Available online at www.sciencedirect.

com

ScienceDirect
Available online at www.sciencedirect.com

ScienceDirect
Procedia Manufacturing 00 (2019) 000–000
www.elsevier.com/locate/procedia

Procedia Manufacturing 51 (2020) 848–855

30th International Conference on Flexible Automation and Intelligent Manufacturing (FAIM2021)

15-18 June 2021, Athens, Greece.
Evaluation of T-joints in aluminium structures under different geometries
F.J.P. Moreiraa, R.D.S.G. Campilhoa,b,*, M.G. Cardosoa, F.J.G. Silvaa
ISEP-School of Engineering, Polytechnic of Porto, Rua Dr. António Bernardino de Almeida, 431, 4200-072, Porto, Portugal
aa

bb
INEGI, Institute of Mechanical Engineering, Rua Dr. Roberto Frias, 4200-465, Porto, Portugal
* Corresponding author. Tel.: +351939526892; fax: +351228321159. E-mail address: [email protected]

Abstract

The adhesive bonding technique is nowadays very popular in industrial applications, and is gradually replacing other more traditional bonding
methods (fastened, welded and riveted joints) due to some advantages. However, its application supposes accurate methods for strength
prediction. As a result, the techniques to predict the strength of adhesive joints has highly evolved. The eXtended Finite Element Method (XFEM)
is a recent innovation implemented over the (Finite Element Method) FEM that enables crack growth to be modelled. However, its application
to adhesive joints is still scarce. This work consists of an experimental and XFEM analysis of aluminium alloy T-joints, adhesively-bonded with
three adhesive types. A parametric study is undertaken regarding the curved adherends’ thickness (tP2 P2), with values between 1 and 4 mm. The
adhesives Araldite®® AV138 (strong but brittle), Araldite®® 2015 (less strong but moderately ductile) and the Sikaforce®® 7752 (with the smallest
strength but highly ductile) were tested. A comparative analysis between the different joints conditions was undertaken by plotting peel (yy) and
shear (xy
xy) stresses. The XFEM predictive capabilities were tested with different damage initiation criteria. It was found that, provided that the
modelling conditions are properly set, accurate numerical results can be found.

©2020
© 2020The TheAuthors.
Authors. Published
Published by Elsevier
by Elsevier Ltd. Ltd.
Thisisisananopen
This open access
access article
article under
under theBY-NC-ND
the CC CC BY-NC-NDlicenselicense (https://creativecommons.org/licenses/by-nc-nd/4.0/)
https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-reviewunder
Peer-review underresponsibility
responsibilityof of
thethe scientific
scientific committee
committee ofFAIM
of the the FAIM
2020.2021.

Keywords: Fracture; Finite element analysis; eXtended Finite Element Method; Bonded joint.

1. Introduction highly efficient when compared to SLJ because of the reduction

of stress concentrations [1]. Although these types of joints are
The use of adhesive bonds greatly increased in industrial the most used in the industry, other types of joints have specific
applications, as they have multiple advantages compared to applications. Stepped-lap configurations can be used in
other more traditional bonding methods (fastened, welded and composite joining due to the easiness to make the step design
riveted joints). The aeronautical, naval, automotive and during the materials’ fabrication process [2]. T-joints find
aerospace industries are good examples where adhesive joints application in the naval and aeronautical industries. In the naval
are widely applied. More uniform distribution of stresses, ease industry, they allow joining panels with the hull [3] and the
of manufacture, possibility of joining different materials and fiberglass hull with anti-flood panels [4]. In the aeronautical
low cost are the main advantages of adhesive bonds. The main industry, they are used to join wing panels and fuselage
disadvantages are related to the requirement of surface sections [5]. Several works were carried out to evaluate T-
preparation, low peel strength and difficulties in quality control joints, using either analytical or numerical techniques [6, 7].
and safety. The most common adhesive joint configurations are The number of approaches to predict the strength of
single-lap joints (SLJ), double-lap joints (DLJ) and scarf joints adhesive joints has increased over the years. Actually,
[1]. SLJ are the most common. However, they develop major analytical and numerical techniques have become more refined
yy peak stresses. DLJ are more difficult to manufacture but, on and with higher accuracy. The FEM allows modelling complex
the other hand, yy stresses greatly diminish. Scarf joints are geometries with precision, due to computational advancements
2351-9789 © 2020 The Authors. Published by Elsevier Ltd.
2351-9789
This is an©open
2020 The Authors.
access Published
article under bytheElsevier Ltd.
CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
This is an openunder
Peer-review access article under the CC
responsibility BY-NC-ND
of the license
scientific https://creativecommons.org/licenses/by-nc-nd/4.0/)
committee of the FAIM 2021.
Peer-review under responsibility of the scientific committee of the FAIM 2020.
10.1016/j.promfg.2020.10.119
 F.J.P. Moreira et al. / Procedia Manufacturing 51 (2020) 848–855 849
2 Moreira et al. / Procedia Manufacturing 00 (2019) 000–000

and Computer Aided Engineering (CAE) tools. The use of being coincident with the mesh or the necessity to correct the
Continuum Mechanics supposes using the obtained stresses or mesh in the crack vicinity [19]. The XFEM is based on the
strains, whose maximum values are used in appropriate failure concept of partition of unity, and its implementation in the
criteria to assess failure. However, this technique has limited FEM can be accomplished by introducing local enrichment
applicability because of stress singularities (which make the functions for the displacements near the crack tip, allowing
predictions dependent on the applied mesh) and neglecting of damage to grow and respective separation between the cracked
fracture mechanics concepts [8]. Actually, in a bonded joint faces [20]. Mubashar et al. [21] carried out a study on the
FEM analysis, stresses near the singular regions increase with damage and failure modelling of adhesively-bonded SLJ with
the mesh refinement, making convergence impossible [9]. spew fillets at the overlap ends, combining two methods:
Traditional Fracture Mechanics-based techniques can be XFEM (to perform the modelling of the crack in the fillet
applied to the study of the behaviour of structures that contain region where the crack path is unknown) and CZM (applied to
defects, such as cracks. These cracks can result from stress model crack progression and damage along the adhesive bond
concentrations, usually located in holes, notches or interfaces interface). The numerical analysis was performed in Abaqus ®.
between different materials. However, it is not mandatory that Aluminium alloy 2024 T3 adherends were bonded with the
the structures to be analysed already have cracks, which is a epoxy adhesive FM73-M, and the adhesive was modelled with
limitation of this method [10]. Cohesive Zone Models (CZM) elastoplastic properties, obtained in tensile tests. This work
were developed to describe damage under static loads in the allowed to conclude that the XFEM is capable of predicting,
cohesive process zone around the crack tip. They are based on with a high degree of precision, the crack onset location and
cohesive elements, which allow connecting solid elements of path within the spew fillet. Moreover, it is possible to combine
two-dimensional (2D) and three-dimensional (3D) structures, the XFEM with CZM to more accurately predict crack
using pre-established traction-separation laws [11]. CZM were initiation and growth in bonded joints, including at the
tested and optimized to promote structural damage initiation interface. However, this technique is limited by the existence
and crack propagation simulations on cohesive and interfacial of a potential discontinuity in the crack at the XFEM-CZM
fracture problems, and delamination in composites. The use of transition. Stuparu et al. [22] conducted a study on the
CZM to model structures enables to create one or more regions combined use of CZM and XFEM for the strength prediction
or interfaces in which damage nucleation and growth is made of bonded joints. The SLJ configuration was tested, with
possible by the softening and release of homologous nodes of aluminium adherends and the adhesive Araldite® AV138. The
the cohesive elements [12]. FEM simulations based on following parameters were used: adherends’ thickness (tP) of 5
continuum mechanics wrongly consider that the solid elements mm, adhesive thickness (tA) of 1, 3 and 5 mm, overlap length
undergo plasticization without taking damage. Damage (LO) of 20 mm and sample width (B) of 25 mm. The numerical
mechanics simulations work by inducing damage to the analysis was done in Abaqus®. The XFEM was used to
elements through the reduction of transmitted loads between simulate failure within the adhesive, considering a strain
solid elements. Thus, it is possible to perform the simulation of criterion (less mesh dependent than stress criteria) for crack
crack growth, in which the cracks can assume a pre-defined onset prediction. Thus, crack initiation/propagation will always
trajectory or an arbitrary trajectory within a finite region [8]. In take place orthogonally to the maximum principal strains. On
Damage Mechanics, a damage parameter is established to the other hand, CZM was equated to simulate an interfacial
cause a change in the response of the constituent materials failure between the adhesive and adherends. Different tA and
through the depreciation of the strength or stiffness, as occurs the positions of initial bonding flaws were tested, resulting in
in adhesive layers, or in composite delaminations, to model modifications of the XFEM crack trajectories, eventually
damage during loading [13, 14]. The insertion of a damage attaining the interface. It was shown that the use of XFEM is
variable in the constitutive law of the material enables well complemented by CZM to promote crack growth after the
simulating damage before and after crack nucleation. Two XFEM crack attained the interface.
types of damage variables can be introduced in the models: This work consists of an experimental and XFEM analysis
variables that empirically depreciate the properties of the of aluminium alloy T-joints, adhesively-bonded with three
materials, without any relation to the damage mechanism, and adhesive types. A parametric study is undertaken regarding tP2,
variables that have a physical significance, by directly relating with values between 1 and 4 mm. The adhesives Araldite®
to the observed type of damage (for example the size of AV138 (strong but brittle), Araldite® 2015 (less strong but
porosities or micro-cavities) [15]. The growth of damage is moderately ductile) and the Sikaforce® 7752 (with the smallest
usually ruled by the load function for static simulations [16] strength but highly ductile) were tested. A comparative analysis
and as a function of the number of cycles for fatigue modelling between the different joints conditions was undertaken by
[17, 18]. The XFEM is a recent variant of the FEM to model plotting y and xy stresses. The XFEM predictive capabilities
damage growth in structures, although it is yet seldom studied were tested with different damage initiation criteria.
within the context of bonded joints. This method uses damage
laws to predict fracture, based on strength concepts to infer 2. Experimental work
damage initiation of damage and deformations for failure.
Comparing the XFEM with CZM, the XFEM has the clear 2.1. Adherends and adhesives
advantage of not requiring the crack to follow a predefined path
by the user. This is because crack propagation occurs freely The T-joints are made of three AW6082 T651 aluminium
inside the material, without the geometry of the discontinuities alloy aluminium adherends bond together. This is a high-
850 F.J.P. Moreira et al. / Procedia Manufacturing 51 (2020) 848–855
Moreira et al. / Procedia Manufacturing 00 (2019) 000–000 3

strength alloy, characterized in a previous work [23]. Fig. 1 toughness or GIC) and the End-Notched Flexure (ENF) test
shows typical stress-strain (σ-ε) curves of this aluminium alloy, (shear fracture toughness or GIIC). Table 1 gives an overview
whose relevant properties in bulk tensile testing are: Young’s of the obtained data, which will be used in this work for input
modulus (E) of 70.1±0.8 GPa, tensile yield stress (σy) of in the numerical simulations. To be noted that the values of
261.7±7.7 MPa, tensile strength (σf) of 324.0±0.2 MPa and yield stress were defined considering a plastic strain of 0.2%.
tensile failure strain (εf) of 21.7±4.2%.
2.2. Experimental details
350

300 Fig. 2 represents the geometry and dimensions of the T-

250 joints. The relevant dimensions are the following: LO=25 mm,
200
B=25 mm, base length LT=80 mm, base thickness tP1=3 mm,
 [MPa]

tP2=1, 2, 3 and 4 mm, L-part length LA=60 mm, L-part radius

150
R=5 mm and tA=0.2 mm.
100 Experimental
50 Numerical approximation

0
0 0.05 0.1 0.15 0.2 0.25


Fig. 1. Experimental and numerical σ-ε curves of the aluminium.

The following adhesives were tested in the T-joint

configuration: Araldite® AV138 (brittle epoxy), Araldite® 2015
(ductile epoxy) and the Sikaforce® 7752 (high-elongation
polyurethane). All adhesives were formerly tested and the
respective properties detailed in references [23, 24]. The tensile
mechanical properties (E, σy, σf and εf) resulted from bulk tests Fig. 2. Representative geometry and dimensions of the T-joints.
to dogbone specimens. The fabrication process for these
specimens followed the indications stipulated in the NF T 76- Specimen fabrication was initiated by cutting/bending the
142 French standard. The shear mechanical properties (Shear adherends to the respective shapes. The straight adherend, used
modulus – G, shear yield stress – τy, shear strength – τf and as the specimen base, is obtained by cutting, in an automated
shear failure strain – ϒf) were estimated by Thick Adherend cutter, a bar to the final dimensions. The L-parts were subjected
Shear Tests (TAST). In this case, the 11003-2:1999 ISO to an identical procedure, but then they were manually bent in
standard was considered for fabrication and testing protocols. a press, such that the end surfaces were at an angle of 90º, and
The TAST specimens were made with DIN C45E steel applying a tool with the chosen R, to produce the geometry
adherends, and curing was undertaken in a rigid mould to depicted in Fig. 2. The surface preparation before bonding
guarantee that the cured specimens are aligned [23]. consisted of manually increasing the roughness by sanding
using emery paper with coarse grain (60 grit), followed by
Table 1. Properties of the adhesives Araldite® AV138, Araldite® 2015 and wiping with cleaning agent (acetone) to eliminate dust and
Sikaforce® 7752 [23, 24]. oxides, thus achieving a strong bond. Joining of the different
parts was accomplished in a jig that positioned the three
Property AV138 2015 7752 adherends in the layout of Fig. 2 and enabled keeping this
Young’s modulus, E [GPa] 4.89±0.81 1.85±0.21 0.49±0.09
position throughout the entire adhesive hardening process. To
Poisson’s ratio,  0.35 a 0.33 a 0.30 a
Tensile yield stress, y [MPa] 36.49±2.47 12.63±0.61 3.24±0.48 assure the specified tA, steel spacers with identical thickness to
Tensile failure strength, f tA were inserted at the edges of the bonded parts are removed
39.45±3.18 21.63±1.61 11.48±0.25
[MPa] after adhesive curing. Before placing the spacers, they were
Tensile failure strain, f [%] 1.21±0.10 4.77±0.15 19.18±1.40 initially coated with Loctite® Frekote 770NC demoulding
Shear modulus, G [GPa] 1.81 b 0.70 b 0.19 b
agent, to guarantee easy removal after the adhesive has cured,
Shear yield stress, y [MPa] 25.1±0.33 14.6±1.3 5.16±1.14
Shear failure strength, f which is essential to prevent damage to the cured adhesive
30.2±0.40 17.9±1.8 10.17±0.64 layers. With the specimens in position after depositing the
[MPa]
Shear failure strain, f [%] 7.8±0.7 43.9±3.4 54.82±6.38 adhesive, and with the spacers providing the correct offset
Toughness in tension, GIC between adherends, grips were used to apply pressure to the set
0.20 c 0.43±0.02 2.36±0.17
[N/mm]
and enable curing to take place. This process was accomplished
Toughness in shear, GIIC
[N/mm]
0.38 c 4.70±0.34 5.41±0.47 during a one-week period. The final step consisted of trimming
a
manufacturer’s data the excess cured adhesive by milling. As a result of this set of
b
Estimated from the Hooke’s law using E and  operations, it was possible to obtain a good representation of
c
estimated in Campilho et al.[23] the theoretical geometry of Fig. 2, with emphasis to the
bondline end geometry positioned at x/LO=0 (x is the horizontal
The toughness properties of the adhesives were estimated coordinate initiating at the bondline end). The specimens were
with the Double-Cantilever Beam (DCB) test (tensile fracture tested as depicted in Fig. 2, i.e. by clamping the edges of the
 F.J.P. Moreira et al. / Procedia Manufacturing 51 (2020) 848–855 851
4 Moreira et al. / Procedia Manufacturing 00 (2019) 000–000

straight adherend and pulling in peel while transversely analysis were comprised of 0.02×0.02 mm elements at the
restraining the upper joint edge. This was done in a Shimadzu overlap edges. The boundary conditions consisted of fixing the
AG-X 100 testing machine, equipped with a 100 kN load cell, base edges to simulate gripping in the testing machine,
at an approximate temperature of 20ºC and testing speed of 1 applying symmetry at the middle of the specimen and pulling
mm/min. Five specimens were fabricated and tested for each the curved adherends’ edges in peel.
joint type. A minimum of 4 valid tests was always assured for
each joint type. 3.2. XFEM formulation

3. Numerical work As an extension to the conventional FEM, the XFEM is

based on the integration of enrichment functions in the FEM
3.1. Numerical simulation details formulation [27]. These functions allow modelling the
displacement jump between crack faces that occur during the
The analysis performed in Abaqus® was 2D, considering a propagation of a crack. The Abaqus® XFEM formulation
geometrically non-linear FEM formulation. The aluminium enables the user to create a pre-crack, or it can initiate cracks
adherends (both base and L-parts) were modelled as solids with in un-cracked regions by using initiation criteria. In this last
the plastic behaviour defined in a previous work for the same scenario, considered in this work, damage initiates and
material [25]. The adhesive layer was also modelled with solid subsequently propagates during the simulation at regions
elements, but with enriched XFEM formulation. The XFEM experiencing stresses and/or strains higher than the
model is presented in the next Section. In all cases, 4-node solid corresponding limiting values. Six crack initiation criteria are
elements with plane-strain conditions were used (with Abaqus® available in Abaqus®. The MAXPS (maximum principal stress)
reference CPE4). A perfect adhesion was considered in the and MAXPE (maximum principal strain) criteria are based on
models between the adherends and adhesive, since the models the introduction of the following functions (by the respective
were built as a single part with different partitions and order)
materials. Two types of meshes were applied: a more refined
mesh to perform a stress analysis in the elastic domain, such   max    max 
that the stress plots are accurate, and a less refined mesh = to f =  or f  0  . (1)
  max    max 
0
promote the XFEM failure analysis. Fig. 3 shows an example
of mesh refinement for a T-joint with tP2=1 mm, with details at
the loaded overlap edge for the stress and failure analyses. σmax and σ0max represent the current and allowable maximum
principal stress. The Macaulay brackets indicate that a purely
compressive stress state does not induce damage. εmax and ε0max
represent the current and allowable maximum principal strain.
Crack growth for the MAXPS and MAXPE criteria is software
defined as orthogonal to the maximum principal stress/strain
direction. As a result of this, and due to the inherent mixed-
mode loading of these joints, the crack grows fast towards the
adherends. For these two criteria, the maximum load (Pm)
estimation was thus considered to take place at the time of first
cracking in the adhesive layer. The MAXS (maximum nominal
stress) and MAXE (maximum nominal strain) criteria are
represented by the following functions, respectively.

 tn ts   n s 
=f max
=  0 , 0  or f max  0 , 0  . (2)
 tn ts   n s 

tn and ts are the current normal and shear traction

Fig. 3. Mesh details at x/LO=0 for the joint with tP2=1 mm: stress analysis and components to the cracked surface. tn0 and ts0 represent the
XFEM strength prediction analyses. respective limiting values. The strain parameters have identical
significance. The quadratic nominal stress (QUADS) and
Mesh bias effects were employed in the models to grade the quadratic nominal strain (QUADE) criteria are based on the
elements’ size, enabling to reduce the total number of elements introduction of the following functions, respectively
but without compromising the accuracy. This was done by
considering a more refined mesh near the adhesive layer and 2 2 2 2
 tn   ts   n   s 
towards its edges, and a coarser mesh in the zones with less f =
 0  + 0 or f =
 0  +  0  . (3)
stress variations [26]. The elements’ side dimensions in the  tn   ts    n  s 
adhesive layer for the XFEM simulations were 0.2×0.2 mm
along the bondline, i.e., only one element through-thickness For the MAXS, MAXE, QUADS and QUADE criteria the
was considered. On the other hand, the models for the stress user can select between horizontal or vertical crack growth (in
852 F.J.P. Moreira et al. / Procedia Manufacturing 51 (2020) 848–855
Moreira et al. / Procedia Manufacturing 00 (2019) 000–000 5

this work horizontal growth, i.e., along the adhesive layers’ especially for the joints bonded with the Araldite® AV138,
length, was selected). All the six aforementioned criteria are failure sometimes took place near to one of the
fulfilled, and damage initiates, when f reaches unity. For adherend/adhesive interfaces Fig. 4 shows an example for the
damage growth, the fundamental expression of the joints with tP2=2 mm), such that visually it resembled an
displacement vector u, including the displacements adhesive failure. However, careful surface inspection including
enrichment, is written as [28]. optical microscope observations revealed that the adherends
that at first hand suffered form an adhesive failure actually were
N covered by a thin layer of adhesive. These findings are
=u  N ( x ) u
i =1
i i + H ( x ) a i  . (4) consistent with previous observations on this particular
adhesive [25]. The fracture surfaces for the joints bonded with
the Araldite® 2015 and Sikaforce® 7752 were smoother,
Ni(x) and ui relate to the conventional FEM formulation,
indicative of ductile fractures, with a clearer evidence of
corresponding to the nodal shape functions and nodal
cohesive failures. L-part adherend plasticization was detected
displacement vector linked to the continuous part of the
in all joints bonded with the Araldite® 2015 and tP2=1 mm, and
formulation, respectively. The second term between brackets,
also with the Sikaforce® 7752 and tP2=1 and 2 mm, although
H(x)ai, is only active in the nodes for which any relating shape
for this last case it was under 0.1%.
function is cut by the crack and can be expressed by the product
of the nodal enriched degree of freedom vector including the
mentioned nodes, ai, with the associated discontinuous shape
function, H(x), across the crack surfaces. The method is based
on the establishment of phantom nodes that subdivide elements
cut by a crack and simulate separation between the newly
created sub-elements. Propagation of a crack along an arbitrary
path is made possible by the use of these phantom nodes that
initially have exactly the same coordinates than the real nodes
and that are completely constrained to the real nodes up to
damage initiation. After being crossed by a crack, the element
Fig. 4. Fracture surfaces for the five sets of T-joints bonded with the
is partitioned in two sub-domains. The discontinuity in the Araldite® AV138 and with tP2=2 mm (for each set the L-parts are at the left
displacements is made possible by adding phantom nodes and the base adherend to the right).
superimposed to the original nodes. When an element cracks,
each one of the two sub-elements will be formed by real nodes 4.2. Stresses in the elastic domain
(the ones corresponding to the cracked part) and phantom
nodes (the ones that no longer belong to the respective part of This Section briefly evaluates y and xy stresses along the
the original element). These two elements that have fully adhesive during the initial stages of the elastic loading, such
independent displacement fields replace the original one. Thus, that the differences between adhesives and tP2 are properly
the crack size increment for a given crack orientation is equal accounted for. The graphics are represented as follows:
to the distance between the cracked element’s edges. From this
point, each pair of real/phantom node of the cracked element is • The x-axis represents the normalized distance along the
allowed to separate according to a suitable damage law up to adhesive length (x/LO), in which x is the horizontal
failure. At this stage, the real and phantom nodes are free to coordinate beginning at the leftmost edge of the
move unconstrained, simulating crack growth. The parameters adhesive layer, i.e. the edge closest to the central portion
introduced in Abaqus® were taken from Table 1. A linear of the joint;
softening XFEM law was considered with an energetic failure • The y-axis represents the normalized y and xy stresses
power law criterion of the type (y/avg and xy/avg, respectively). avg is the average xy
stress along the adhesive mid-thickness for the
  respective tP2.
 GI   GII 
  + 1,
 = (5)
 GIC   GIIC  Between the three adhesives, stresses in the elastic loading
stage are identical, with the sole difference residing in the
in which α is the damage law exponent (α=1 for linear normalized peak values, since these increase in proportion with
softening). the adhesive stiffness (E values compared in Table 1) [29].
Thus, graphically, only the results for one adhesive are
4. Results presented, in this case for the Aradite® 2015 (Fig. 5), which has
the middle stiffness. It can be first observed that y stresses are
4.1. Experimental failure modes prevalent over xy stresses, which was expected due to the
expected peel loading, although xy of non-negligible
All failures took place beginning with crack propagation at magnitude also develop due to the sliding of the L-part over the
x/LO=0 and growing towards the other edge. After failure, the base adherend. Independently of tP2, peak stresses or at least
fracture surfaces were inspected and cohesive failures were stress perturbations exist at the vicinity of x/LO=0 and 1 mm,
found for all adhesives and tP2. However, in some cases, which consist of the stress singularity regions.
 F.J.P. Moreira et al. / Procedia Manufacturing 51 (2020) 848–855 853
6 Moreira et al. / Procedia Manufacturing 00 (2019) 000–000

200 stresses near x/LO=1. This modification in the τxy stress plots is
related to the aforementioned behaviour of σy stresses, i.e., due
150
to the stiffening effect of the L-part, which is not accompanied
100 by the transverse deformation of the base adherend, which
y /avg

subsequently leads to large adhesive strains near x/LO=1.

50 Moreover, the variation of τxy stress plots can also influence the
failure onset and growth, identically for that discussed for σy
0
stresses.
-50
0 0.2 0.4 0.6 0.8 1 4.3. Experimental strength
x/LO
a) 1 mm 2 mm 3 mm 4 mm
The experimentally obtained Pm for the T-joints bonded
100
with the three adhesives are presented in this Section. Fig. 6
reports the average Pm vs. tP2 curves, including the standard
50
deviation of the experiments.
xy /avg

0 6

5
-50
4

Pm [kN]
3
-100
0 0.2 0.4 0.6 0.8 1
2
x/LO
b) 1 mm 2 mm 3 mm 4 mm 1

0
Fig. 5. Normalized σy (a) and τxy (b) stresses along the adhesive mid-thickness 0 1 2 3 4
for the T-joints bonded with the Araldite® 2015. tP2 [mm]
AV138 2015 7752
σy stresses were identical between the three adhesives,
although with the aforementioned differences in magnitude of Fig. 6. Experimental Pm vs. tP2 curves for the three adhesives.
peak stresses. The curves depicted in Fig. 5 (a), relating to the
Araldite® 2015, showed intermediate peak values between the It can be found that Pm always increases, and by a large
other two adhesives. For the joints with tP2=1 mm, the critical amount, with tP2, irrespectively of the adhesive. The two
region was undoubtedly x/LO=0, due to the small stiffness of Araldite® adhesives show an increasing growth with tP2, while
the L-part, arising from its low thickness. For the other tP2, σy the Sikaforce® 7752 has a marked linear evolution of the Pm
stresses are either close to zero or compressive at this location. vs. tP2 curve. The percentile Pm increase for the T-joints with
The increase of tP2 gradually reduces the harmful effect of σy tP2=2, 3 and 4 mm, over the T-joint with tP2=1 mm, was
peak stresses at x/LO=0, but gradually loads the other edge respectively of 75.0%, 173.8% and 419.3% (Araldite ®
(x/LO=1), which is deemed to occur because of the growing AV138), 81.2%, 197.8% and 403.7% (Araldite ® 2015) and
thickness and stiffness of the L-part. Actually, this change in 110.9%, 227.1% and 358.6% (Sikaforce ® 7752). This marked
deformation pattern of the L-part, namely the significant Pm improvement with tP2 is mainly due to the σy stress levelling
reduction of its bending, works against the natural curvature effect that is visible in Fig. 5 (a) near x/LO=0, which
developing in the straight adherend due to the tensile pulling. corresponds to the stress initiation site. Between adhesives, the
This gives rise to tensile σy peak stresses near x/LO=1, while the Sikaforce® 7752 clearly outperforms the other two adhesives,
other edge tends to become lightly loaded. This effect can even despite being the less strong amongst the three adhesives
be responsible by an alteration of the failure path if tP2 reaches (Table 1). The Pm improvement of the T-joints bonded with this
a given limit. On the other hand, σy stresses tend to span for a adhesive, over those bonded with the Araldite® AV138, ranges
bigger region of the bonding length with the increase of tP2, between 213.8% (tP2=4 mm) and 328.0% (tP2=2 mm). On the
which can be responsible for a Pm improvement. other hand, compared against the joints bonded with the
Identically to that found for σy stresses, the behaviour of τxy Araldite® 2015, the Pm improvement varies between 78.5%
stresses has large similarities between adhesives, although with (tP2=4 mm) and 128.1% (tP2=2 mm). It is also visible in Fig. 6
peak magnitude variation between adhesives due to their that Pm for the joints bonded with the Araldite® 2015 are higher
stiffness differences. The T-joints bonded with the Araldite® than those with the Araldite® AV138. In fact, depending on tP2,
2015, whose curves are shown in Fig. 5 (b), presented a middle Pm may almost double that obtained with the Araldite® AV138
behaviour between the three tested adhesives. The smallest τxy adhesive, although this adhesive has higher strengths. The
peak stresses are those of the joints with tP2=1 mm and, for this improvement is minimum for tP2=4 mm (78.4%) and maximum
joint, τxy stresses peak close to x/LO=0, while being negligible for tP2=3 mm (97.2%). These results show that, in a test
for the remainder of the bonding portion. On the other hand, geometry that is mainly loaded in peel, and which promotes
the increase of tP2 lightly increases τxy peak stresses near stresses to be concentrated in small areas, flexible and ductile
x/LO=0, concurrently with a progressive increase of τxy peak adhesives behave best. In one hand, the flexibility tends to
854 F.J.P. Moreira et al. / Procedia Manufacturing 51 (2020) 848–855
Moreira et al. / Procedia Manufacturing 00 (2019) 000–000 7

increase the area along which stresses are being transferred for the QUADE (tP2=2 mm), +214.7% for the MAXE (tP2=4
(Fig. 5). On the other hand, the ductility permits plasticization mm) and +160.5% for the MAXPE (tP2=2 mm).
of the adhesive at the stress concentration sites while the 3

regions in the vicinity becomes loaded, resulting in an overall

improved behaviour.
2

Pm [kN]
4.4. Numerical evaluation of the XFEM initiation criterion
1
The XFEM initiation criteria described in Section “XFEM
formulation” are evaluated against the experimental data, by
directly comparing Pm with the experiments. At this stage, the 0
linear energetic criterion, considering α=1 in equation 0 1 2 3 4
tP2 [mm]
Erro! A origem da referência não foi encontrada., is
considered in all simulations. As it was previously discussed, a) Exp MAXS QUADS QUADE MAXE MAXPS MAXPE

6
the use of the MAXS, MAXE, QUADS and QUADE criteria
results in crack onset and growth parallel to the adhesive layer,
while MAXPS and MAXPE criteria leads to cracking 4
perpendicular to the maximum principal stresses or strains,

Pm [kN]
which subsequently makes the crack grow in the direction of
the adherends. Fig. 7 shows the Pm comparison for all tP2 2
between the different XFEM initiation criteria and the
experiments for the adhesives Araldite® AV138 (a), Araldite®
2015 (b) and Sikaforce® 7752 (c). 0
0 1 2 3 4
For the Araldite® AV138, the QUADS and MAXS criteria tP2 [mm]
are closest to the experimental points, and the respective curves b) Exp MAXS QUADS QUADE MAXE MAXPS MAXPE

are practically overlapped in the figure. Averaged over the 20

experiments, the maximum relative deviations were +8.8%
(tP2=2 mm) and +9.5% (tP2=2 mm), respectively, for these two 16
criteria. The MAXPS criterion revealed to be unsuited, in the
12
manner that it was used, since Pm highly underestimates the
Pm [kN]

tests (up to -82.5% for tP2=1 mm). Oppositely to this behaviour, 8

the strain-based criteria (QUADE, MAXE and MAXPE)
overshoot the experimental data, with emphasis to the QUADE 4
and MAXE. The highest offsets for these criteria were
0
+113.3% (QUADE), +117.3% (MAXE) and +31.3% 0 1 2 3 4
(MAXPE), in all cases for tP2=2 mm. tP2 [mm]
Qualitatively, the Pm predictions for the Araldite® 2015 by c) Exp MAXS QUADS QUADE MAXE MAXPS MAXPE

the six criteria agree with those of the Araldite® AV138. Thus,
the QUADS and MAXS criteria are quite close to the Fig. 7. Experimental and numerical Pm comparison, considering different
XFEM initiation criteria, for the T-joints bonded with the Araldite® AV138
experimental values, with a negligible difference between (a), Araldite® 2015 (b) and Sikaforce® 7752 (c).
them. The maximum Pm deviations were, in both cases,
obtained for tP2=4 mm, attaining -8.9% (QUADS) and -8.6% It was shown in a previous work [23] that damage initiation
(MAXS). The MAXPS criterion was offset up to -81.3% (tP2=1 is ruled by the adhesive layer’s stresses rather than the strains
mm), whilst the maximum deviations for the strain-based (which also vary by a large amount between adhesives). On the
criteria attained maximums of +110.7% for both QUADE and other hand, using strain-based criteria can result in major
MAXE (tP2=1 mm), and +55.2% the MAXPE criterion (tP2=2 deviations to the real joint behaviour, with an over prediction
mm). tendency. This is why the QUADS and MAXS criteria
Finally, the results for the Sikaforce® 7752 were much alike generally work very well. The QUADE and MAXE criteria,
those of the former adhesives although, in this case, even the being based on strains, naturally present wrong Pm results and
QUADS and MAXS criteria showed bigger variations to the should not be considered in the design process of bonded joints.
experimental Pm (up to -13.3% for the QUADS, considering The MAXPS and MAXPE criteria, due to its intrinsic
tP2=4 mm, and +13.3% for the MAXS, considering tP2=1 mm). formulation, are unable to promote a realistic damage growth
Identically, the curves for these two criteria overlap. Due to the path, since the crack growth direction is ruled by the maximum
aforementioned approximations, the MAXPS criterion showed principal stresses or strains, in a sense that crack initiates and
Pm values much below the expected, with a maximum grows perpendicularly to the principal directions. As a result,
deviation of -85.8% (tP2=3 mm). The strain-based criteria adherend detaching by the adhesive through all the adhesive
significantly over predicted Pm, in line with the previous layer is rendered unfeasible because the mixed-mode loading
adhesives. The maximum offsets, all by excess, were +197.8% induced in the adhesive results in short crack growth in the
adhesive before the crack hits the adherend interface. Since Pm
 F.J.P. Moreira et al. / Procedia Manufacturing 51 (2020) 848–855 855
8 Moreira et al. / Procedia Manufacturing 00 (2019) 000–000

was assessed by the damage initiation load, i.e., at the time the [8] K. Duan, X. Hu, Y.-W. Mai, Substrate constraint and adhesive
thickness effects on fracture toughness of adhesive joints, Journal of
first crack appears in the model, the results do not match the Adhesion Science and Technology 2004;18:39-53.
real joint behaviour. Moreover, since stresses and strains in [9] S.K. Panigrahi, B. Pradhan, Three Dimensional Failure Analysis and
FEM modelling are traditionally mesh dependent [11], the Damage Propagation Behavior of Adhesively Bonded Single Lap Joints in
Laminated FRP Composites, Journal of Reinforced Plastics and
direct use of initiation criteria to assess failure, i.e., without
Composites 2007;26:183-201.
failure criteria, should promote mesh-dependent Pm. [10] P.a.W.B. Weißgraeber. Crack Initiation at Weak Stress Singularities
– Finite Fracture Mechanics Approach: Procedia Materials Science; 2014.
5. Conclusions [11] L.F.M. da Silva, R.D.S.G. Campilho. Advances in numerical
modelling of adhesive joints. Heidelberg, Germany: Springer; 2012.
[12] Q.D. Yang, M.D. Thouless, Mixed-mode fracture analyses of
This work presented an experimental and numerical plastically-deforming adhesive joints, International Journal of Fracture
assessment of the behaviour of adhesively-bonded T-joints 2001;110:175-87.
[13] H. Khoramishad, A.D. Crocombe, K.B. Katnam, I.A. Ashcroft,
between aluminium adherends, considering different geometric Predicting fatigue damage in adhesively bonded joints using a cohesive
conditions (tP2) and adhesives with different characteristics zone model, International Journal of Fatigue 2010;32:1146-58.
with respect to the strength and ductility. The experimental [14] L. Daudeville, P. Ladevèze, A damage mechanics tool for laminate
delamination, Composite Structures 1993;25:547-55.
analysis showed that, for the particular joints conditions tested, [15] G.Z. Voyiadjis, P.I. Kattan. Damage Mechanics. New York: Marcell
i.e., a predominantly peel loading with major peak stresses, the Dekker; 2005.
most ductile although less strong Sikaforce ® 7752 is the one [16] P. Raghavan, S. Ghosh, A continuum damage mechanics model for
that presents better results for all tP2. Increasing tP2 highly unidirectional composites undergoing interfacial debonding, Mechanics of
Materials 2005;37:955-79.
increased Pm for all adhesives. The difference between [17] M. Imanaka, T. Hamano, A. Morimoto, R. Ashino, M. Kimoto,
adhesives was clarified by a σy and τxy stress analysis being Fatigue damage evaluation of adhesively bonded butt joints with a rubber-
performed to the adhesive layer, which showed that the σy peak modified epoxy adhesive, Journal of Adhesion Science and Technology
2003;17:981-94.
stresses ruling the failure process are inversely proportional to [18] M.M.A. Wahab, I.A. Ashcroft, A.D. Crocombe, S.J. Shaw, Prediction
the adhesives’ stiffness. Thus, the corresponding stress plots of fatigue thresholds in adhesively bonded joints using damage mechanics
are more uniform and enable spreading the loads more evenly, and fracture mechanics, Journal of Adhesion Science and Technology
2001;15:763-81.
which highly benefits Pm. Adding to this, the Sikaforce® 7752 [19] S. Mohammadi. Extended finite element method for fracture analysis
is highly ductile, thus permitting the adhesive to undergo of structures. New Jersey, USA: Blackwell Publishing; 2008.
plasticization at the highest stresses zones, whilst lightly loaded [20] N. Moës, J. Dolbow, T. Belytschko, A finite element method for crack
growth without remeshing, International Journal for Numerical Methods
regions increase load transfer, before failure. The XFEM
in Engineering 1999;46:131-50.
analysis applied to the initiation criterion enabled to conclude [21] A. Mubashar, I.A. Ashcroft, A.D. Crocombe, Modelling damage and
that, for all the adhesives, the QUADS and MAXS criteria were failure in adhesive joints using a combined XFEM-cohesive element
the most adequate. The MAXPS criterion was inadequate, in methodology, The Journal of Adhesion 2014;90:682-97.
[22] F. Stuparu, D.M. Constantinescu, D.A. Apostol, M. Sandu, A
view of the simplification taken to estimate Pm. All strain-based combined cohesive elements - XFEM approach for analyzing crack
criteria (QUADE, MAXE and MAXPE) overshot Pm by a large propagation in bonded joints, The Journal of Adhesion 2016;92:535-52.
amount for the three adhesives, and should not be considered [23] R.D.S.G. Campilho, M.D. Banea, A.M.G. Pinto, L.F.M. da Silva,
A.M.P. de Jesus, Strength prediction of single- and double-lap joints by
as well. standard and extended finite element modelling, International Journal of
Adhesion and Adhesives 2011;31:363-72.
References [24] T.M.S. Faneco. Caracterização das propriedades mecânicas de um
adesivo estrutural de alta ductilidade [Tese de Mestrado]: Instituto
Superior de Engenharia do Porto; 2014.
[1] E.M. Petrie. Handbook of Adhesives and Sealants. New York: [25] S.L.S. Nunes, R.D.S.G. Campilho, F.J.G. da Silva, C.C.R.G. de
McGraw-Hill; 2000. Sousa, T.A.B. Fernandes, M.D. Banea, et al., Comparative failure
[2] S. Akpinar, The strength of the adhesively bonded step-lap joints for assessment of single and double-lap joints with varying adhesive systems,
different step numbers, Composites Part B: Engineering 2014;67:170-8. The Journal of Adhesion 2016;92:610-34.
[3] G. Di Bella, C. Borsellino, E. Pollicino, V.F. Ruisi, Experimental and [26] T.A.B. Fernandes, R.D.S.G. Campilho, M.D. Banea, L.F.M. da Silva,
numerical study of composite T-joints for marine application, International Adhesive selection for single lap bonded joints: Experimentation and
Journal of Adhesion and Adhesives 2010;30:347-58. advanced techniques for strength prediction, The Journal of Adhesion
[4] R.S. Trask, S.R. Hallett, F.M.M. Helenon, M.R. Wisnom, Influence of 2015;91:841-62.
process induced defects on the failure of composite T-joint specimens, [27] M.G. Pike, C. Oskay, XFEM modeling of short microfiber reinforced
Composites Part A: Applied Science and Manufacturing 2012;43:748-57. composites with cohesive interfaces, Finite Elements in Analysis and
[5] F. Bianchi, T.M. Koh, X. Zhang, I.K. Partridge, A.P. Mouritz, Finite Design 2015;106:16-31.
element modelling of z-pinned composite T-joints, Composites Science [28] Abaqus®. Documentation of the software Abaqus®. Dassault
and Technology 2012;73:48-56. Systèmes. Vélizy-Villacoublay 2013.
[6] L.A. Burns, A.P. Mouritz, D. Pook, S. Feih, Bio-inspired design of [29] C.C.R.G. de Sousa, R.D.S.G. Campilho, E.A.S. Marques, M. Costa,
aerospace composite joints for improved damage tolerance, Composite L.F.M. da Silva, Overview of different strength prediction techniques for
Structures 2012;94:995-1004. single-lap bonded joints, Journal of Materials: Design and Application -
[7] T. Yang, J. Zhang, A.P. Mouritz, C.H. Wang, Healing of carbon fibre– Part L 2017;231:210-23.
epoxy composite T-joints using mendable polymer fibre stitching,
Composites Part B: Engineering 2013;45:1499-507.