two common tooth faults; spalling and breakage. k1 h

Bending, fillet-foundation and contact deflection are taken where kc2 is the tooth stiffness of cracked gear 2, the
into account by Chaari et al. [15] to derive an analytical for- subscript sp denotes the single pair, kh is the Hertzian con-
mulation of time varying gear mesh stiffness. Chen and tact stiffness, k1 is the series combination of bending stiff-
Shao [16] proposed an analytical model to investigate the ness (kb1), axial compressive stiffness (ka1) and shear
effect of gear tooth crack on the gear mesh stiffness for stiffness (ks1) of uncracked gear 1 and can be calculated
tooth crack propagations along tooth width and crack as [12,13,24],
depth. Chen and Shao [17] proposed a model that estab-
lishes the relationship between the gear tooth errors and 1
k1 ¼ 1
ð2Þ
the total mesh stiffness. Pandya and Parey [18] simulated kb1
þ k1a1 þ k1s1
the crack propagation in spur gear tooth for different gear
The instantaneous single pair mesh stiffness for
parameters like pressure angle, fillet radius and backup
uncracked pair in the direction of force can be written as
ratio and its influence on mesh stiffness. Pandya and
[12,13],
Parey [19] computed the total time varying mesh stiffness
for different contact ratios and compared the change in the 1
ktðspÞ ¼ ð3Þ
total effective gear mesh stiffness. Liang et al. [20] used 1
þ k1 þ k12
k1 h
potential energy method to analytically evaluate the mesh
stiffness of a planetary gear set. They used modified can- where k2 is the series combination of bending stiffness
tilever beam model and derived the analytical equations (kb2), axial compressive stiffness (ka2) and shear stiffness
of the bending, shear and axial compressive stiffness. (ks2) of uncracked gear 2 and can be calculated as [12,13],
Zhou et al. [21] developed a modified mathematical model 1
for simulating gear crack with linear growth from root of a k2 ¼ 1
ð4Þ
kb2
þ k1a2 þ k1s2
pinion and improved potential energy method for calculat-
ing mesh stiffness with considering gear body deforma- The total effective mesh stiffness for double pair kt(dp) is
tion. Zheng et al. [22] analysed the influence of friction a parallel combination of instantaneous single pairs mesh
and variations of damping ratio on dynamic transmission stiffness in the direction of force that can be written as
error. Their results show that the anti-backlash gear can [12,13],
increase the composite mesh stiffness as compared with
ktðdpÞ ¼ ktðspÞ þ kcðspÞ ð5Þ
the mesh stiffness of the normal gear pair. Mohammed
et al. [23] presented a new method for calculating the where kt(sp) and kc(sp) are the total effective mesh stiffness
mesh stiffness of cracked tooth root for the purpose of of single pair for healthy gear pair and cracked gear pair
vibration based fault analysis. respectively. The subscript dp denotes the double pair.
From the above literature it has been found that only
analytical and numerical methods are available to calcu- 2.1. Tooth stiffness of cracked gear
late the mesh stiffness. However, Pandya and Parey [24]
proposed the experimental method of mesh stiffness cal- Considering a cracked spur gear tooth of crack length a,
culation by photoelasticity. They presented the variation face width B and applied load per unit face width W. The
in mesh stiffness at three points of single tooth contact detailed derivation of tooth stiffness of cracked pinion is
pair. Their study does not indicate the mesh stiffness vari- given in [10,24] as,
ation from engagement to disengagement of the gear
tooth. In the present study, the work of Pandya and Parey WB
kc2 ¼ ð6Þ
[24] has been extended to full mesh cycle. Moreover, a d
new method to calculate the SIF has been proposed in
where W = F/B and F is the applied load on normal to the
the present study which simplifies the measurement of
tooth profile and d is the deflection of gear tooth in the
SIF without using shape factor for gear tooth profile.
direction of force F. This deflection d is derived on the basis
of strain energy release rate and it includes the contribu-
tion of all the deformations i.e. bending deformation, axial
2. Mesh stiffness calculation of cracked spur gear pair
compressive deformation and shear deformation [24].
The SIF is assumed to be linear function of applied force
In the present study, single tooth contact pairs and dou-
F. The expression of displacement d is given as [10,24],
ble tooth contact pairs of gear system are investigated. In
Z a  
single pair, two teeth are meshed and share equal force. 2B
In case of double pair, two pairs share the total force and d¼ K 2I þ K 2II da ð7Þ
EF 0
N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452 441

where E is the modulus of elasticity of gear material. KI and from the tooth root is x and the central line of the tooth.
KII are the Mode-I SIF and Mode-II SIF respectively. The More detail about hx and a1 calculation are given in [12,13].
Mode-I SIF only has been considered in the present study,
the displacement expression can be written as [24], 2.3. Calculation of SIF by photoelasticity
Z a
2B
d¼ ðK 2I Þda ð8Þ The SIF is a constant parameter which can be used to
EF 0
calculate the stress components near the crack tip. The
SIF depends on the applied stress, type of cracks and geom-
2.2. Stiffness of uncracked gear etry of the specimen. The general expression of SIF is given
as [24–26],
In the present study, the stiffness of uncracked gear tooth pffiffiffiffiffiffi
has been calculated by considering all the components of K I ¼ r paY ð16Þ
total tooth stiffness as bending stiffness (kb), axial compres- where KI, r, a and Y are the Mode-I SIF, applied stress, crack
sive stiffness (ka), shear stiffness (ks) and Hertzian contact length and shape factor. The shape factor Y depends on
stiffness (kh). Stiffness due to fillet foundation is not consid- geometry of the specimen and crack length. The shape fac-
ered for simplicity. These stiffness components are calcu- tor plays an important role to evaluate the SIF especially
lated based on potential energy method [12,13,24], for gear tooth profile because shape factor is changed when
The Hertzian contact stiffness kh remains constant dur- force moves on the tooth profile and also by load sharing
ing the complete mesh cycle [12,13,24] and can be written ratio as in double tooth contact pair. Pandya and Parey
as, [24] used Eq. (16) to evaluate KI, they calculated r by pho-
EBp toelasticity experiment and shape factor was evaluated
kh ¼ ð9Þ analytically [8,24]. Analytical expression of shape factor
4ð1  t2 Þ
is laborious and time consuming.
where E, B and t are the modulus of elasticity, tooth face In the present work, an alternative method of SIF calcu-
width and Poisson’s ratio respectively. lation is used to extract the SIF from the modified
The bending, axial compressive and shear stiffness can Westergaard’s equations by using photoelasticity experi-
be calculated by these expressions as given in [12,13,24], ment. The advantage of this method is that there is no need
F2 F2 F2 to calculate the shape factor Y for a complicated geometry
kb ¼ ; ka ¼ ; ks ¼ ð10Þ like gear tooth profile. These modified Westergaard’s equa-
2U b 2U a 2U s
tions can be employed to edge crack, centre crack and
where F is the contacting normal force on the tooth profile finite body for Mode-I loading. These equations near the
and Ub, Ua and Us are the bending, axial compressive and crack tip can be written as [7],
shear potential energies respectively. These potential ener- 8 9 8 9 8 9
gies can be expressed as given in [12,13,24], < rx >
> = KI h
>
< 1  sin 2h sin 3h
2 >
= >
< rox >
=
Z ry ¼ pffiffiffiffiffiffiffiffi
ffi cos 1 þ sin 2h sin 3h
2
þ 0
d 2 >
: > 2pr 2> >
; > >
Ub ¼
½F b ðd  xÞ  F a  h
dx ð11Þ sxy ; :
sin 2h cos 3h
2
:
0
;
0 2EIx
ð17Þ
Z d
F 2a where rx, ry and sxy are the stress in x direction, stress in y
Ua ¼ dx ð12Þ
0 2EAx direction and shear stress respectively at a point r and h
near the crack tip as shown in Fig. 1. KI is the Mode-I SIF
Z d
1:2F 2b and rox is a constant stress term which is added to the rx
Us ¼ dx ð13Þ component in the classical Westergaard’s equations. Eq.
0 2GAx
where x is the distance of the section from the tooth root, Ix
is the area moment of inertia of the section at a distance x,
d is the distance between the contact point and the tooth
root, h is the distance between the contact point and the
central line of the tooth, Ax is the section area at a distance
x and G is the shear modulus.
The Fb and Fa can be calculated by resolving the force F
as [12,13,24],
F b ¼ F cos a1 and F a ¼ F sin a1 ð14Þ
where a1 is the angular variable [12].
The Ix and Ax can be calculated as [12,13],
1 3
Ix ¼ ð2hx Þ B; Ax ¼ 2hx B ð15Þ
2
where hx is the distance between the point on the tooth
curve corresponding to the section where the distance Fig. 1. Stress components near the crack tip.
442 N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452

(17) can be used for calculating KI using photoelasticity [7]. to the cross section of the beam at the root of the gear
For high accuracy the data collection should be collected tooth where the crack is generated. So the biaxiality ratio
near to the crack tip and crack should not be reached up is zero in the study and rox becomes equal to applied stress
to boundary [7]. This method can also be applied for long r. The r can be evaluated by photoelasticity method. The
cracks such as a/w = 0.9 where w is the width of plate, stress optic law can be written as [1]
but the zone of data collection should be very small as
Nfr
r/a < 0.03 and the angle of data collection from the crack r1  r2 ¼ ð19Þ
h
axis should be in the range of 73° < h < 139° [7]. In the pre-
sent study this method is applied on short cracks as where r1 and r2 are the maximum and minimum principle
a/w < 0.217 where w is the width of the tooth at root in stresses respectively. N is the isochromatic fringe orders, fr
the present study and the zone of data collection was car- is the material fringe constant and h is the model thick-
ried out within r/a < 0.412 at h = 90°, so the Eq. (17) can be ness. Now r2 is zero in present case due to unidirectional
applied to the problem considered in this study with suffi- applied stress. So the constant stress term rox becomes
cient accuracy. equal to r1 or r. Therefore the rox can be calculated by
The rox can be expressed in terms of biaxiality ratio k as photoelasticity method as,
[7] Nfr
rox ¼ ð20Þ
rox ¼ ð1  kÞr ð18Þ h
In Eq. (20) the isochromatic fringe order N is counted at
where k is the ratio of two applied principal stresses in x
the tooth root when crack is absent.
and y directions and r is the applied principal stress in y
The principal stresses at a point r and h from the crack
direction. If the ratio becomes unity, the modified
tip as shown Fig. 1 can be written as,
Westergaard’s equations become classical Westergaard’s
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
equations [7]. rx þ ry r  r 2
x y
In the present study, the gear tooth is assumed as a can- r1 ¼ þ þ s2xy ð21Þ
2 2
tilever beam and the applied stress acting in unidirectional

Fig. 2. Flow chart of mesh stiffness calculation methodology for cracked gear by photoelasticity technique.
N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452 443

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rx þ ry r  r 2 The whole methodology used for mesh stiffness calcula-
x y
r2 ¼  þ s2xy ð22Þ tion in the present study can be explained by a flow chart
2 2
as shown in Fig. 2.
where r1 and r2 are the maximum and minimum principal
stresses respectively.
The difference between two principal stresses from (21) 3. Experimental studies
and (22) can be written as,
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3.1. Photoelastic model
r  r 2
x y
r1  r2 ¼2 þ s2xy ð23Þ
2 Polycarbonate birefringent material was used for mak-
ing the specimen of spur gears of standard involute profile.
From (19) and (23) the expression can be written as,
Advanced CNC milling machine was used for obtaining
 2 
Nfr rx  ry  2 2 high accuracy. In this work the same material has been
¼ þ sxy ð24Þ
2h 2 used as given in [24]. Material fringe constant (fr) of this
material is 4.95 N/mm/fringe. The main parameters of
With the help of Eqs. (17) and (24), KI can be computed. gears are given in Table 1. Fig. 3 shows the specimens used
for experimental work.
Table 1
Main parameters of gears.
3.2. Description of photoelasticity setup and loading
Parameters Gear 1 and gear 2 arrangement
No of teeth 13
Diametral pitch 1.587/in. The photoelastic setup is shown in Fig. 4. A plane polar-
Module 16 mm
iscope is employed for this experiment. A plane polar-
Pressure angle 20°
Face width 6 mm iscope is constructed by four elements; light source,
Contact ratio 1.44 polarizer, model and analyzer. White light source is used
Young’s modulus 200 GPa as a light source and a digital camera is used for capturing
Poisson’s ratio 0.3 the images of isochromatic fringes. For loading and mount-
ing of gears, special type of fixtures is designed. For apply-
ing the torque on gear 1 in anticlockwise direction, a
300 mm long lever is attached to the gear 1. The load can
be hanged at the end of this lever. For measuring the angu-
lar displacement of gear 2, a big protector is attached on
the fixtures having least count of 0.5°. The gear 1 is free
to rotate on its axle and gear 2 is also free to rotate on
its axle but can be fixed at any angle of rotation with the
help of nuts and bolts. For this, a D-shape slot is provided
to fix the gear 2 at different angular positions.

3.3. Force measurement

Let P is the applied load at the end of the lever having


length L. The applied torque T on the gear 1 in anticlock-
Fig. 3. Gear specimens used for testing. wise direction can be calculated as,

Fig. 4. Photoelastic setup (a) loading and fixing arrangements of gears on photoelastic setup. (b) Photoelastic unit ready for observations.
444 N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452

T ¼PL ð25Þ
Torque T on the gear 1 can also be expressed as,

T ¼ F t  R1 ð26Þ

where Ft is the tangential force at the pitch point of the


pitch circle and R1 is the pitch circle radius of gear 1.
From (25) and (26) Ft can be calculated.
The normal force F on the tooth having involute profile
can be calculated as,

Ft
F¼ ð27Þ
cos u

where Ft is tangential force and u is pressure angle as


Fig. 5. Resolution of force on gear tooth. shown in Fig. 5.
In the present study, the values of known parameters
are as following: P = 66 N, L = 300 mm, R1 = 104 mm and
pressure angle u = 20°. Finally, the normal force F on the
tooth profile is calculated as 202.6 N. In single tooth con-
tact the load sharing is taken as the normal force F but in
case of double tooth contact the load sharing factor is
chanced from engagement to highest point of single tooth
contact (HPSTC) 0.2F to 0.7F and from lowest point of sin-
gle tooth contact (LPSTC) to disengagement as 0.7F to 0.2F
for gear 2 as shown in Fig. 6 [27].

4. Isochromatic fringe orders

Isochromatic fringe patterns are shown in Figs. 7–13.


These images have been captured by a digital camera.
After that images have been processed in greyscale for
counting the fringe orders. For calculating the applied
Fig. 6. Load-sharing factor from engagement to disengagement for gear 2. stress at the tooth root (without crack) isochromatic fringe

Fig. 7. Isochromatic fringe patterns of healthy gears for double tooth contact pairs in images (a–f) at angular displacement b = 0°, 2.5°, 5°, 7.5°, 10°, 12°
respectively.
N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452 445

patterns are captured at different angular displacement as engagement to highest single tooth contact point (HSTCP),
shown in Figs. 7–9. The starting point of engagement is (b = 0–12°). Similarly, Fig. 8 shows the fringes during single
considered at the end of the tooth tip of gear 2. Fig. 7 shows tooth contact pair, (b = 13–25°) and Fig. 9 shows the
isochromatic fringes during double tooth contact pair from fringes during double tooth contact pair from lowest single

Fig. 8. Isochromatic fringe patterns of healthy gears for single tooth contact pair in images (a–f) at angular displacement b = 13°, 15°, 17.5°, 20°, 22.5°, 25°
respectively.

Fig. 9. Isochromatic fringe patterns healthy gears for double tooth contact pair in images (a–f) at angular displacement b = 26°, 28.5°, 31°, 33.5°, 36°, 39°
respectively for double tooth pair contact.
446 N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452

Fig. 10. Isochromatic fringe patterns near the crack tip of cracked gear tooth (a = 2 mm) in images (a–r) at angular displacement b = 0°, 2.5°, 5°, 7.5°, 10°,
12°, 13°, 15°, 17.5°, 20°, 22.5°, 25°, 26°, 28.5°, 31°, 33.5°, 36°, 39° respectively.

tooth contact point (LSTCP) to disengagement (b = 26–39°). Now, cracks are introduced at the tooth root of the gear
Observed fringe orders N at the tooth root for healthy gear 2 and the isochromatic fringes near the crack tip were cap-
tooth are tabulated with angular displacement of gears in tured from engagement to disengagement. Observed fringe
Table 2. patterns for 2 mm crack are shown in Fig. 10 at different
N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452 447

Fig. 11. Isochromatic fringe patterns near the crack tip of cracked gear tooth (a = 3 mm) in images (a–r) at angular displacement b = 0°, 2.5°, 5°, 7.5°, 10°,
12°, 13°, 15°, 17.5°, 20°, 22.5°, 25°, 26°, 28.5°, 31°, 33.5°, 36°, 39° respectively.

angular displacement. Obtained fringe patterns are used should be negative. Similarly, Figs. 11–13 show the fringe
for counting the fringe orders N near the crack tip at r patterns for 3 mm, 4 mm and 5 mm crack lengths during
and h coordinate. For simplicity, all the fringe orders are full mesh cycle, eighteen observations were taken for each
counted at h = 90°. After digital image processing these crack during engagement to disengagement. With the help
data are tabulated in Table 3. In isochromatic patterns of these fringe patterns SIFs are computed from engage-
the loops are found backward tilted that shows the rox ment to disengagement.
448 N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452

Fig. 12. Isochromatic fringe patterns near the crack tip of cracked gear tooth (a = 4 mm)in images (a–r) at angular displacement b = 0°, 2.5°, 5°, 7.5°, 10°, 12°,
13°, 15°, 17.5°, 20°, 22.5°, 25°, 26°, 28.5°, 31°, 33.5°, 36°, 39° respectively.

5. Results and discussion parameters, deflection of cracked tooth in the direction of


force was calculated from (8). The deflections of cracked
The photoelasticity experiment was conducted to calcu- gear tooth for full mesh cycle are shown in Table 4. The
late applied stress and Mode-I SIF. With the help of these stiffness of cracked gear tooth was calculated from (6) by
N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452 449

Fig. 13. Isochromatic fringe patterns near the crack tip of cracked gear tooth (a = 5 mm) in images (a–r) at angular displacement b = 0°, 2.5°, 5°, 7.5°, 10°,
12°, 13°, 15°, 17.5°, 20°, 22.5°, 25°, 26°, 28.5°, 31°, 33.5°, 36°, 39° respectively.

Table 2
Isochromatic fringe orders (N) at tooth root for healthy gear tooth.

b 0° 2.5° 5° 7.5° 10° 12° 13° 15° 17.5° 20° 22.5° 25° 26° 28.5° 31° 33.5° 36° 39°
N 6 5 5 4 4 3 7 7 7 6 6 6 3 3 4 4 5 5
450 N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452

Table 3
Photoelasticity experimental data near the crack tip.

b (°) Crack
2 mm 3 mm 4 mm 5 mm
N r (mm) h (°) N r (mm) h (°) N r (mm) h (°) N r (mm) h (°)
Double tooth contact 0 6 0.90 90 7 0.85 90 7 0.90 90 7 1.03 90
2.5 6 0.83 90 7 0.80 90 7 0.87 90 7 0.98 90
5 5 0.90 90 6 0.82 90 6 0.93 90 6 1.10 90
7.5 5 0.89 90 5 0.92 90 5 0.98 90 6 0.88 90
10 4 0.98 90 5 0.86 90 5 0.94 90 6 0.81 90
12 4 0.97 90 5 0.81 90 5 0.90 90 5 1.03 90
Single tooth contact 13 9 0.84 90 9 0.94 90 10 0.91 90 10 1.01 90
15 8 0.98 90 9 0.86 90 9 1.10 90 10 0.92 90
17.5 8 0.94 90 9 0.75 90 9 1.03 90 10 0.88 90
20 8 0.84 90 8 0.92 90 9 0.98 90 10 0.99 90
22.5 8 0.91 90 8 0.95 90 8 1.15 90 10 0.77 90
25 8 0.95 90 8 0.99 90 8 1.10 90 10 0.79 90
Double tooth contact 26 4 0.75 90 5 0.86 90 5 0.90 90 5 0.96 90
28.5 4 0.89 90 5 0.89 90 5 0.92 90 5 0.96 90
31 5 0.60 90 5 0.93 90 5 0.96 90 6 0.72 90
33.5 5 0.65 90 5 1.02 90 5 1.02 90 6 0.82 90
36 5 0.84 90 6 0.77 90 6 0.79 90 6 0.91 90
39 5 0.91 90 6 0.87 90 6 0.92 90 6 1.03 90

Table 4 putting the value of deflection. Stiffness of uncracked gear


Deflections of cracked gear tooth at different angular displacements (b). tooth was calculated by potential energy method as
b (in degrees) Deflections for different crack lengths (mm) described in Section 2.2. After that, instantaneous mesh
a = 2 mm a = 3 mm a = 4 mm a = 5 mm
stiffness was evaluated for single tooth contact pair and
double tooth contact pair from (1) and (5) respectively.
0 0.00082 0.00155 0.00219 0.00313
For crack length a = 1 mm, it was difficult to count the
2.5 0.00049 0.00092 0.00134 0.00189
5 0.00028 0.00054 0.00082 0.00122 fringe orders near the crack tip for keeping the r/a ratio
7.5 0.00021 0.00033 0.00048 0.00073 small. Therefore, experiment was performed on crack
10 0.00013 0.00026 0.00038 0.00056 lengths starting from 2 mm and up to 5 mm by 1 mm
12 0.00011 0.00020 0.00028 0.00041
increment. Experimentally calculated Mode-I SIF for differ-
13 0.00033 0.00055 0.00085 0.00118
15 0.00031 0.00051 0.00086 0.00108
ent crack lengths at 18 points during mesh cycle are shown
17.5 0.00030 0.00044 0.00081 0.00103 in Fig. 14. The values of SIF in Fig. 14 are given in
pffiffiffiffiffiffiffiffiffi
20 0.00026 0.00042 0.00073 0.00092 MPa mm. It has been observed that the values of SIFs
22.5 0.00028 0.00044 0.00069 0.00086
for double tooth contact pair are less as compared with sin-
25 0.00029 0.00046 0.00064 0.00088
26 0.00008 0.00020 0.00028 0.00038
gle tooth contact pair. The highest SIF for 2 mm crack has
pffiffiffiffiffiffiffiffiffi
28.5 0.00011 0.00024 0.00035 0.00045 been observed around 24 MPa mm for single tooth con-
31 0.00014 0.00034 0.00046 0.00060
pffiffiffiffiffiffiffiffiffi
tact pair and 16:5 MPa mm for double tooth contact pairs
33.5 0.00019 0.00044 0.00062 0.00085
as shown in Fig. 14. When the crack length is increased the
36 0.00035 0.00068 0.00093 0.00135
39 0.00057 0.00116 0.00163 0.00230 SIF also increased. Higher SIF during b = 13–25°, indicates
that the stresses are higher for single tooth contact pair.

Fig. 14. Variation in Mode-I SIF for full mesh cycle by photoelasticity method.
N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452 451

Spur gear specimens without tooth profile modification pairs has been observed higher than single tooth contact
have been tested with accuracy and the tooth tip contact pair in case of photoelasticity method. When crack length
effect has been observed at the time of engagement and is increased the mesh stiffness reduction has been
disengagement. The tooth tip contact increases the stresses observed higher by photoelasticity method. Trend of mesh
at the engagement and disengagement positions; hence stiffness variation during single tooth contact pair and
the SIF is also higher at these points. double tooth contact pairs has been found a good match
The deflections are given in millimetres (mm) in Table 4 for higher crack lengths. For 2 mm crack the mesh stiffness
for cracked tooth which is in contact during meshing for during single tooth contact pair is decreasing and for
different crack lengths i.e. 2 mm, 3 mm, 4 mm and 5 mm. longer cracks (4 mm and 5 mm) it is increasing from
From Table 4, it is clear that the deflections are increasing HPSTC to LPSTC in both the methods as shown in Fig. 15.
with the increase in crack lengths. The variation in deflec- It has been observed that while comparing the experi-
tion of cracked gear tooth with angular displacement of mental results with Wu et al. [13] results, the mean mesh
gears have been observed as deflection of cracked gear stiffness of experiments is lower than analytical values. As
tooth is higher during single tooth contact pair than double the crack length increases percentage difference between
tooth contact pairs. It can be seen that the deflections is experimental values and analytical values is increasing.
higher during single tooth contact pair due to higher load Table 5 shows the percentage difference in mean mesh
because of load sharing ratio equal to one. At the time of stiffness between experimental and analytical values. The
engagement and disengagement the deflections are higher mean mesh stiffness for 2 mm crack has been observed
due to tooth tip contact which increases the contact stres- by photoelasticity experiment as 4.991  105 N/mm
ses as well as stress concentrations at the crack tip. These whereas by Wu et al. [13] the mean mesh stiffness is
deflections were used to calculate the stiffness of the obtained as 5.886  105 N/mm. The percentage difference
cracked tooth. in mean mesh stiffness between two methods is
Mesh stiffness has been calculated by photoelasticity 15.21%. Similarly for 3 mm, 4 mm and 5 mm cracks the
method and potential energy method of Wu et al. [13]. percentage differences are tabulated in Table 5.
The comparative study of mesh stiffness by two methods
is shown in Fig. 15. This study shows the trend of mesh Table 5
stiffness variation by two methods and found almost same Comparison of mean mesh stiffness in N/mm (105).
but lower values of mesh stiffness has been observed by
Methods 2 mm 3 mm 4 mm 5 mm
photoelasticity experiment than potential energy method crack crack crack crack
(Wu et al. [13]). It has been observed that the difference
Wu et al. [13] 5.886 5.850 5.810 5.767
in mesh stiffness values between two methods is small
Photoelasticity 4.991 4.233 3.761 3.463
for 2 mm crack and it is higher when crack is increased. Percentage difference 15.21% 27.63% 35.26% 39.94%
The variation in mesh stiffness during double tooth contact

Fig. 15. Mesh stiffness comparison by photoelasticity experiment and Wu et al. [13] for different crack lengths.
452 N.K. Raghuwanshi, A. Parey / Measurement 73 (2015) 439–452

6. Conclusion [9] R. Shao, F. Dong, W. Wang, P. Jia, Influence of cracks on dynamic


characteristics and SIF of gears, Eng. Fail. Anal. 32 (2013) 63–80.
[10] S.R. Daniewicz, J.A. Collins, D.R. Houser, The stress intensity factor
Photoelasticity technique has been used to measure the and the stiffness for a cracked spur gear tooth, J. Mech. Des. 116
SIF for different crack lengths. Modified Westergaard’s (1994) 697–700.
[11] D.C.H. Yang, J.Y. Lin, Hertzian damping, tooth friction and bending
equations were used to calculate the SIF experimentally.
elasticity in gear impact dynamics, J. Mech., Transm., Automat. Des.
The mesh stiffness values were computed for single tooth 109 (2) (1987) 189–196.
contact pair as well as double tooth contact pairs for full [12] X.H. Tian, Dynamic simulation for system response of gearbox
including localized gear faults, Master’s Thesis, University of Alberta,
mesh cycle. The results obtained by photoelasticity exper-
Edmonton, Alberta, Canada, 2004.
iments have been compared with analytical method. The [13] S. Wu, M.J. Zuo, A. Parey, Simulation of spur gear dynamics and
experimental values of mesh stiffness have shown good estimation of fault growth, J. Sound Vib. 317 (3–5) (2008) 608–624.
trend of mesh stiffness variation with established Wu’s [14] F. Chaari, W. Baccar, M.S. Abbes, M. Haddar, Effect of spalling or
tooth breakage on gearmesh stiffness and dynamic response of a
analytical method from engagement to disengagement one-stage spur gear transmission, Eur. J. Mech.-A/Solids 27 (4)
but the difference in the magnitude of mesh stiffness (2008) 691–705.
between two methods has been found large. Wu’s model [15] F. Chaari, T. Fakhfakh, M. Haddar, Analytical modelling of spur gear
tooth crack and influence on gear mesh stiffness, Eur. J. Mech.-A/
does not include the effect of rim deformation, involute Solids 28 (2009) 461–468.
profile error, etc. In our photoelasticity experiment the [16] Z. Chen, Y. Shao, Dynamic simulation of spur gear with tooth root
photoelasticity setup frame is considered as a rigid body. crack propagating along tooth width and crack depth, Eng. Fail. Anal.
18 (8) (2011) 2149–2164.
However, the efficiency of photoelasticity experiment to [17] Z. Chen, Y. Shao, Mesh stiffness calculation of a spur gear pair with
calculate the mesh stiffness is quite high. This method tooth profile modification and tooth root crack, Mech. Mach. Theory
includes rim deformation, bending deformation, axial 62 (2013) 63–74.
[18] Y. Pandya, A. Parey, Simulation of spur gear tooth for different gear
deformation, shear deformation, etc. Further, the efficiency parameters and its influence on mesh stiffness, Eng. Fail. Anal. 30
can also be improved with the modification in the photoe- (2013) 124–137.
lasticity setup by taking rigid structure and by using digital [19] Y. Pandya, A. Parey, Failure path based modified gear mesh stiffness
for spur gear pair with tooth root crack, Eng. Fail. Anal. 27 (2013)
photoelasticity. The photoelasticity technique can be used
286–296.
as an alternate to analytical method for calculation of mesh [20] X. Liang, M.J. Zuo, M. Pandey, Analytically evaluating the influence of
stiffness of cracked gears. In future, this work will be crack on the mesh stiffness of a planetary gear set, Mech. Mach.
extended for planetary gears. Theory 76 (2014) 20–38.
[21] X. Zhou, Y. Shao, Y. Lei, M. Zuo, Time-varying meshing stiffness
calculation and vibration analysis for a 16 degree of freedom
References dynamic model with linear crack growth in a pinion, J. Vib. Acoust.
134 (2012). 011011-1-11.
[1] J.W. Dally, W.F. Riley, Experimental Stress Analysis, McGraw-Hill, [22] Z. Yang, J.Z. Shang, Z.R. Luo, Effect analysis of friction and damping
New York, 1978. on anti-backlash gear based on dynamics model with time-varying
[2] J.F. Doyle, J.W. Phillips, Manual on Experimental Stress Analysis, fifth mesh stiffness, J. Cent. South Univ. 20 (2013) 3461–3470.
ed., Society for Experimental Mechanics, 1989. [23] O.D. Mohammed, M. Rantatalo, J.O. Aidanpaa, Improving mesh
[3] E.A. Patterson, Digital photoelasticity: principles, practice and stiffness calculation of cracked gears for purpose of vibration-based
potential, Strain 38 (2002) 27–39. fault analysis, Eng. Fail. Anal. 34 (2013) 235–251.
[4] I.A. Wang, P. Wang, Complete fringe order determination in digital [24] Y. Pandya, A. Parey, Experimental investigation of spur gear tooth
photoelasticity using fringe combination matching, Strain 39 (2003) mesh stiffness in the presence of crack using photoelasticity
121–130. technique, Eng. Fail. Anal. 34 (2013) 488–500.
[5] T.J. Dolan, E.L. Broghamer, A photoelasticity study of stresses in gear [25] K. Prashant, Elements of Fracture Mechanics, Tata McGraw Hill, New
tooth fillets, University of Illinois, Bulletin No. 335, vol. 31, 1942. Delhi, 2009.
[6] L. Jianfeng, A. Xing, T. Zhiren, Measurement of gear tooth dynamic [26] T.L. Anderson, Fracture Mechanics, third ed., Taylor & Francis, Boca
deformation using dynamic speckle photography, Chin. Sci. Bull. 43 Raton, 2005.
(6) (1998) 514–518. [27] G.K. Sfantos, V.A. Spitas, T.N. Costopoulos, G.A. Papadopoulos, Load
[7] K. Ramesh, S. Gupta, A.A. Kelkar, Evaluation of stress field parameter sharing of spur gears in mesh an analytical and experimental study,
in fracture mechanics by photoelasticity – revisited, Eng. Fract. Technical Report, National Tech. Univ. Athens, No. TR-SM-0303,
Mech. 56 (1) (1997) 25–45. March 2003.
[8] B. Abersek, J. Flasker, Stress intensity factor for cracked gear tooth,
Theoret. Appl. Fract. Mech. 20 (1994) 99–104.