Engineering Failure Analysis 121 (2021) 105110

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Fatigue failure modeling and life expectancy of the dipper-teeth

assembly of a mining shovel
Muhammad Azeem Raza a, *, Samuel Frimpong b
a
University of Engineering and Technology, Lahore, Pakistan
b
Missouri University of Science and Technology, Rolla, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Large mining shovels are used to achieve economic bulk production in surface mining operations.
Shovel dynamics The suspended payload for these large capacity mining shovels combined with dipper weight, and
Virtual prototype harsh digging environment may result in severe stress loading of the shovel front-end assembly.
Crack propagation
Material flaws, high stresses and harsh operating environment can initiate cracks in the dipper-
Stress intensity factor
Fatigue life estimation
teeth assembly that can propagate to critical lengths resulting in fatigue failure, unscheduled
downtimes, costly unplanned repairs, and downstream processing circuit problems. The existing
shovel research has been restricted to shovel dynamic modeling and stress measurement for the
boom only and the fatigue life estimation has been ignored so far. In this research, we provided a
framework for modeling the fatigue failure based on crack propagation studies for the dipper-
teeth assembly. We proposed a step-by-step scheme that includes the modeling of the forma­
tion resistive forces, kinematic and dynamic modeling of the shovel front-end assembly, virtual
prototyping, and crack propagation and life expectancy studies. We implemented this framework
for a large mining shovel. A virtual P&H 4100XPC shovel prototype was built in ANSYS (R15)
software for stress and fatigue failure modeling studies. We introduced crack at various locations
of the dipper. These cracks were incremented gradually. Crack propagation simulation studies
show that a 100 mm crack-length is a critical crack-length for the shovel dipper. A 75 mm bottom-
plate crack can propagate to the critical length in 16 days. These initial but pioneering insights
can provide a scientific basis for shovel maintenance and care and contribute towards shovel
health and longevity.

1. Introduction

Excavation is one of the basic activities in surface mining operations and is primarily performed through large capacity shovels that
offer high production capacities, lower unit production costs, and better ruggedness. The capital investment in cable shovels can be as
high as $30 million. The excavating capacities of the shovels have seen an increasing trend over the years to achieve economy of scale,
and the modern day cable shovels have capacities of 120 tons per scoop [1,2]. Along with the enhanced productivity advantages, these
large capacity shovels present challenges for stress loading and fatigue failure. The presence of large payload, combined with the
dipper self-weight, results in extreme stress loading of the whole shovel front-end assembly, especially the dipper-teeth assembly. The
dipper-teeth assembly is also subject to the formation resistive forces, majority of which are dynamic in nature. Shovel loading is a

* Corresponding author.
E-mail addresses: [email protected] (M. Azeem Raza), [email protected] (S. Frimpong).

https://doi.org/10.1016/j.engfailanal.2020.105110
Received 22 July 2020; Received in revised form 9 November 2020; Accepted 17 November 2020
Available online 23 November 2020
1350-6307/© 2020 Elsevier Ltd. All rights reserved.
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

Nomenclature

β constant inclination of dipper from crowd-arm

da/dN crack growth rate per cycle as defined by Paris Law
a crack length
Nf number of cycles to failure
K stress intensity factor
m material constant for Paris Law
c material constant for Paris Law
C(ϴ,Ḯ˙ ) generalized Coriolis and centripetal torque
D(ϴ) generalized inertia matrix
G(ϴ) generalized gravity torque
m1, m2 mass of crowd-arm and dipper, respectively
β constant inclination of link 3 from link 2 (inclination of X4 from X3)
θe inclination of coordinate frame 4′ from coordinate frame 3.
L1 length of crowd-arm from pivotal point to connection point between arm and dipper
L2 length between dipper tip and connect point of arm and dipper
si, ci Sinθi and cosθi, respectively
di offset distance of the gravity center in link i
Fn, Ft normal & tangential cutting resistive forces on dipper tip
F’ cable shovel breakout force
Fload(Fn, Ft) formation resistive forces
ODEs ordinary differential equations

cyclic process and those repeated loading and unloading cycles induce fatigue stresses in shovel components. Cracks appear over time
in the dipper-teeth assembly that may propagate to critical lengths under conditions of high stress loading resulting in fatigue failure of
front-end components. Shovel breakdown time is a significant and frequent contributor to total shovel breakdown time [3]. Fatigue
failure of shovel components may cause unplanned breakdowns and downtimes resulting in lower operational efficiency and higher
production costs.

2. Fatigue failure modeling procedure

The authors presented a review of the formative resistive forces and strategy for fatigue failure modeling of shovel front end
components in an earlier study [4]. In this research, the fatigue failure modeling of a shovel front-end assembly is presented as a four-
step procedure (Fig. 1):

1. Modeling and computations of dynamic formation resistive forces during shovel digging;
2. Kinematic and dynamic modeling of the shovel front-end assembly;
3. Stress profiling of the shovel’s dipper-teeth assembly through virtual prototyping and numerical simulation; and
4. Crack propagation modeling for fatigue life estimation of components.

The focus of this paper is the crack propagation and fatigue life estimation. The first three steps have been reported in a previous
research and are briefly described here.

2.1. Formation resistive forces

A review of the shovel resistive forces is given in Raza and Frimpong [4] and was also discussed by Blouin, Hemami and Lipsett [5],
Hemami, Goulet and Aubertin [6], and Takahashi, Hasegawa and Nakano [7]. The model proposed by Hemami, Goulet and Aubertin
[6] consists of six forces (f1 – f6), and is a comprehensive model for resistive forces acting on the shovel dipper during the excavation
cycle as shown in Fig. 2.
f1: Resistive force due to payload
f2: Resistive force due to compression of the material moving towards dipper
f3: Frictional force due to material sliding into the dipper
f4: Resistive force due to penetration of the dipper tip and sidewalls into formation
f5: The inertia force of the payload
f6: The resistive force to move empty dipper. This force is modeled as part of f1.
Force f1 can be modeled using models proposed by either Hemami [8] or Awuah-Offei, Frimpong and Askari-Nasab [9]. Force f2 can
be set to zero by selecting a proper trajectory where the bottom of the bucket stays clear of the material and does not compress the
material [8]. The dipper cutting force (f3) in soil can be estimated by the empirical model developed by Zelenin, Balovnev and Kerov

2
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

Fig. 1. Flow chart of shovel fatigue failure and life expectancy modeling.

[10]. Force f4 can be modeled as a part of f3 [8,9]. The f5 force can be ignored (or zero) if the dipper is assumed to move at a constant
speed through the muck pile [9]. Force f6 is simply the weight of the dipper and is a known force as it depends both on the dipper’s
dimensions and the excavated material. This can be modeled as part of f1.

2.2. Kinematica and dynamic modeling

A kinematic model of a mining shovel describes the motions of the shovel front-end. It provides a framework to include the resistive
forces and to build a dynamic model to calculate the torques and forces on individual components. Collectively, kinematic and dy­
namic models describe the full motion and forces during the shovel digging cycle.
An iterative procedure is generally adopted to transfer the motions and forces from point to point (or link to link) for the kinematic
modeling. Table 1 gives the structural parameters required for this modeling using the Denavit-Hartenberg (D-H) notation [11].
Kinematic and dynamic models were developed by Daneshmend, Hendricks, Wu and Scoble [12], Frimpong, Hu and Awuah-Offei
[13] and Awuah-Offei, Frimpong and Hooman [14] that included the dynamic payload forces using Newton-Euler or Lagrange for­
mulations. A detailed description of the step-by-step process for the dynamic modeling basics and procedures is available in
[11,13,15–18].
The general dynamic model for the shovel can be defined as in Eq. (1) [18,19].

D(Θ̇)Θ̈ + C(Θ, Θ̇)Θ̇ + G(Θ) = F − Fload (Ft , Fn ) (1)

3
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

Fig. 2. Forces on a dipper during excavation [6,17].

Table 1
Structural parameters for shovel kinematic modeling [17].
Link i Joint Description αi ai di θi

1 Saddle – Boom joint 0 0 0 θ1

2 Saddle –Dipper-handle joint 90 a1 0 0
3 Dipper-handle – Dipper joint − 90 0 d2 0
4 Dipper Tip 0 L2 0 β

Fig. 3. Structural parameters and external forces during shovel digging [17] .

4
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

where
D(Θ) = massmatrix

C(Θ, Θ̇) = centrifugalandcoriollisterms

G(Θ) = gravityterms
Using the information given in Fig. 3 and Table 1, the Newton-Euler iterative method was used to define the reverse kinematic
model for a shovel as in Eqs. (2) and (3). These equations define the dipper arm extension and the rotation around the saddle block for a
given dipper-tip coordinates. Alternatively, these two equations are the required dipper arm extension and the rotation to properly
place the dipper tip at the required coordinates on the digging trajectory [17].
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
d2 = a2 sβ + p2x + p2y − a21 − a22 − 2a1 a2 cβ + (a2 sβ )2 (2)

( √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ ) ( )
θ1 = Atan2 a2 sβ − d2 , ± p2x + p2y − (a2 sβ − d2 )2 − Atan2 py , px (3)

The formation resistive forces, as explained in 2.1, are used as input to the kinematic model to compute the required crowd arm
force and hoisting torque required to overcome the resistive forces at the saddle block. The resultant dynamic model, computed using
the Newton-Euler iterative algorithm, is given in Eq. (4). This model can be programed in a solver, such as Matlab/Simulink, to
compute the hoisting torque (and crowd force requirements for the cable shovel dipper.
⎡ ( ) ⎤
d2
0 m2 + m3 d2
{ } ⎢
⎢ 2 ⎥{ }
⎥ d̈2
F1
=⎢ ⎢ ( ( ) (
⎥
) ⎥ θ̈
τ1 ) 2 2 2
⎣ a3 ( ) a d a ⎦ 1
− (m2 − m3 )a1 + m3 Izz1 + Iyy2 + Izz3 + m1 1 + m2 2 − a21 − m3 a21 − 3 − d22
2 4 4 4
⎡ ( a (a )) ⎤
1 3
2(m2 + m3 )θ̇1 − m1 + m2 a1 + m3 + a1 θ̇1 { }
⎢ 2 2 ⎥ ḋ2
+⎢⎣ ( d2 ) ⎥
⎦ θ̇1 +
2 m2 + m3 d2 θ̇1 − 2m3 a1 d2 θ̇1
2
⎡ ⎤
(m1 + m2 + m3 )gs1 { }
⎣ ((m ) ) ( a ( a3 ) ) ⎦ d2
2 1
+ m3 gs1 + m1 − m3 a1 − + a1 m2 gc1 θ1
2 2 2
[ ]
Fa
+ Ấ (4)
Fa − a1 Fb + a3 Fd + a3 Fe

2.3. Numerical simulation

A numerical simulation model was developed in Matlab/Simulink to solve the Eq. (4) using the embedded numerical Runge-Kutta
algorithm [20]. The simulation model is a collection of sub-models that compute dipper trajectory, the extension and rotation of the
crowd-arm, and the external formation resistive forces on the dipper during the digging cycle. For this purpose, a virtual working
bench was created (Fig. 4), and the dynamic forces were computed during the digging process using the shovel dimensions presented in

Fig. 4. Virtual test bench for dipper trajectory modeling.

5
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

Table 2
P&H 4100XPC Shovel Dimensions.
Dipper Capacity (m3) 28.44
Dipper Inner Width (m) 4.6241
Mass of Dipper (kg) (link 3) 14,802
Mass of Crowd-arm (kg) 40,134
Mass of Saddle Block (kg) 16,207
Saddle Block Length (m) 1.1162
Length from Crowd-Arm to Dipper-Lip (m) 3.3004

Table 2. The output of the numerical simulation model was used as input to the virtual prototyping for stress profiling of the dipper.

2.4. Shovel virtual prototyping and stress profiling of Dipper-Teeth assembly

A virtual prototype of the shovel front-end assembly was developed in ANSYS (R15) software environment as shown in Fig. 5 [17]
where the dimensions for the shovel were estimated from a 3D model of P&H XPC 4100. The front-end geometry was simplified, too
and the geometric complexities were ignored.

2.5. Dipper stress profiling

A transient analysis was performed in ANSYS Workbench R15 for stress profiling of the dipper-and teeth. Von-Mises stresses were
computed for the dipper-teeth assembly using the prototype and using the resistive forces as the input. The resulting von-Mises stresses
for the dipper are presented in Fig. 6.
Fig. 6 shows that the sidewall experiences high stress during the excavation process. The top corner is under a high stress than the
rest of the sidewall. The stress on the rest of the sidewall is fairly uniform, except for the bottom corner, which shows lighter stress load.
These areas are the focus for the fracture analysis.

2.6. Fatigue and fracture life modeling of dipper

The dipper-teeth assembly engages directly with the formation, and the variations in cutting and weight resistive forces results in
stress loading of the shovel front-end assembly. The shovel digging process is cyclic and is generally a continuous process for shovel

Fig. 5. A simplified 3D virtual prototype of cable shovel’s front-end assembly.

6
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

Fig. 6. Stress profile (equivalent von-Mises stresses) of the dipper [17] .

operation. This cyclic digging and loading processes result in fatigue cracking of the front-end components. The initiated fatigue
cracks, and the existing cracks from manufacturing flaws, can propagate to critical lengths and may result in component failures. These
failures can also lead to expensive repairs and shovel downtimes. Pearson, Hannen, Soderberg and construction [21] reported a sudden
breaking down of a large barge-mounted hydraulic excavator boom due to fatigue cracks reaching a critical length.
Metal fatigue is a complex metallurgical phenomenon and is dependent on multiple factors including metal microstructure and
operating external environment. Common fatigue failure methods include stress life, strain life, and fracture mechanics methods.
Given that the shovel is subject to high stresses with pre-existing cracks from internal flaws or high impacts, this research used the
fracture mechanics approach to estimate a crack propagation life. The initial crack lengths are either known (such as welds,
manufacturing defects) or assumed as part of the process. Fracture mechanics approaches are applied to estimate the crack propagation
rates and fatige lives.
Fatigue failure of metals generally have three modes: - Mode-I (crack opening), Mode-II (in-plane shear or crack opening), and
Mode-III (out-of-plane shear or crack twist). Metal failure can also result from mixed-mode fatigue. Mode-I fatigue research has

Fig. 7. A representative crack growth curve with three identifiable regions: (i) Crack Initiation; (ii) Crack Propagation; and (iii) Rapid Growth
to Failure.

7
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

dominated the fatigue analysis and life-expectancy studies. A typical crack growth curve has three identifiable regions as (i) crack
initiation, (ii) crack propagation, and (iii) rapid crack growth to failure (Fig. 7). Any distinction between the initial two phases is nearly
impossible to make. The plastic behavior around the notches can be attributed to the crack-initiation phase. A major portion of crack
life is spent during the crack propation phase. A number of models are available to predict the crack propagation phase (the middle
region on the curve) [22]. Paris and Erdogan [23] developed a model, known as Paris’ Law in equation (5), is a commonly used
technique for crack life estimation.
da
= C(ΔK)m (5)
dN
The crack growth rate is defined by the gradient of the middle linear region of the curve in Fig. 7. The factor ‘K’ in Eq. (5) is the
stress intensity factor (SIF). Numerical and finite element-based approaches are commonly used to compute SIFs for complicated
geometries. Yin, Grondin, Obaia and Elwi [24] used finite element method to estimate the crack growth rate and estimated the fatigue
life for corner cracks in the steel welded box section of the shovel boom.
The fatigue life in Eq. (6) can then be computed for a known crack-length by integrating Eq. (6) [25].
∫af
da
Nf = (6)
C(ΔK)m
ai

Fig. 8. Side-wall crack locations and orientations. a) top-corner, b) center of sidewall.

8
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

3. Experimentation for fatigue crack propagation

A number of parameters control the crack failure modeling. These parameters include the crack geometry, shape, orientation and
stress environment. A series of experimentation is run to investigate the crack-propagation of pre-defined cracks at the selected lo­
cations of the dipper in specific orientations.

3.1. Crack locations and geometries

Cracks can appear at any location of the dipper-teeth assembly or at any stressed location of the shovel front-end. The dipper-teeth
assembly is huge in size and there can be an infinite number of possible combinations of crack-parameters (geometry, shape, orien­
tation, and stress environment). The simulations are computationally expensive and require longer run times. Given the expensive
computations, analyzing cracks at all the locations is not optimal. Therefore, experimentation for crack failure and life estimation was
designed at the selected locations of the dipper-teeth assembly. These locations (Fig. 6) include highly stressed dipper regions, the
centers of sidewall and bottom-plate. These selected, critical locations were the focus areas for further crack and life-estimation
experimentation.
A separate series of experiments was run at each selected location of the dipper-teeth assembly. Pre-defined, semi-elliptical cracks
were introduced at these locations, and were gradually increased with fixed increments. The crack-depths were maintained at half of
the crack-lengths for each crack increment. The semi-elliptical crack shape represents the crack front (Figs. 8 and 9). Contours were
created around this crack-front (Fig. 10), and J-integrals were computed over each contour. These J-integral values were later used to
compute the SIFs at the crack-tip.
For the sidewall experiments, semi-elliptical cracks were introduced at the center of the sidewall and the top corner (Fig. 8). For
each location, cracks were simulated in two orthogonal orientations (parallel and normal to the bottom-plate) as shown in Fig. 8. These
cracks were gradually incremented from a quarter of an inch to three inches with a step size of 0.5 in. The depth of these cracks was
incremented from one-eighth of an inch to 1.5 in. (half of the crack length).

Fig. 9. Bottom-plate crack locations and orientations. a): closer to side-wall b): center of bottom-plate.

9
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

Fig. 10. Crack geometry definition computations of J-integral in ANSYS.

Similarly, for the bottom-plate experiments, semi-elliptical cracks were introduced at the center of bottom-plate and at a second
location closer to the sidewall (Fig. 9). For the central location, cracks were simulated in two orthogonal orientations which are parallel
and normal to the sidewall. While for the location closer to sidewall, only one orientation of the crack was simulated. This direction
was selected along the maximum stress intensity direction as was established during the stress profiling of the dipper. These two crack
locations, orientations, and the dipper stress profile are shown in Fig. 9.
These cracks were gradually incremented from a quarter of an inch to three inches with a step size of 0.5 in. The depth of these
cracks was incremented to half of the crack-length in every simulation.

3.2. Crack definition in ANSYS

The crack definition is explained in Fig. 10 for this experimentation. For each crack, a localized coordinate frame was assigned at a
point that served as the center of the semi-elliptical crack. The crack-length (major radius of ellipse) and the crack-depth (minor radius

Table 3
Experimentation for fracture modeling and analysis.
Crack Location Description

Side-Wall Crack Description: SIF computations and life-estimations for cracks at selected locations for the sidewall
Variable: crack size, Orientation
Sizes: 0.25′′ , 0.5′′ , 0.75′′ 1.00′′ , 1.5′′ , 2.0′′ , 3.0′′
Orientation: Parallel to Bottom-Plate, & Normal to Bottom-Plate
No. of Experiments: 28
Bottom-Plate Description: SIF computations and life-estimations for cracks at selected locations for crack in the bottom-plate.
Crack Variable: crack size, Orientation
Sizes: 0.25′′ , 0.5′′ , 0.75′′ , 1.00′′ , 1.5′′ , 2.0′′ , 3.0′′
Orientation: Parallel to Bottom-Plate; Normal to Bottom-Plate
No. of Experiments: 21
Significance To establish the relationship between SIF and crack length (a vs ΔK)
To establish relationship between rate of change of crack-length and SIF (Δa/ΔN vs ΔK)
To estimate the remaining life of component (Nf)
Expected Results The SIF should increase with crack length and should show a logarithmic relationship between rate of change of crack-length and SIF (Δa/ΔN
vs logΔK should be a straight line) as in Paris Law [23].
Cracks in one orientation should grow more rapidly than the other orientation

10
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

of ellipse) were defined along the Z and positive X directions, respectively. The crack width was measured along the Y direction. The
center and the two radii defined the shape of the crack front.
Every crack front was then divided into multiple segments with six circular contours generated around the crack front. The circular
contours were further divided into segments as well. These divisions of contours and crack font became the node locations for the
computation of J-integrals in the finite element model. The SIFs were then computed from these J-integrals, at the crack tips, for life
estimation.
Initial crack definition (length and depth) was a critical aspect in the crack modeling and dipper life-estimation because the SIFs
were dependent upon these. Estimating the SIFs variations with size is the most important and critical aspect for life-estimation and
modeling. Generally, the relationship between SIF and crack length is not linear.
The experiment simulations were conducted in ANSYS R15 environment. Missouri S&T’s high computing facility was used to run
these computationally expensive simulations. The series of experimentations conducted for the fatigue crack-simulations and life-
estimations are summarized in Table 3.

3.3. Numerical computations of SIFs

Stresses were computed at every node around the crack front to generate the stress profile around the crack front. The stress profile
was used to compute the SIFs for each simulated crack. The SIF was calculated at each node of the crack front.
Similarly, simulations are performed for the dipper sidewall. The important fracture modeling parameters for the three repre­
sentative cracks are summarized in Table 4. The geometry of these crack-fronts is semi-elliptical and is defined by major and minor
axes along the Z and X directions, respectively.

4. Experimentation, results and discussions

Before fracture mechanics evolved high-cycle fatigue using the S-N curves was the only method considered for the metal fatigue
failure. Structural failure due to fatigue cracking is now considered one of the major failure types in metal structures. The local stresses
at the crack-tips of these cracks can be high enough to cause the crack to grow to critical extents. At critical extents the material
behavior becomes plastic at the crack-tip and it leads to rapid, often brittle, failure. Before the critical length, the material behavior is
elastic and linear elastic theories are applicable. Computations of SIF at the crack-tip is necessary to predict the crack growth. SIFs for
many simple geometries and loading situations are available in literature (Loadkimidis and Theocaris, 1978; Raju and Newman, 1997;
Sih, 1973; Tada et al., 1973). For complex geometries and stress loading conditions, numerical methods are the only way to compute
the reliable SIF values.

4.1. Finite element computation of SIFs

There are three approaches to compute the SIFs using finite element models. These approaches include displacement, the stress,
and energy methods. The energy method is one of the most commonly used methods for SIFs. This method computes J-integrals over a
closed path around the crack-tip to compute SIFs [26,27]. The energy release rate, computed through this J-integral, is proportional to
the mode-I (K1) SIF [26,27].
In this research, J-integrals were computed, using finite element method and a fine mesh, around the pre-defined crack tips using
ANSYS R15. SIFs were computed, using plane stress conditions, and an average value of SIFs for five contours (contour 2 to 6) were
used for further fatigue analysis. Both the plane stress and plain strain conditions are possible for the computation of SIFs. Here, plane
stress conditions are assumed considering the depth of the crack in relation to the wall thickness. Further, the plane stress conditions
will not restrict the crack propagation through wall thickness. The J-integral values for the first contour were ignored, being very close
to the tip, for accurate results. SIFs were computed for all the cracks at the selected locations as identified in Fig. 6. Crack lengths were
incremented gradually at these selected locations and SIFs were computed for each of the crack-size. The results were later used to
generate the crack-growth curves and for life-expectancy of dipper components. The following descriptive nomenclature was used for
the cracks at the selected locations of dipper-teeth assembly and is shown in Fig. 11. These locations were selected to have a thorough
understanding of the SIF variation with crack length for the sidewall and bottom-plate. Two orthogonal directions were selected to
model the crack orientation impact on SIF.
The SIF variation curves, at each crack-tip, were obtained through a curve fitting process. These variation curves for the simulated
cracks at the center and at the top-corner locations of sidewall are shown in Fig. 12 while for the bottom plate cracks are shown in
Fig. 13. The corner cracks (I and II) represented a higher and steeper increase in the SIFs as compared with the SIFs of cracks at the
center of sidewall (cracks III and IV). This may be attributed to the higher stress concentration at the corner locations. Further, both the
vertical cracks showed higher values than the horizontal cracks, indicating faster propagation rates for the vertical cracks. The crack-V
was located in a high stress region and consequently. therefore, the associated SIFs represented a steep increase with crack size. Given
the high SIFs, these cracks are expected to propagate rapidly.
The SIF variation equations at the crack-tips, at selected locations, are given as Eq. (7) through Eq. (13) for cracks I to V,
respectively.

SIF = 7.23E − 04a3 − 1.06E − 01a2 + 5.17a + 2.66E + 01 (7)

11
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

Table 4
Fracture modeling parameters of representative cracks.
Parameter Description Crack-1 Crack-2 Crack-3

Location Bottom Plate Side Plate Corner Tooth

Major Axis (m) Length of crack 0.0762 0.0762 0.0254
(3-inches) (3-inches) (1-inches)
Minor Axis (m) Depth of crack 0.0254 0.0254 0.0127
(1-inch) (1-inch) (0.5-inch)
Mesh Contours No. of contours around crack-front 6 6 6
Circumferential Divisions No. of division of contour circle 8 8 8
Crack-Front Divisions No. of nodes to define crack-front 18 18 18

Fig. 11. Crack nomenclature for life estimation.

SIF = 3.17E − 04a3 − 5.53E − 2a2 + 3.46E + 00a + 2.61E + 01 (8)

SIF = 4.04E − 04a3 − 6.32E − 02a2 + 3.57E + 00a + 2.25E + 01 (9)

SIF = 3.31E − 04a3 − 5.21E − 02a2 + 2.99E + 00a + 1.85E + 01 (10)

SIF = 4.98E − 04a3 − 8.75E − 02a2 + 5.14a + 2.96E + 01 (11)

SIF = 7.28E − 04a3 − 1.67E − 2a2 + 1.31E + 00a + 1.43E + 01 (12)

SIF = 9.71E − 05a3 − 1.98E − 02a2 + 1.29E + 00a + 1.06E + 01 (13)

4.2. Crack propagation curves

The fatigue crack propagation can be modeled by integrating Eq. (5) as Eq. (6). Eq. (6) has three important input parameters
including two material constants (c, m), and “ΔK”. The two material constants need to be accurately established in a laboratory
following standard procedures. Values for these material constants for commonly used materials may be found in literature as well. The

12
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

Fig. 12. Stress intensity factors for side-wall crack-tips.

Fig. 13. Stress intensity factors for the bottom-plate crack-tips.

‘c’ values generally vary between 3 and 4. For this research, we used the material constants suggested by Yin, Grondin, Obaia and Elwi
[24] during a crack growth analysis for shovel boom cracks and reported the material constants (‘c’ and ‘m’) in laboratory settings
using a standard procedure (ASTM standard E1820). Dipper is assumed to be made of the same steel for this research. With these
parameters, Eq. (6) takes the form of Eq. (14) that is solved numerically using Gauss-Legendre quadrature in MATLAB.
∫af
da
Nf = (14)
5.89− 12 (ΔK)3.27
ai

The output from the Eq. (14) represents the number of cycles (Nf) required for a crack to propagate from an initial length (ai) to a
final length (af). The cycles count is not a useful output for shovel digging operation. A more useful parameter for field operations is the

13
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

number of days for crack-propagation, because it can be used to check and set safe maintenance intervals. Therefore, the number of
cycles must be converted into days.
For this purpose, one digging cycle is assumed to be equivalent to one fatigue cycle. Using this approach total number of fatigue
cycles will be equivalent to the number of digging cycles of the shovel. The equivalent damage can then be estimated, in number of
days, following the Palmgren-Miner’s Rule [28]. The shovel cycle time can be used to estimate the number of days adjusted from
shovel operational efficiency. A complete duty cycle (digging, loading, crowding) for a large shovel averages about 30 s. Using a 30-
second cycle time the total number of digging (and hence fatigue cycles) are calculated as 2730 cycles per day with 95% operational
efficiency. This computation closely matches with the field observations recorded by Yin, Grondin, Obaia and Elwi [29] whereby they
counted 2880 cycles per day for a cable shovel working continuously over a period of two weeks. For this research, 2800 cycles per day
are assumed to convert the number of cycles to number of days.
Crack propagation curves for the Crack-I to Crack-V are shown in Fig. 14. The Crack-V shows the highest propagation rate, while
the Crack-IV shows the slowest. The crack curves VI and VII are in low stress regions and consequently their respective propagation
curves were very slow and hence are not included in further fatigue life estimation.
The Crack-V shows the highest crack growth rate. This was expected, as the crack is in the higher stress region and is slected along
the stress contour (Fig. 11). The Crack-I is also at a high stress region, however, that region is smaller as compared with the region for
Crack-V. The vertical Cracks (I and III) show higher propagation rates as compared with their horizontal counter parts (Cracks II and
IV). The Crack-IV, being horizontal and at a low stress region shows the least propagation of the five cracks. Two important deductions
can be made from these cracks. First, the cracks in the high stress region have higher propagation rates. Second, the vertical cracks in
the side-wall have higher propagation rates than the horizontal cracks.

4.3. Remaining life expectancy of dipper components

The remaining useful life for the cracked dipper components can be estimated with a knowledge of critical crack lengths for dipper
material. The critical crack-length represents the Phase-III of the crack curve (Fig. 7) and represents the length where the material at
the crack tip bechaves like a pastic material with rapid crack growth to failure. Standardized tests like fatigue toughness tests are
generally used to define the critical lengths for the metals. In some cases, the critical length limit may also be implemented based on
field operating conditions or using the crack-growth curves.
A maximum crack-length or crack-growth rate can be implemented for a shovel component in the field. This maximum crack-length
limit could be based on the available crack measurement instrument or technique during the scheduled maintenance. Similarly, a
maximum limit on crack propagation rate can be set based on the scheduled maintenance interval. This will ensure that no crack grows
to a critical limit before the next scheduled maintenance. Yin, Grondin, Obaia and Elwi [24] and Yin, Grondin, Obaia and Elwi [29]
used a field limit of 4.0 mm/day leading to a 204 mm crack length for a boom crack operating in oil-sands formations.
From Fig. 14, it is noted that the crack propagation rates become very high once the crack has propagated to a certain length. A
critical crack length limit can be set based on these propagation rates. For Cracks I to V this critical length may be set at around 100-mm
length. Crack growth curves are more influenced by the initial crack length (ai) than the final crack length (af). The estimated life for

Fig. 14. Crack propagation curves for the cracks I to V. Life expectancy and critical crack-length estimation for cracks with initial lengths of 50 mm.

14
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

cracks, with initial crack of 50 mm, is plotted in Fig. 14. The crack growth curves are not linear, and the crack grows rapidly as it gets
longer. Further, the figure shows an estimated life of 38 days for a 50 mm crack (Crack-V). However, the remaining life is reduced to 16
days once the crack has grown to 75 mm.

5. Conclusions

Shovel loading is subject to dynamic resistive forces. These variations result in severe stress loading of large shovel front-end
components, especially the dipper-teeth assembly. This may lead to sudden component failure and unplanned and costly down­
time. This research has provided a framework for fatigue failure modeling of a large mining shovel. A simpliefied virtual prototype for
the front-end assembly of a large mining shovel (P&H XPC 4100) was built using a scaled model. The virtual prototype of the shovel
front-end assembly may be effectively utilized for fatigue modeling and life-expectancy of dipper-teeth assembly. For this purpose,
representative cracks were modeled at selected locations of dipper side and bottom plates. The J-integrals were computed for each
crack-tip to calculate SIFs. These cracks are incremented in small step sizes and the resulting SIFs are computed for each increment. The
results showed that the SIFs had cubic relationship with the crack lengths and were dependent on the local stresses and crack ori­
entations. The cracks in the sidewall had higher crack-propagation rates than most of the cracks in other locations because of the higher
stresses. The results also showed that crack propagation rates were dependent on orientations. The vertical cracks on the sidewalls had
higher growth rates than their horizontal counterparts in the same region. Crack propagation rates were not linear and increased
rapidly with the crack initial length. A 100 mm crack length is critical for dipper components. At 100 mm crack length, the crack
propagation rates are high, and the material can be unstable. Further, the estimated life of 38 days for a 50 mm crack (Crack-V).
However, the remaining life is reduced to 16 days once the crack has grown to 75 mm for the same material.
The research used simplified geometry for a mining shovel with dimensions estimated from a 3D scaled model of pH 4100 XPC
shovel. The actual wall dimensions, material properties, and the detailed geometry were not used so these results cannot be directly
applicable to PH4100 XPC or similar large mining shovels. This research has provided a framework for shovel dipper-teeth fatigue
modeling and analysis, whose results are useful for shovel designs and manufacturing for the health and longevity of next generation
mining shovels. The framework developed in this research lays a foundation for a detailed analysis and crack-life estimation for the
dipper-teeth assembly. In the future, we plan to extend this research and map the dipper-teeth assembly for a range of crack lengths
using the actual model and material properties. The fatigue-life map would help the field engineers and operators for estimating more
realistic maintenance intervals after field testing and validation of the model. Further, this would help the manufactures and equip­
ment design engineers by providing a scientific basis for shovel health and longevity that has been largely experience-based so far.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

References

[1] Komatsu, Electric Rope Shovels, in: 2018.

[2] Caterpillar, Electric Rope Shovels, in, Caterpillar, 2018.
[3] S.K. Roy, M.M. Bhattacharyya, V.N.A. Naikan, Maintainability and reliability analysis of a fleet of shovels, Trans. Inst. Min. Metall. (2001) A163–A171.
[4] M.A. Raza, S. Frimpong, Cable shovel stress & fatigue failure modeling-causes and solution strategies review, J. Powder Metall. Min. (2013).
[5] S. Blouin, A. Hemami, M. Lipsett, Review of resistive force models for earthmoving processess, J. Aerospace Eng. 14 (2001) 102–111.
[6] A. Hemami, S. Goulet, M. Aubertin, Resistance of particulate media to excvation: application to bucket loading, Int. J. Surface Min. Reclamation Environ. 8
(1994) 125–129.
[7] H. Takahashi, M. Hasegawa, E. Nakano, Analysis on the resisitive forces acting on the bucket of a load-hual-dump machine and a wheel loader in the scooping
task, Adv. Robot. 13 (1999) 97–114.
[8] A. Hemami, An approximation of the weight of the loaded material during the scooping operation of a mechanical loader, Trans. Can. Soc. Mech. Eng. 18 (1994)
191–205.
[9] K. Awuah-Offei, S. Frimpong, H. Askari-Nasab, Formation excavation resistance modeling for shovel dippers, Int. J. Min. Miner. Eng. 1 (2009) 127–146.
[10] A.N. Zelenin, V.I. Balovnev, I.P. Kerov, Machines for Moving the Earth, Amerind Publishing New Delhi, India, 1985.
[11] A.J. Koivo, Fundamentals for Control of Robotic Manipulators, Jhon Wiley & Sons, Inc., 1989.
[12] L. Daneshmend, C. Hendricks, S. Wu, M. Scoble, Design of a Mining Shovel Simulator, Innovative mine design for the 21st century: Proceedings of the
International Ccongress on Mine Design, Ontario, Canada, (1993) 551-561.
[13] S. Frimpong, Y. Hu, K. Awuah-Offei, Mechanics of cable shovel-fromation interactions in surface mining excavations, J. Terramech. Elsevier Science Ltd. (2005)
15–33.
[14] K. Awuah-Offei, S. Frimpong, A.-N. Hooman, Formation excavation resistance modeling for shovel dippers, Int. J. Min. Minera. Eng. 1 (2009) 127–146.
[15] A.J. Koivo, Kinematics of excavators (backhoes) for transferring surface material, J. Aerospace Eng. 7 (1994) 17–32.
[16] A.J. Koivo, M. Thoma, E. Kocaoglan, J. Andrade-Cetto, Modeling and control of excavator dynamics during digging operation, J. Aerospace Eng. 9 (1996)
10–18.
[17] M.A. Raza, S. Frimpong, Mechanics of electric rope shovel performance and reliability in formation excavation, in: Lagrangian mechanics, Intech, 2017, pp.
107–133.
[18] J.J. Craig, Introduction to Robotics: Mechanics and Control, 2nd ed., Addison Wesley, 1996.
[19] S. Frimpong, Y. Hu, H.J.I.J.o.G. Inyang, Dynamic modeling of hydraulic shovel excavators for geomaterials, 8 (2008) 20–29.
[20] I. Simulink, V.P.B.J.R.m.r.h.h.t. Tutorials, References http://www.mathworks. com/help/releases, in, 2020.
[21] J.E. Pearson, W.R. Hannen, E.J.P.p.o.s.d. Soderberg, construction, Development of fatigue monitoring system for a hydraulic excavator, 9 (2004) 221–226.
[22] A. Fatemi, L.J.I.j.o.f. Yang, Cumulative fatigue damage and life prediction theories: a survey of the state of the art for homogeneous materials, 20 (1998) 9–34.
[23] P. Paris, F. Erdogan, A critical analysis of crack propagation laws, (1963).

15
M. Azeem Raza and S. Frimpong Engineering Failure Analysis 121 (2021) 105110

[24] Y. Yin, G. Grondin, K. Obaia, A.J.J.o.C.S.R. Elwi, Fatigue life prediction of heavy mining equipment. Part 2: Behaviour of corner crack in steel welded box
section and remaining fatigue life determination, 64 (2008) 62–71.
[25] J. Bannantine, J. Comer, J.L. Handrock, Fundamentals of metal fatigue analysis, 1990.
[26] J.R. Rice, A path independent integral and the approximate analysis of strain concentration by notches and cracks, 1968.
[27] J.R. Rice, Mathematical analysis in the mechanics of fracture, Fract.: Adv. Treatise 2 (1968) 191–311.
[28] M. Miner, Cumulative fatigue damage, J Appl Mech 12 (1945) A159–A164.
[29] Y. Yin, G. Grondin, K. Obaia, A.J.J.o.C.S.R. Elwi, Fatigue life prediction of heavy mining equipment. Part 1: Fatigue load assessment and crack growth rate tests,
63 (2007) 1494–1505.

16