Aerospace Science and Technology 23 (2012) 224–232

Contents lists available at ScienceDirect

Aerospace Science and Technology

www.elsevier.com/locate/aescte

Advanced modelling of bird strike on high lift devices using hybrid

Eulerian–Lagrangian formulation
I. Smojver ∗ , D. Ivančević
Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb, I. Lučića 5, HR-10000 Zagreb, Croatia

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

Article history: The work presented in this paper deals with application of explicit finite element analyses in order to
Received 1 March 2011 predict bird strike induced impact damage on high lift devices of typical large transport aircraft. The
Received in revised form 21 July 2011 hybrid Eulerian–Lagrangian finite element formulation has been applied as to efficiently model the impact
Accepted 24 July 2011
of a highly deformable object on an inboard flap structural model. As structural elements of aeronautical
Available online 5 August 2011
structures are commonly assembled from structural items manufactured from a variety of materials, a
Keywords: damage prediction methodology on aeronautical structures has to be able to model damage in metallic
Aeronautical structure materials as well as in composite and sandwich structures. The presented methodology employs a
High lift devices progressive damage model for composite materials which is based on Hashin’s failure initiation criterion.
Damage modelling Prediction of damage in metallic structural items is based on the equivalent plastic strain, while strain
Composite structures rate effects are included in the hardening behaviour. An important part of the bird strike modelling
Bird strike procedure is the application of a bird replacement material with constitutive response which replicates
Structural impact
the forces induced by an impact of a real bird. The bird material is modelled by an equation of state, with
properties that match the pressure–density relations of water and air mixtures. The bird replacement
material model and Eulerian impactor model have been validated in a comparison with experimental
results available in the references. In order to demonstrate the ability of the presented procedure to
simulate bird strike induced damage on actual high lift devices, an example of a possible impact scenario
is investigated. The input parameters for the simulation have been selected as to simulate a bird strike
used to verify the compliance with damage tolerance certification requirements.
© 2011 Elsevier Masson SAS. All rights reserved.

1. Introduction experiments. Abaqus/Explicit has been used in this work to de-

velop an accurate damage prediction procedure.
A very common example of foreign object impact in air trans- In order to efficiently employ numerical methods in these sim-
port is bird strike. Due to increased density of the air transport ulations, a set of numerical difficulties have to be addressed. The
and changing migration routes of flocking birds [6], this impact greatest difficulty that arises is modelling of the extreme bird ma-
loading still presents a significant threat to the safety of airplanes, terial deformation which occurs during the impact. Excessive de-
and needs to be accounted for in the certification phase of certain formation and even bird material separation have to be correctly
aircraft components. Accurate simulation of the bird strike is still replicated in order to accurately transfer the impacting loads on
a challenge when applying numerical methods due to the complex the structure. The application of the traditional Lagrangian formu-
physical phenomena which need to be correctly simulated in order lation for bird deformation modelling is not suitable, since dis-
to predict the response of the impacted structure. tortion of bird finite elements leads to a drop of the stable time
Nonlinear explicit finite element analyses enable prediction of increment of the explicit integration algorithm and, eventually, to
damage caused by the foreign object impact without the need for numerical errors caused by excessively distorted finite elements.
costly and time consuming experiments. This ability is particularly Despite these shortcomings, this formulation has been used for
useful in the certification phase of the design process, in which
bird strike simulation in [9,2,13,19] as it is the computationally
the compliance with certification requirements has to be demon-
more efficient technique compared to the more advanced formu-
strated. Numerical methods and techniques are therefore still being
lations like hybrid approaches and SPH (Smooth Particle Hydrody-
improved in order to enhance the accuracy of bird impact sim-
namics).
ulations and, consequently, reduce the requirements for gas-gun
Hybrid approaches employ Eulerian finite element discretiza-
tion of the impactor, with traditional Lagrangian description in
* Corresponding author. Tel./fax: +385 1 6168 267. modelling of the impacted structure. Thus, the bird material
E-mail address: ismojver@fsb.hr (I. Smojver). flows relative to an Eulerian mesh, thereby avoiding large mesh

1270-9638/$ – see front matter © 2011 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ast.2011.07.010
 I. Smojver, D. Ivančević / Aerospace Science and Technology 23 (2012) 224–232 225

distortions. The impacting loads are transferred to the Lagrangian recomputed in each time increment, as the material flows through
mesh of the impacted structure through an Eulerian–Lagrangian the mesh.
coupling algorithm [24]. This bird strike modelling approach has The application of the CEL method in bird strike simulations
been used in e.g. [22,15,20]. allows the bird to be modelled as Eulerian material, while the
The SPH, a more recent approach to the bird strike modelling impacted structure is represented by traditional Lagrangian finite
problem, is a mesh free method based on the Lagrangian formu- elements. Utilization of this technique avoids numerical difficulties
lation in which the finite elements have been replaced by a set associated with Lagrangian bird models as there are no restric-
of discrete, mutually interacting particles. As SPH is a mesh-free tions on the Eulerian material deformation. The initial position of
method, it is well suited for simulations in which large material the bird material is determined by the Abaqus/CAE volume frac-
deformations are expected to occur. This method has been used tion tool. In Abaqus/CAE, the geometry of the Eulerian material
for bird strike simulations in [8,7,12]. Hybrid approaches and SPH has to be introduced as a reference part, whose only purpose is to
formulation significantly improve bird strike analysis stability, but define the initial material boundaries by calculating the initial vol-
a major disadvantage of these approaches is the prolonged com- ume fractions for all Eulerian finite elements. The volume fraction
putational time needed to simulate a couple milliseconds of a real tool also creates a node set containing nodes in the area of the
time in which the impact takes place. material geometry, allowing assignment of initial conditions to the
The work presented in this paper is a continuation of the work Eulerian material.
presented in [19] and [20]. As already demonstrated in [20], ap- The interaction between the Eulerian material and Lagrangian
plication of the hybrid approach resulted in physically improved structure is achieved by an extension of the general contact algo-
modelling of the fluid-like bird deformation at the high velocity rithm. The contact is created between Lagrangian mesh surfaces
impact. The Abaqus hybrid Coupled Eulerian–Lagrangian formula- and Eulerian material surfaces, which are automatically computed
tion thus has been utilized in this work as well. and tracked during the analysis. No restrictions are placed on the
The main focus of the current paper is on the improvement of mutual position of the nodes of different finite element meshes,
the bird strike loading imposed on the impacted structure. Numer- as there is no interaction between the nodes of the Eulerian and
ical bird strike approaches treat the bird as a mixture of water and Lagrangian finite element meshes. Abaqus uses penalty contact al-
air with constitutive behaviour governed by an Equation Of State gorithms to introduce coupling between Eulerian and Lagrangian
(EOS). However, the Mie–Grüneisen EOS, which has been used in meshes, as this approach uses the simplest computational level
[19] and [20], cannot include porosity effects of water to air mix- and increases robustness, as described in [3]. There are two impor-
tures which are necessary for realistic bird material simulation. tant restrictions on the dimensions and mesh size of the Eulerian
Therefore, the polynomial equation of state has been included via model in CEL analyses. The size of the volume enclosing Eulerian
the Abaqus user subroutine VUMAT. The coefficients of the poly- elements must be sufficiently large to prevent loss of material dur-
nomial EOS have been varied as to replicate the time dependent ing the analysis. The loss of material leads to a loss of kinetic
contact pressures measured at gas-gun experiments in which real energy and could under some conditions lead to numerical insta-
birds have been fired against rigid targets, thereby validating the bilities. The second restriction is placed on the mesh size of the
improved material models. Eulerian finite element model. A very fine mesh of the Eulerian
Finally, bird strike damage prediction procedures have to be grid is necessary to efficiently capture the contact between Eule-
able to simulate damage on complex aeronautical structures. Such rian material surfaces and Lagrangian elements in order to prevent
structures are mostly manufactured from a variety of materials, in- physically unacceptable penetration of the bird material through
cluding aluminium alloys, composite materials and sandwich struc- the target finite element mesh. The CEL bird strike modelling tech-
tures. A procedure that accurately simulates bird strike damage has nique has already been employed in [20], resulting in improved
to be able to take into account damage modelling on all these ma- physical modelling of the fluid-like bird deformation compared to
terials. the Lagrangian model used in [19].
The damage prediction procedure described in this work is pri-
marily focused on high lift devices. The paper presents results of 3. Bird material validation
bird impact simulation on a detailed finite element model of a typ-
ical large airliner flap model. The problem of finding a material that successfully simulates
real birds during bird strike research arises in both numerical and
2. Numerical bird strike modelling technique experimental investigation methods. Gas-gun experiments with
real birds, although providing the best method to realistically sim-
This work employs the Abaqus hybrid modelling technique – ulate an actual bird impact, lack the ability to provide repeatable
Coupled Eulerian–Lagrangian (CEL) approach. This formulation of- test results as different birds have different anatomic structures
fers the ability to simulate fluid–structure interaction in which and thus the effect on the impacted target differs in every single
the structure is formulated by traditional Lagrangian formulation, test. To overcome this shortcoming, the authors of [23] analyzed
whereas the fluid is modelled as the Eulerian material. The Eule- different homogeneous substitute materials which could be used
rian finite element model in CEL analyses is represented by a sta- in gas gun experiments in order to be able to validate experimen-
tionary cube containing multi-material EC3D8R volume elements, tal results. A further disadvantage of experimental test methods is
which may be completely or partially occupied by material [1]. the high cost involved in gas gun experiments on actual aeronau-
The Eulerian material is able to move through the stationary mesh tical components. The results published in [23] are an important
and interact with the Lagrangian structure. The ratio by which the reference for scientific research in the field of soft body impact,
EC3D8R elements are occupied with bird material is the Eulerian as it provides important test results as well as equations which
Volume Fraction (EVF). Completely filled elements have the EVF theoretically explain the load distribution and pressure time de-
equal to one, whereas completely void materials have EVF values pendency involved in a bird impact.
equal to zero. As the material flows through the stationary mesh, As the bird strike usually occurs at higher velocities, very high
elements obtain diverse values of the EVF, which enables track- stresses are generated in the bird material. These stresses greatly
ing of the material motion. The Eulerian material boundaries are exceed the material’s strength, leading to a fluid-like deformation
independent of the finite element mesh and don’t have to match of the bird material. This observation allows hydrodynamic theory
element geometry at any time during the analysis, but have to be to be applied in the theoretical analysis of the bird strike [23]. As
226 I. Smojver, D. Ivančević / Aerospace Science and Technology 23 (2012) 224–232

Fig. 1. Hugoniot curves of porous water and approximated homogenized bird materials.

the strength and viscosity in this approach are neglected, a simple The work in this paper presents the continuation of the work
pressure vs. density equation of state can be used to describe the presented in [19] and [20] with significantly enhanced bird mod-
constitutive behaviour of the bird material in numerical analyses. elling through the improvement of the bird material constitutive
The conclusion of Willbeck’s experimental results in [23] is that behaviour by means of application and verification of different
the most suitable bird substitute material was gelatine with 10% EOS properties explained in detail in this paragraph. A generally
porosity and a density of 950 kg/m3 . Another important conclusion accepted benchmark procedure used to validate numerical bird re-
was that the most appropriate shape of the impacting body is a placement materials [22,15,12] has been followed in this paper
cylinder with hemispherical ends and a length to diameter ratio as well. This validation method consists of a comparison of the
equal to 2. contact pressures, obtained by numerical simulation of the bird
Numerical bird replacement materials usually have the proper- impacts against a rigid plate, with experimental results published
ties of water, as real birds mostly consist of water and trapped air in [23]. A similar validation of the bird material has been already
inside internal cavities. The constitutive behaviour of the fluid-like published in [19]. These results have been obtained with the La-
materials is governed by hydrodynamic material models, whose grangian bird model and the Abaqus built-in Mie–Grüneisen equa-
volumetric strength is calculated using equations of state. An in- tion of state. The properties of the equation of state were taken
complete equation of state, which doesn’t cover heat conduction from [5] to match the behaviour of water during high velocity im-
effects, determines values of the hydrostatic pressure depending pact. The results showed good correlation with theoretical values
on a combination of two internal variables: density ρ and specific of Hugoniot and stagnation pressures, but the values of Hugoniot
internal energy e, volume V and temperature T , or volume V and pressure differ greatly from the values obtained in Willbeck’s re-
specific internal energy e after [11] sults.
In order to improve the time dependent pressure response at
the impact, in this paper porosity effects of water to air mixtures
p = p (ρ , e ) ∨ p ( V , T ) ∨ p ( V , e ). (1) have been taken into account by programming the polynomial
equation of state using Abaqus/Explicit user material subroutine
As crash-worthiness problems are mostly not affected by heat con-
VUMAT. According to [23], porosity has a significant effect on the
duction effects, incomplete equations of state are accurate enough
shock velocity and compressibility of soft body impactors and, con-
to be used for bird strike problems. Willbeck’s results show that
sequently, needs to be considered in order to realistically replicate
the time dependence of the pressure values at the impacted plate
the forces generated at an impact of real bird. Porosity decreases
shows a recognizable pattern with three distinct regions. Imme-
the shock velocity in the material, resulting in the lower Hugoniot
diately after the initial contact very high values of pressure arise.
and stagnation pressures although the effect on stagnation pres-
The peak pressure value (also called Hugoniot pressure) has the
sures is not as pronounced. Substitute materials applied in the
theoretical value [23]
numerical analyses usually have porosities (α ) of 10–15%, as for
example in [2,15,12]. According to Willbeck’s theory of mixtures
p H = ρ0 U S (U 0 )U 0 . (2) the effect of porosity on the pressure to density relation is ac-
counted for by the relation [23]
The second stage is characterized by release waves which decrease
 −1/ B  −1/γ
pressure values [23]. After several reflections of release waves, a ρ0 p p
region of stable and constant pressure is established. The steady = (1 − α ) +1 +α , (4)
ρ A p0
flow stage is characterized by the stagnation pressure
where ρ0 and p 0 are the initial density and pressure, respectively,
1 while γ is the ratio of specific heats of the air. The empirical con-
p stag = ρ0 U 02 . (3) stants A and B are defined by Eqs. (5) and (6), where s is the
2
 I. Smojver, D. Ivančević / Aerospace Science and Technology 23 (2012) 224–232 227

niot curves of the homogenized materials with 10 and 15% poros-

ity. Hugoniot curves describe the unique pressure to density rela-
tionship achievable behind the shock wave. The Hugoniot curves of
the approximated material properties along with the ones for wa-
ter with varying porosity after Eq. (4) have been plotted in Fig. 1.
The CEL model used to validate the EOS properties has been
shown in Fig. 2. In order to efficiently capture the Hugoniot pres-
sure, the results are written to the output file every 2.5 μs. The
Eulerian model consists of 500 000 ECD8R elements, enclosing a
volume with dimensions 1.4 × 1.4 × 0.7 m, while the element size
has been biased by a factor of 5 towards the centre of the Eulerian
mesh as to create a finer mesh at the impact location. In order to
compare the results with the experimental values, the initial ve-
locity of the bird material has been set to 116 m/s.
Fig. 2. CEL model for bird material validation. Fig. 3 summarizes the results of the validation simulation for
different EOS properties as to illustrate enhancements of the new
coefficient which relates impacting and shock velocities, while c 0 bird models compared to the EOS used in the previous work
is the speed of sound in the material [19,20]. Beside the properties of the polynomial EOS with coeffi-
cients which match Hugoniot curves of porous water, Eq. (7), two
ρ0 c02 additional EOS properties have been tested in this validation pro-
A= , (5)
(4s − 1) cedure. The first EOS employs only the coefficient C 1 in Eq. (7)
B = 4s − 1. (6) to simulate a nonporous polynomial EOS which has been used for
bird replacement materials in [16] and [10]. The Hugoniot curve
The effect of porosity on the pressure vs. relative specific volume of this EOS has been also plotted in Fig. 1. The second EOS is the
relation, after Eq. (4), is shown by Fig. 1. The polynomial equation Abaqus built-in Mie–Grüneisen EOS which has been used as bird
of state which has been used for bird material modelling in [2,22, substitute material in [19] and [20]. The Mie–Grüneisen EOS has
12,10] has the form the form
 
p = C 0 + C 1 μ + C 2 μ2 + C 3 μ3 , (7) ρ0 c02 η Γ0 η
p= 1− + Γ0 ρ0 e , (9)
where μ is a dimensionless parameter which is defined in terms (1 − sη)2 2
of the ratio of initial to current density with material properties c 0 = 1480 m/s, Grüneisen’s Gamma
ρ Γ0 = 0 and s = 0 as defined in [5].
μ= − 1. (8) The pressure values in Fig. 3 have been normalized by the stag-
ρ0 nation pressure in order to compare the values with experimental
As in [2], the coefficients C 0 –C 3 in the polynomial EOS (Eq. (7)) results. It can be concluded that the pressure temporal responses
have been varied in order to achieve a reasonable fit of the Hugo- for all validated EOS properties follow the general trend observed

Fig. 3. Pressure–time history plots for different EOS properties.

228 I. Smojver, D. Ivančević / Aerospace Science and Technology 23 (2012) 224–232

items made of CFRP have been modelled as unidirectional lam-

inates. The upper and lower skins are additionally strengthened
with 10 CFRP stringers, of which 6 are positioned along the up-
per skin and 4 are positioned on the lower flap skin. The distance
between front and rear spar is 525 mm, resulting in intervals be-
tween stringers of 105 mm for the lower and 75 mm for the upper
flap skin. Flap leading and trailing edges have been designed as
sandwich structures consisting of CFRP face layers for the lead-
ing edge and Al2024 alloy for the trailing edge face layers. Nomex
honeycomb cores have been used in both sandwich structures.
Due to thickness variations of different flap structural items,
the finite element model has been assembled by a combination
of beam, conventional and continuum shell elements and three-
dimensional solid elements. The total number of elements in the
model is 109 585, of which 69 696 are conventional shell, 16 020
Fig. 4. Eulerian bird material deformation upon impact.
continuum shell, and 23 665 solid elements. Local reinforcements
of the front spar are modelled with 204 beam elements. Most of
in experimental results with distinct Hugoniot, pressure release the flap geometry is suitable for two-dimensional discretization,
and stagnation pressure stages. The theoretical Hugoniot pressure whereas three-dimensional finite elements have been employed to
value for an impact at 116 m/s is 93.6 MPa, after [14], which has model sandwich structures. To efficiently connect finite elements
given normalized value of 14.9. Experimental results for real birds having different nodal degrees of freedom, kinematic constraints
show significantly lower peak pressure values, reaching normalized had to be introduced at the interface of different elements. These
pressure value of 3.5. Fig. 3 shows that all tested EOS properties kinematic couplings have been achieved by the tie surface con-
show significantly higher Hugoniot pressure values than the ex- straints which are imposed to the nodes at the juncture of con-
perimental values reported in [23]. The same conclusion has been tinuum shell elements of the sandwich structure and conventional
reported in [22,12]. The discrepancy between numerical and ex- shell elements of the main flap structure.
perimentally obtained results could be explained by the inability Compared to [19] and [20], detailed description of structural el-
of the measuring equipment which has been used in the experi- ements in the vicinity of impact point has been illustrated, what
ments to capture the instantaneous peak pressure [14]. The Mie– has not been done in the previous work due to the material and
Grüneisen EOS of the nonporous material predicts the theoretical structural complexity of this aeronautical component. Fig. 5 shows
pressure value very accurately (also reported in [19]), but overes- a detail of the flap structure finite element model in the area of
timates the experimental values, with normalized peak pressure the analyzed impact, with assigned section properties, while the
values reaching 13.81. The inclusion of the porosity effects lowers geometry and structural layout of the complete flap model are
the peak pressures to 9.1 for the polynomial EOS with 10% porosity shown by Fig. 6. The gradually varying lower flap composite skin in
and 9.9 for EOS properties with 15% porosity. The EOS suggested the impact area consists of two different composite layups – a 14-
in [16] and [10] reaches normalized peak pressure values of 12.2. layer layup (t = 1.75 mm) and an 18-layer layup (t = 2.25 mm).
The stagnation pressure values for all validated EOS properties The 14-layer layup is applied also to the complete upper skin at
oscillate around the theoretical and experimental value (normal- the location of impact. The skin stringers consist of 21 CFRP layers
ized to 1.00). The Mie–Grüneisen EOS produces pressure oscilla- with symmetrical layup resulting in a thickness of 2.625 mm. The
tions with relatively high frequency and amplitude. The best match stacking sequences of composite parts along with the thicknesses
of stagnation pressure has been achieved by the polynomial EOS of metallic structural items are shown by Fig. 5. The 0◦ oriented
with 10% porosity as it produces the least amplitude of pressure layer corresponds to the flap span direction and the first layer in
oscillation. Due to the fact that the loads imposed on the structure the laminate stacking sequence is located on the internal compos-
at a bird impact are the best reproduced by a polynomial EOS with ite skin.
10% porosity, it has been decided to use this EOS as bird material
for further work. The deformation of the Eulerian impactor on the 5. Material modelling
rigid plate is shown in Fig. 4. The impactor deformation closely re-
sembles gas-gun experiments published in [9,15,12] and numerical The analyzed flap model consists of structural items manufac-
results for bird replacement materials published in [15,12,16]. tured from a variety of metallic and composite materials resulting
in a need to employ various failure and damage models. The pre-
4. Flap structure model sented work simulates a bird strike in the flap at the joint of
the composite skin and metallic rib. High velocity impacts, like
The structure subjected to bird strike analyzed in this work is bird strikes, usually result in deformation rates in the intermedi-
an inboard flap of a typical large transport aircraft. The flap fi- ate strain rate regime (maximum value of equivalent strain rate for
nite element model is the same as the model analyzed in [19] the a bird strike at the analyzed velocities reaches approximately
and [20] and thus only basic pieces of information about the mod- 200 s−1 ). Strain effects for aluminium alloy have been taken into
elling approach are given in this work. Dimensions of the modelled account and implemented through the VUMAT user subroutine,
structure are span of 4.38 m and chord of 1.05 m. The flap struc- while these effects have been neglected in the previous research
tural layout consists of front, rear and auxiliary spars and 14 ribs. [19,20].
These structural items have been manufactured out of aluminium The strain rate dependency has been included for the Al 2024-
alloys. Front and rear spar, as well as ribs which comprise struc- T3 alloy which is used for the impacted rib, while strain rate
tural items used to connect the flap to the wing structure and effects are neglected in other structural items in order to reduce
hydraulic actuators, are made of Al7075 aluminium alloy, while computational time. Generally, the yield stress depends on strain,
the remaining ribs and auxiliary spar are made of Al2024 alloy. The strain rate and temperature
flap skins are made of CFRP with variable number of unidirectional
layers resulting in gradually variable skin thickness. The structural σ = f (ε, ε̇, T ). (10)
 I. Smojver, D. Ivančević / Aerospace Science and Technology 23 (2012) 224–232 229

Fig. 5. Detail of the flap model in the area of impact.

The VUMAT subroutine has been used to model stain rate effects. failure criterion is suitable for highly dynamic problems and is
The implemented constitutive model includes a Von Mises yield based on the accumulated equivalent plastic strain calculated as
criterion and an algorithm that takes into account strain rate sen-
sitivity by enforcing the Cowper–Symonds law. The elastic–plastic t
pl pl 2
behaviour has been defined as a power law [17] ε̄ = ε̄ +
0 ε̇ pl : ε̇ pl dt , (14)
3
0
σ (ε) = a + b(ε p )n . (11)
pl
where ε̄0 is the initial value of equivalent plastic strain and ε̇ pl is
The parameters a, b and n for the Al 2024-T3 alloy are taken from the equivalent plastic strain rate [1]. An element is assumed to fail
[17]. The mathematical description of the assigned hardening law when the damage parameter calculated as
has the form

 1/ p ε̄0pl +
ε̄ pl
σn ε̇ ω= , (15)
=1+ , (12) ε̄ plf
σy D
pl
where σn is the dynamic yield stress, σ y is the static yield exceeds the value of 1. In Eq. (15) ε̄ f is the strain at failure,
ε̄ pl
strength, and ε̇ is the equivalent strain rate. The parameters D is the plastic strain increment and the summation is performed
and p of the Cowper–Symonds law for the Al 2024-T3 are taken over all increments in the analysis.
from [9]. Combining Eqs. (11) and (12) results in the final form of Strain rate effects on the mechanical properties of the com-
the hardening rule posite material have been neglected, as these effects are rela-
  1/ p  tively insignificant for CFRP composites, after [4]. Modelling of
 n
 ε̇ failure and damage in the composite part of the flap structure has
σ (ε, ε̇ ) = a + b(ε p ) 1+ . (13)
D been achieved by Abaqus built-in Hashin’s failure initiation criteria
which accounts for the following failure modes: fibre rupture in
All necessary parameters needed to define the constitutive be- tension, fibre buckling and kinking in compression, matrix crack-
haviour of the Al 2024-T3 alloy are summarized in Table 1. If the ing under transverse tension and shearing, matrix crushing under
equivalent plastic strain in the material exceeds the failure strain transverse compression and shearing [1]. Mechanical properties of
rate for the Al 2024-T3 alloy, the element is assumed to fail and is unidirectional CFRP layers are taken from [21] and listed in Table 2.
therefore excluded from the analysis. Continuum damage mechanics principles are employed to
The constitutive responses of the remaining aluminium alloys model damage in the composite material. Degradation of CFRP me-
have been modelled as elastic–plastic, with mechanical properties chanical properties is accounted for by using damage parameters,
taken from [18]. Additionally, the shear failure element deletion which modify the initial undamaged elasticity matrix. Fibre (d f ),
criterion has been employed to model possible damage. The shear matrix (dm ) and shear (d s ) damage parameters reflect the current

Table 1
Properties of Al 2024-T3 alloy.

E [GPa] ν a [MPa] b [MPa] n D [s−1 ] p εfail ρ [kg/m3 ]

72.4 0.33 277 485 0.55 1.28 · 105 4.0 0.18 2780

Table 2
CFRP mechanical properties, after [21].

ρ E1 E2 G 12 ν12 S 1+ = S 1− S 2+ S 2− S 12
[kg/m3 ] [GPa] [GPa] [GPa] [MPa] [MPa] [MPa] [MPa]
1600 181 10.3 7.17 0.28 1500 40 246 68
230 I. Smojver, D. Ivančević / Aerospace Science and Technology 23 (2012) 224–232

bird at the expected operating velocities. As high lift devices are

extended only during the take-off and landing phases of the flight,
flap structures are exposed to bird strikes at relatively low veloc-
ities. Taking this into account, this work presents results of a bird
strike simulation which included a bird mass used in certification
(after FAR 25.571) at 90 m/s. Compared to [19] and [20], the mass
of the impactor has been the same, although with significantly dif-
ferent velocity (90 m/s vs. 100 m/s in [19,20]), in order to more
realistically simulate landing flight conditions. Besides, the location
Fig. 6. CEL model of the analyzed impact. of the analyzed impact has been changed compared to the previ-
ous work, as to simulate an impact on a part of the inboard flap
state of damage, having values ranging from 0 for an undamaged having different composite layup of the impacted skin. The bird
state, to 1 for a completely degraded material. The damaged elas- material used in this simulation is an EOS with properties which
ticity matrix has the form replicate the Hugoniot curve of a water and air mixture with 10%
⎡ ⎤ porosity, as described in Section 3. Taking into account the approx-
(1 − d f ) E 1 (1 − d f )(1 − dm )ν21 E 1 0
1⎢ ⎥ imate landing glide slope, wing incidence angle and aircraft pitch
Cd = ⎢ (1 − d f )(1 − dm )ν12 E 2
⎣ (1 − dm ) E 2 0 ⎥,
⎦ angle, the aircraft angle of attack has been assumed to be 15◦ .
D
0 0 (1 − ds )G 12 D The flap has been assumed to be in the maximal deflection posi-
tion (40◦ ), resulting in a total angle of attack of 55◦ with regard to
(16)
the flap reference plane. The velocity component in the flap span
where D = 1 − (1 − d f )(1 − dm )ν12 ν21 . In the previous expression, direction is zero, as the inboard flap is perpendicular to the air-
E 1 , E 2 , G 12 , ν12 and ν21 are unidirectional ply material properties plane’s plane of symmetry. The initial position of the bird material
defined in main material coordinate system. with regard to the Eulerian model and the flap structure is shown
Damage parameter evolve by a law defined as by Fig. 6.
f The Eulerian model consists of 498400 ECD8R elements. The
δeq (δeq − δeq
0
) dimensions of the cube containing these elements is 0.9 × 0.8 ×
d= , (17)
f
δeq (δeq − δeq
0
) 0.5 m, as to assure that the Eulerian material doesn’t protrude the
mesh during the analyzed 6 ms, which would lead to numerical
0
where δeq is the initial equivalent displacement at which Hashin’s difficulties as described in Section 2.
f The deformation of the 1.81 kg impactor at an initial velocity
damage initiation criterion has been reached and δeq is the dis-
placement at which the material is completely damaged according of 90 m/s is shown in Fig. 7. The analyzed impact is simulated for
to this failure mode. An element is removed from the mesh when 6 ms of real time, after which bird material has lost its kinetic en-
all material points reach the critical damage value, as to prevent ergy and no further increase of damage at the impacted structure
infinite straining of elements with degraded stiffness properties. is expected. As shown by the contours of equivalent stresses, the
flap structure is exposed to very high stresses which are spread to
6. Results a wide area of the lower composite skin.
The analyzed impact scenario causes severe damage and even
The ability to predict bird strike induced damage without costly partial penetration of the lower composite skin and stringers. Ele-
and time consuming experiments is the most valuable in the pre- ments which reach the criterion for element removal are removed
liminary phase of design of aircraft structural components. Accord- from the analysis as explained in Section 5. These elements are
ing to FAR 25.571 – “Damage-tolerance and fatigue evaluation of also removed from the contact detection algorithm, thereby creat-
structure”, the design of the flap structure must ensure that catas- ing a gap in the flap finite element mesh through which the Eu-
trophic failure does not occur at an impact with a 4 lb (1.81 kg) lerian material can penetrate, and interact with the interior struc-

Fig. 7. Bird material deformation upon impact (EVF = 0.65). Contours of Von Mises stress are shown on the impacted structure [Pa].
 I. Smojver, D. Ivančević / Aerospace Science and Technology 23 (2012) 224–232 231

Fig. 8. Penetration of the flap skin (upper skin removed, EVF = 0.5). Contours of Von Mises stress [Pa] are shown on the impacted structure at t = 5.5 ms.

Fig. 9. Contours of equivalent plastic strain (left-hand side image) and displacement magnitude (right-hand side image) [m].

tural elements. The penetration of bird material through the flap only a few elements have been removed from the analysis. These
skin is shown in Fig. 8, which depicts the contours of Von Mises elements have reached the criterion for element removal as the
stress at t = 5.5 ms. The advantages of the Eulerian impactor, com- damage parameters of all material points in the element have
pared to the impactor model used in [19], are best demonstrated reached the critical value. Matrix tensile damage is observed in the
by Figs. 7 and 8, as material separation and extreme deformation complete impacted skin and stringers, while compressive matrix
could not be modelled by the Lagrangian bird model. damage is predicted only in the stringers. The applied compos-
The impacted rib is subjected to a very large elastic and in- ite damage prediction model is not able to predict delamination
elastic deformation. The values of equivalent plastic strain reach damage. The primary concern of the work presented in this pa-
locally the failure value resulting in removal of these elements, as per is the prediction of damage of a complex structure caused
depicted in Fig. 9, left-hand side image. The highest values of plas- by a foreign body impact at relatively high velocities. The ex-
tic strain are reached at the stringer cut-outs. The right-hand side pected failure mode at such loading conditions is penetration of
image of Fig. 9 shows the contours of displacement magnitudes the structure, and for this purpose prediction of delamination is
at the end of the analysis (t = 6 ms). It can be concluded that the not relevant. Furthermore, the increased computational cost of the
rib and lower skin undergo large displacements, which cause buck- complex three-dimensional meshes, required for appropriate mod-
ling of the impacted rib. The displacement contours show the state elling of the interlaminar failure, prevents analyses on complex
of deformed structure at the end of the analysis. This state is not structures such as the one analyzed here.
the final appearance of deformed structure, as the rib still has to
undergo the elastic return. Simulation of the elastic return of the 7. Conclusions
impacted rib would require analysis of further approximately 6 ms.
As the analysis of 1 ms on the complex model requires approx- This work describes a bird strike damage prediction method-
imately 3 h of computational time on a 4 CPU desktop working ology which is based on nonlinear explicit finite element meth-
station, and taking into account that the damage on the rib is mea- ods in combination with complex structural finite element models,
sured by values of plastic strain, simulation of the elastic return various failure and degradation modes and an Eulerian impactor
has not been considered in the presented analysis. model. Realistic bird strike load prediction is a particular subject of
The extent of damage caused to the lower composite skin and the work in this paper. In order to replicate the pressure imposed
stringers is depicted in Fig. 10. The critical damage modes are ma- on the structure subjected to a real bird impact, a VUMAT subrou-
trix and fibre tensile failure. Although large portions of the skin tine has been employed to replicate the constitutive behaviour of a
and stringers have reached the critical failure initiation indexes, water and air mixture with 10% porosity. An equation of state with
232 I. Smojver, D. Ivančević / Aerospace Science and Technology 23 (2012) 224–232

Fig. 10. Contours of maximum values of Hashin’s failure indices.

approximated polynomial coefficients has been validated by com- [9] M. Guida, F. Marulo, M. Meo, M. Riccio, Analysis of bird impact on a composite
parison with experimental results published in the literature. The tailplane leading edge, Appl. Compos. Mater. 15 (2008) 241–257.
obtained results show that application of the new EOS properly [10] A.G. Hanssen, Y. Girard, L. Olovsson, T. Berstad, M. Langseth, A numerical model
for bird strike of aluminium foam-based sandwich panels, Int. J. Imp. Eng. 32
predicts the stagnation pressure. On the other hand, the peak pres-
(2006) 1127–1144.
sure value is still overestimated, although to a much lower extent [11] S. Hiermaier, Structures Under Crash and Impact, Springer Science+Business
then the Mie–Grüneisen EOS which has been used in the previous Media, New York, 2008.
work. The ability of presented methodology to simulate the bird [12] A.F. Johnson, M. Holzapfel, Modelling soft body impact on composite struc-
strike on a complex aeronautical structure is demonstrated in an tures, Compos. Struct. 63 (2003) 103–113.
impact simulation on a typical large airliner flap structure. The flap [13] E. Kirtil, D. Pestal, A. Kollofrath, N. Gänsicke, J. Mendler, Simulating the impact
behaviour of composite aircraft structures, in: ABAQUS User’s Conference, 2002.
structure is able to withstand the impact of the bird mass used for
[14] M.A. Lavoie, A. Gakwaya, M. Nejad Ensan, D.G. Zimcik, Review of existing nu-
certification requirements without complete penetration or loss of
merical methods and validation procedure available for bird strike modelling,
load carrying ability. in: International Conference on Computational & Experimental Engineering and
Sciences, 2007.
Acknowledgements [15] M.A. Lavoie, A. Gakwaya, M. Nejad Ensan, D.G. Zimcik, D. Nandlall, Bird’s sub-
stitute test results and evaluation of available numerical methods, Int. J. Imp.
This research has been financially supported by the Ministry of Eng. 36 (2009) 1276–1287.
Science, Education and Sports, Republic of Croatia through Tech- [16] M.A. McCarthy, J.R. Xiao, C.T. McCarthy, A. Kamoulakos, J. Ramos, J.P. Gallard,
V. Melito, Modelling of bird strike on an aircraft wing leading edge made from
nology Project TP-07/0120-30 “Numerical Modeling of Damage in
fibre metal laminates – Part 2: Modelling of impact with SPH bird model, Appl.
Aeronautical Structures” and the scientific project “Numerical Mod- Compos. Mater. 11 (2004) 317–340.
elling of Nonisotropic Continua”. [17] M.A. McCarthy, J.R. Xiao, N. Petrinic, A. Kamoulakos, V. Melito, Modelling of
bird strike on an aircraft wing leading edge made from fibre metal laminates
References – Part 1: Material modelling, Appl. Compos. Mater. 11 (2004) 295–315.
[18] Metallic Materials and Elements for Aerospace Vehicle Structures MIL-HDBK-5J,
[1] Abaqus Analysis User’s Manual, Version 6.8, Dassault Systémes, 2008. Department of Defence Handbook, Washington, DC, 2003.
[2] A. Airoldi, B. Cacchione, Modelling of impact forces and pressures in Lagrangian [19] I. Smojver, D. Ivancevic, Numerical simulation of bird strike damage prediction
bird strike analyses, Int. J. Imp. Eng. 32 (2006) 1651–1677. in airplane flap structure, Compos. Struct. 92 (2010) 2016–2026.
[3] D.J. Benson, S. Okazava, Contact in a multi-material Eulerian finite element for- [20] I. Smojver, D. Ivancevic, Bird strike damage analysis in aircraft structures using
mulation, Comput. Methods Appl. Mech. Eng. 193 (2004) 4277–4298. Abaqus/Explicit and coupled Eulerian–Lagrangian approach, Compos. Sci. Tech-
[4] W.J. Cantwell, J. Morton, The impact resistance of composite materials – a re- nol. 71 (2011) 489–498.
view, Composites 22 (1991) 347–362.
[21] G.S. Springer, Mechanics of Composite Structures, Cambridge University Press,
[5] M. Chizari, L.M. Barrett, S.T.S. Al-Hassani, An explicit numerical modelling of
Cambridge, 2003.
the water jet tube forming, Comp. Mater. Sci. 45 (2009) 378–384.
[22] C.H. Tho, M.R. Smith, Accurate bird strike simulation methodology for BA609
[6] P.F. Echenfelder, High speed flight at low altitude: Hazard to commercial avia-
tiltrotor, presented at the American Helicopter Society 64th Annual Forum,
tion?, in: Birdstrike Committee Proceedings, 2005.
2008.
[7] S. Georgiadis, A.J. Gunnion, R.S. Thomson, B.K. Cartwright, Bird-strike simula-
tion for certification of the Boeing 787 composite movable trailing edge, Com- [23] J.S. Wilbeck, Impact behavior of low strength projectiles, Air Force Materials
pos. Struct. 86 (2008) 258–268. Laboratory, Technical Report AFML-TR-77-134, 1977.
[8] M. Guida, F. Marulo, M. Meo, A. Grimaldi, G. Olivares, SPH – Lagrangian study [24] J.A. Zukas, T. Nicholas, H.F. Swift, L.B. Greszczuk, D.R. Curran, Impact Dynamics,
of bird impact on leading edge wing, Compos. Struct. 93 (2011) 1060–1071. Krieger Publishing Company, Malabar, FL, 1992.