Example Problem 1

Helius:MCT™ Version 2.0 for Abaqus

July, 2009

Abstract

The objective of this example problem is to demonstrate the modeling process for conducting a failure
analysis of a large composite structure using Helius:MCT. This example will demonstrate the effects of
through-the-thickness mesh density, finite element type, and failure criteria type on the progressive failure
response of the composite structure.

For questions, comments or further information, contact Firehole Technologies at

support@fireholetech.com

Legal Notices

Copyright 2009, Firehole Technologies, Inc.

Helius:MCT is a trademark of Firehole Technologies, Inc. Any use of the Helius:MCT trademark requires the prior
written consent of Firehole Technologies, Inc.

Abaqus/Standard is a trademark of Dassault Systemes S.A. and Dassault Systemes SIMULIA Corp.
 Table of Contents
E1.1 Introduction ......................................................................................................................................... 3
E1.2 Problem Description ............................................................................................................................ 4
E1.3 Modeling Comparison ......................................................................................................................... 7
E1.3.1 Linear Elastic Analysis ................................................................................................................. 7
E1.3.2 First Failure Analysis ................................................................................................................... 9
E1.3.3 Effect of Through-Thickness Mesh Density on Progressive Failure Response ......................... 11
E1.3.4 Effect of Element Type on Progressive Failure Response ......................................................... 14
E1.4 Comparison of Helius:MCT with currently available Abaqus failure criteria .................................. 17
E1.4.1 Comparison of Helius:MCT with linear elastic Abaqus failure criteria. .................................... 17
E1.4.2 Comparison of Helius:MCT with Abaqus’ progressive damage model ..................................... 20

Page 2 of 22
 Helius:MCT™ Example Problem 1

Composite Conic

E1.1 Introduction

The objective of this example problem is to demonstrate the modeling process for conducting a failure
analysis of a large composite structure using Helius:MCT (herein referred to as Helius). This example
will demonstrate the effects of through-the-thickness mesh density, finite element type, and failure criteria
type on the progressive failure response of the composite structure. For detailed instructions on how to
use Helius, please refer to the Helius:MCT User’s Guide and Tutorials 1 and 2.

In a typical composite structure, catastrophic global structural failure is precipitated by the onset and
subsequent growth (or spreading) of localized matrix and fiber constituent failures. Consequently, three
different types of failure are discussed in this example problem.

1. Local Failure (failure at a Gaussian integration point):

A. Matrix Constituent Failure

B. Fiber Constituent Failure

2. Global Failure (significant, discrete reduction in global stiffness of the structure):

This example problem involves imposing a set of quasi-static monotonically increasing loads on a
large composite structure. The overall response of the composite structure to these applied loads
is characterized by the displacement at certain points on the structure. Global structure failure is
identified by the appearance of a significant discontinuity in the structure's load/displacement
relationship indicating a large decrease in the overall stiffness of the structure.

Page 3 of 22
E1.2 Problem Description

The example problem consists of a conical, composite sandwich structure that is subjected to a quasi-
static axial compressive load that is monotonically increased until global structure failure occurs. The
example problem is similar to the load-controlled tests that are part of the composite structure’s flight
qualification testing. Figure 1 shows the finite element mesh used to represent the conical composite
structure, the load head and the load head adapter.

180°
Load Head

270° 90°
Load Head
Adapter

0°

Conical Sandwich
Composite Structure

Figure 1: Test Assembly

Four actuators are used to apply concentrated vertical compressive forces to a load head that sits on top of
the conic structure (Figure 1). Table 1 lists the compressive forces applied to each of the four positions
on the load head. As seen in Table 1, these forces result in a uniform vertical compressive load where
100% loading corresponds to a total vertical compressive force of 200,000 lbs.

Page 4 of 22
 Table 1: Load scheme (uniform compression)

Actuator Azimuth
Load Level Total Load
000 090 180 270
0% 0 0 0 0 0
50% -250 -250 -250 -250 -1000
100% -500 -500 -500 -500 -2000
- Loads are in kips
- Loading is quasi-static and is linearly ramped up
- Negative sign indicates a compressive force

An adapter is used to connect the load head to the conic structure (Figure 1). For simplicity, the adapter
and load head are considered rigid when compared to the composite conic, therefore, an arbitrarily large
elastic modulus was assigned to the isotropic material used for both the load head and the adapter. As
seen in Figure 1, an access door is cut through the side of the conic structure. The boundary conditions
for the conic structure are provided by constraining the entire bottom surface of the conic structure using
fixed boundary conditions (displacements in the 1, 2 and 3 directions are constrained to have zero
displacement).

The composite conic structure features a sandwich panel construction. The through-the-thickness profile
of the sandwich panel (Figure 2) is uniform in both the hoop and axial directions on the conic. The inner
and outer composite facesheets of the sandwich construction have a [(90/0)4] layup as measured from the
inside surface (0° is in the axial direction, 90° in the hoop direction). Each of the composite plies is
0.0075” thick and composed of carbon/epoxy AS4-3501-6. The 1” thick core is an isotropic foam
material (Rohacell 110 WF).

Outer (Bag) Surface

8 plies Outer Facesheet [(90/0)4] - AS4-3501-6 0.06"

Core - Rohacell 110 WF 1.0" 1.12"

8 plies Inner Facesheet [(90/0)4] - AS4-3501-6 0.06"

Inner (Tooling) Surface

Figure 2: Sandwich panel profile

Material properties for AS4-3501-6 and Rohacell 110 WF are provided in Figure 3. Helius:MCT
assumes all composite lamina are transversely isotropic, so a transversely isotropic material model will be
used for AS4-3501-6.

Page 5 of 22
 Property Table
AS4-3501-6 Rohacell_110WF
[psi ] [psi ]
E11 1.83E+07 E 2.61E+04
E22 = E33 1.62E+06 ν 0.286
ν12 = ν13 0.282
ν23 0.545
G12 = G13 9.51E+05
G23 5.25E+05
+
S11 2.83E+05
-
S11 2.15E+05
+ +
S22 = S33 6.96E+03
- -
S22 = S33 2.90E+04
S12 = S13 1.15E+04
S23 7.25E+03

Figure 3: Material properties for AS4-3501-6 and Rohacell_110WF

Figure 4 provides the dimensions for the composite conic structure.

Figure 4: Schematic for cross-section of composite conic (dimensions are in inches)

Page 6 of 22
E1.3 Modeling Comparison

A series of different finite element models was created to demonstrate the effect of three key modeling
issues on the predicted failure response of the composite structure:
• Number of elements distributed through the thickness of the sandwich laminate
• Type of finite element
• Type of failure criteria

Note: In this study, the surface mesh density of all finite element models remained constant.

E1.3.1 Linear Elastic Analysis

Abaqus version 6.8 provides a three-dimensional continuum element (e.g., C3D8R) that can utilize a
multilayer composite lay-up. Previous versions of Abaqus did not have this functionality. The use of
continuum elements that explicitly account for all six stress components (σ11, σ22, σ33, σ12, σ13, σ23) ensures
that the material failure criteria can utilize the transverse stress components (σ33, σ13, σ23) that are
neglected (or only approximated) in conventional or continuum shell elements. To demonstrate the
importance of explicitly accounting for the transverse stress components, a simple linear elastic analysis
is performed to allow comparison of the magnitudes of the in-plane and transverse stress components in
the most highly stressed region of the model.

Model(s)

linear_elastic_solid.inp
• Element type: C3D8R
• Through-thickness mesh density:
o Composite Facesheets: 1 element
o Foam Core: 1 element

Results

Figure 5 displays the stress state for a point in the outside surface ply near the upper left-hand corner of
the access door of the composite conic. The fiber constituent failure criterion is dependent on three stress
components (σ11, σ12, σ13). Examination of Figure 5 reveals that the transverse shear stress (σ13) is
negligible in comparison to the in-plane stresses (e.g., σ13 is only 0.5% of σ11). This indicates that
transverse stresses will not provide a significant contribution to fiber failure. On the other hand, matrix
constituent failure is driven by all six stress components. By comparing the magnitude of the out-of-
plane stresses (σ33, σ13, σ23) with the magnitude of the in-plane stresses (σ22, σ12), we can see that the out-
of-plane stresses will make a significant contribution to the matrix failure criterion. In order to accurately
predict failure in the structure, matrix failure must be captured properly, because after a matrix failure
event, its stiffness is reduced and stress redistribution causes an increased stress state in the fibers.

Because the transverse stresses are not negligible in certain regions of large composite structures, Firehole
Technologies recommends that solid elements should be used wherever possible. This will be
demonstrated when solid elements are compared with continuum shell elements later in this
documentation (Section E2.2.4).

Page 7 of 22
 Value Normalized Normalized Normalized
Stress Component (σij / σ11) (σij / σ22) (σij / σ12)
[psi]
σ11 267,057 100.0% 1045.6% 1575.6%
σ22 25,540 9.6% 100.0% 150.7%
σ33 3,623 1.4% 14.2% 21.4%
σ12 16,949 6.3% 66.4% 100.0%
σ13 1,274 0.5% 5.0% 7.5%
σ23 932 0.3% 3.6% 5.5%

Figure 5: Stress state of selected element

Page 8 of 22
E1.3.2 First Failure Analysis

As a first attempt at a progressive failure analysis, a through-thickness mesh density of 1 element per
facesheet and 1 element for the core, for a total of 3 elements through the thickness of the sandwich
construction is used. In this case, the C3D8R (reduced integration element) is used. Note that it is
possible to use a single element through the thickness of the entire laminated sandwich structure;
however, due to the relatively high compliance of the thick foam core, the use of a single element through
the laminate thickness yields very inaccurate results.

Model(s)

1elemFace_1elemCore.inp
• Element type: C3D8R
• Through-thickness mesh density:
o Composite Facesheets: 1 element
o Foam Core: 1 element

Results

Table 2 displays the load level at which each type of failure event is predicted. The criteria used to
predict localized matrix failure and localized fiber failure are described in detail in the Helius:MCT
Theory Manual. In Table 2, it should be emphasized that the matrix failure and fiber failure load levels
indicate the load at which a localized matrix or fiber failure is first detected. Note that this first instance
of matrix failure or fiber failure occurs at a single Gaussian integration point within one of the material
plies of one of the elements of the model. In a large composite structure containing thousands (or
millions) of Gaussian integration points, a very large number of localized constituent failures are
necessary in order to detect any appreciable change in the overall stiffness of the composite structure.

Global structural failure can be defined in many different ways, but for the purpose of this example
problem, global failure is defined as a large discontinuity in the composite structure’s overall vertical
load-displacement curve. The overall vertical deformation of the composite structure is quantified by
using the vertical displacement at the load application point labeled 0° in Figure 1. Because the load head
and adapter are considered rigid, a large discontinuity in the load-displacement curve is indicative of very
rapid growth (spreading) of localized material failures that occur during a particular load increment,
resulting in a large degradation of the overall stiffness of the composite structure. This definition of
global structural failure is chosen since most experimental tests are stopped at this point to prevent
damage to expensive test equipment.

Figure 6 shows the overall vertical load-deflection curve for the composite structure. Note that the
overall response of the structure appears to be linear up to a load level of 56%, at which time a global
structural failure occurs. As seen in Table 2, the first localized matrix failure occurred at a load of 49%
and the first localized fiber failure occurred at a load of 49%. However, all of the localized failures that
occurred between the load range of 49% and 56% were not sufficient to produce a visually detectable
change in the composite structure’s overall load-deflection response. When the load is increased from
56% to 57%, a very large cascade of localized failures occurs. This failure cascade is significant enough
to destroy approximately 92% of the vertical stiffness of the composite structure (the vertical
displacement increases by a factor of 12 during the load increment). This type of behavior, where the
overall response of the composite structure remains approximately linear up to global structural failure is
fairly common among structures composed of brittle composite materials.

Page 9 of 22
 Table 2: Failure events for composite conic

# Elements # Elements 1 1 1,2

Model Element Type Matrix Failure Fiber Failure Global Failure
(facesheet) (core)
1elemFace_1elemCore C3D8R 1 1 49% 49% 57%

Notes: 1: Load percentage is on a 0-100% scale

2: Global failure is defined as a large discontinuity in the load-vertical displacement curve for
the 0° application point of the load head (Figure 1)

Global Failure

Figure 6: Vertical load-displacement curve for the 0° actuator point

This first failure analysis is simply intended to provide a base-level prediction that can be used for
comparing the results from subsequent modeling efforts.

Page 10 of 22
E1.3.3 Effect of Through-Thickness Mesh Density on Progressive Failure
Response

The progressive failure response of the composite structure was predicted using four finite element
models that differed only in the number of elements distributed through the thickness of the sandwich
panel, i.e., the materials, element type, in-plane mesh density and boundary conditions were exactly the
same in each model. The four different levels of through-the-thickness mesh density are listed below.
Each of the four models uses C3D8R elements.

• Original model: 1 element to capture core and 1 element per facesheet that captures the entire
layup (3 elements through-thickness).
• 1st modification: Increase the core mesh density to allow 4 elements to capture the core profile (6
elements through-thickness).
• 2nd modification: Increase the facesheet density to allow 2 elements to capture the layup of each
facesheet (8 elements through-thickness).
• 3rd modification: Increase the facesheet density to allow 4 elements to capture the layup of each
facesheet (12 elements through-thickness).

Ideally, a very fine through-thickness mesh would be utilized to provide the most accurate loading
distribution and deformation throughout the model, but run time costs often limit mesh densities.

Model(s)

1elemFace_4elemCore.inp
• Element type: C3D8R
• Through-thickness mesh density:
o facesheets: 1 element
o core: 4 elements

2elemFace_4elemCore.inp
• Element type: C3D8R
• Through-thickness mesh density:
o facesheets: 2 elements
o core: 4 elements

4elemFace_4elemCore.inp
• Element type: C3D8R
• Through-thickness mesh density:
o facesheets: 4 elements
o core: 4 elements

Page 11 of 22
Results

Table 3 displays the load level at which each type of failure event is predicted by the four models. Figure
7 displays the overall vertical load-displacement curves for the composite structure.

Table 3: Failure events for composite conic

# Elements # Elements 1 1 1,2

Model Element Type Matrix Failure Fiber Failure Global Failure
(facesheet) (core)
1elemFace_1elemCore C3D8R 1 1 49% 49% 57%
1elemFace_4elemCore C3D8R 1 4 48% 55% 64%
2elemFace_4elemCore C3D8R 2 4 52% 52% 60%
4elemFace_4elemCore C3D8R 4 4 59% 60% 60%

Notes: 1: Load percentage is on a 0-100% scale

2: Global failure is defined as a large discontinuity in the load-vertical displacement curve for
the 0° application point of the load head (Figure 1)

Figure 7: Vertical load-displacement curve for the 0° actuator point

Page 12 of 22
Comparing mesh density results, the following observations result:

1. As the through-the-thickness mesh density is increased, the initiation of localized matrix and fiber
failure is predicted to occur at higher load levels.

- As we progressively increase the number of elements through the thickness of the laminate,
the transverse shear stiffness of the model decreases faster than the in-plane stiffness of the
model. Consequently, as the through-the-thickness mesh density is increased, the model
tends to exhibit an increase in transverse shear deformation at the expense of in-plane
deformation. The net result of this trend is that the magnitude of the peak in-plane stresses
tends to decrease as the through-the-thickness mesh density is increased, thus localized
failure is predicted to occur at higher load levels.

2. Global failure is only slightly affected by mesh density.

- Despite the fact that increasing the through-the-thickness mesh density results in delayed
local failure initiation for both the matrix and the fiber, the predicted global structural failure
load is not significantly affected by the through-the-thickness mesh density. This observation
suggests that the escalation from local failure initiation to global structural failure occurs
faster as the through-the-thickness mesh density is increased. It is reasonable to expect local
failures to progress into global failures more rapidly for a higher density mesh (as seen in
Table 3). A denser mesh results in a more refined failure path for a structure due to the
increase in number of Gaussian integration points for failure criterion to be evaluated at. A
more refined failure path facilitates a more rapid progression of failure, hence the rapid
progression of local failures into global failures.

Page 13 of 22
E1.3.4 Effect of Element Type on Progressive Failure Response

The progressive failure response of the composite structure was predicted using three finite element
models that differed only in the type of element; i.e., the materials, mesh density and boundary conditions
were exactly the same in each model. Three Abaqus element types were tested: C3D8R, C3D8 and
SC8R. Each of the three models used the same through-the-thickness mesh density, namely, four
elements were used for the foam core and each composite facesheet was divided into two elements for a
total of eight elements through the thickness of the sandwich laminate.

Model(s)

C3D8.inp
• Element type: C3D8
• Through-thickness mesh density:
o Each composite facesheet: 2 elements
o Foam Core: 4 elements

SC8R.inp
• Element type: SC8R
• Through-thickness mesh density:
o Each composite facesheet: 2 elements
o Foam Core: 4 elements

Results

Table 4 displays the load level at which each type of failure event is predicted by the three models.
Figure 8 displays the overall vertical load-displacement curves for the composite structure predicted by
the three models.

Table 4: Failure events for composite conic

# Elements # Elements 1 1 1,2

Model Element Type Matrix Failure Fiber Failure Global Failure
(facesheet) (core)
2elemFace_4elemCore C3D8R 2 4 52% 52% 60%
C3D8 C3D8 2 4 41% 47% 55%
SC8R SC8R 2 4 42% 43% 68%

Notes: 1: Load percentage is on a 0-100% scale

2: Global failure is defined as a large discontinuity in the load-vertical displacement curve for
the 0° application point of the load head (Figure 1)

Page 14 of 22
 Figure 8: Vertical load-displacement curve for the 0° actuator point

Comparing reduced integration continuum elements with fully integrated continuum elements (C3D8R
vs. C3D8), the following observations and explanations are given:

1. A fully integrated element type results in local failure events (matrix and fiber failure) occurring
at lower load levels.

- The fully integrated element utilizes more Gaussian integration points than the reduced
integration element. Even if both elements predict the same element average stresses, the
fully integrated element will predict a higher local peak stress than the reduced integration
element simply because it has more Gaussian integration points and it contains Gauss points
that are closer to the element's boundaries (where the linear stress distribution attains a
maxima). Consequently, the fully integrated element will predict localized failure initiation
at a lower load level than the reduced integration element.

2. Global failure occurs at a lower load level with the fully integrated continuum element.

Page 15 of 22
 - The difference in the global failure load predicted by the two elements is due primarily to
difference in the local failure initiation load. The reduced integration element, with a single
Gauss point per material ply, provides a more discretized representation of failure cascading.
In other words, when failure occurs in a material ply of a reduced integration element, the
stiffness of the entire ply is reduced and a relatively large amount of load re-distribution
occurs. In contrast, when failure occurs at one of the Gauss points in a material ply of a fully
integrated element, only part of the material ply experiences a stiffness reduction and a
relatively small amount of load re-distribution occurs.

- One might be tempted to think that a fully integrated element inherently provides a more
realistic progressive failure response since it contains numerous integration points where the
failure criteria is tested and stiffness reduction is imposed. This is not true because the
integration points of a fully integrated element are not the most accurate locations for
computing stress. In fact, the most accurate locations for computing stress are the reduced
integration points1. Consequently, the reduced integration elements evaluate failure and
stiffness reduction at fewer points, but the evaluation itself is more accurate because the stress
state is more accurate at the reduced integration points. Therefore, the use of more
integration points per element does not necessarily result in a more accurate analysis.

In comparing the failure response of the reduced integration continuum elements (C3D8R) with reduced
integration continuum shell elements (SC8R), the following observations and explanations are given:

1. The continuum shell elements predict local failure initiation at lower load levels (42%) than the
reduced integration continuum elements (52%).

- Even though both elements make use of the exact same set of Gaussian integration points, the
SC8R elements predict higher in-plane stress components than the C3D8R elements. This
discrepancy is due to differences in the transverse shear and transverse normal stiffnesses of
the two elements. The C3D8R element obtains its transverse stiffnesses by simply
integrating the stiffness of the individual material plies over the volume of the element.
However, the SC8R element obtains its transverse stiffnesses as direct user input that applies
to the element as a whole. This scenario effectively precludes the possibility of both elements
exhibiting the same transverse stiffness. Any change in the transverse stiffness of an element
will necessarily result in a different division of the element's total strain energy into in-plane
and out-of-plane components. Therefore, if the two elements have different transverse
stiffnesses, then they will likely have different in-plane stress components. In this particular
problem, the initial matrix failure is primarily driven by the in-plane shear stress which is
19% larger in the SC8R element than in the C3D8R element; consequently, the SC8R
element predicts earlier localized matrix failure.

1
See Section 14.2 of The Finite Element Method by O.C. Zienkiewicz and R.L. Talyor, Fifth Edition,
Butterworth Heinemann, Oxford, 2000.
 2. Global failure occurs at a higher load level with the continuum shell element even though local
failure occur at much lower load levels.

- Even though the SC8R element predicts initiation of localized constituent failure at a lower
load level than the C3D8R element, the SC8R element predicts that global structural failure
occurs at a higher load level (68%) than the C3D8R element (60%). The fact that the SC8R
predicts a more gradual failure cascade process is due primarily to the fact that localized
material failures do not affect the transverse stiffnesses (E33, G13, G23) in the SC8R element
(Abaqus requires these stiffnesses to be constant in the SC8R element). Since the transverse
stiffnesses do not experience any degradation, the SC8R elements can more easily
accommodate load re-distribution without causing additional localized failures. To
summarize, the C3D8R and SC8R elements exhibit very different failure cascade behavior.

E1.4 Comparison of Helius:MCT with currently available Abaqus failure criteria

Abaqus provides five different failure criteria that can be used to predict composite material failure in a
linear elastic analysis. In addition Abaqus provides one type of damage evolution model that can be
applied to composite materials. In this section, the various Abaqus failure criteria and damage evolution
criteria will be compared with Helius:MCT by using each one to simulate the failure response of the
composite conic structure.

E1.4.1 Comparison of Helius:MCT with linear elastic Abaqus failure criteria.

In Abaqus Standard, five failure criteria are provided for use in linear elastic analyses: four stress-based
criteria and one strain-based criterion. The practical utility of these linear elastic failure criteria is quite
limited because they only predict the occurrence of localized failure, not the consequences of localized
failure. In other words, when one of the Abaqus failure criteria predicts that failure occurs, there is no
accompanying stiffness reduction. Therefore, the processes of load re-distribution and progressive failure
are not represented. In contrast, Helius:MCT not only predicts localized failure, it also predicts localized
stiffness reduction. In further contrast, the Abaqus linear elastic failure criterion utilizes the homogenized
composite state of stress or strain to predict failure of the homogenized composite material, whereas
Helius:MCT utilizes constituent average stress to independently predict failure of each constituent
material.

In this example problem, the composite conic can be assumed to be linearly elastic prior to global failure.
In that respect, failure initiation predicted by Helius:MCT can be compared with the failure initiation
predicted by the different Abaqus failure criteria. The failure criteria provided by Abaqus are applicable
to generally orthotropic materials, but for the purpose of this example problem, a transversely isotropic
material model is used for both the Abaqus failure criteria and Helius. One of the shortcomings of the
Abaqus linear elastic failure criteria is that they are based entirely upon an assumed condition of plane
stress; consequently they can only be used in plane stress 2-D continuum elements or shell elements.
While Helius:MCT uses a general 3-D state of stress in its failure criteria, continuum shell elements
(SC8R) are used in the comparison in order to accommodate the limitations of the Abaqus linear elastic
failure criteria.

In the case of composite laminates, laminate failure is taken to be first ply failure. For brevity, only the
max-stress and Tsai-Wu failure criteria (stress-based failure criteria) are compared with Helius (stress-

Page 17 of 22
based). For more information on the max-stress and Tsai-Wu failure criteria formulation, see section
18.2.3 of the Abaqus Analysis User’s Manual.

Model(s)

SC8R.inp
• Failure criteria: Helius:MCT™
• Element type: SC8R
• Through-thickness mesh density:
o facesheets: 2 elements
o core: 4 elements

stressBasedFailure.inp
• Failure criteria:
o Max-Stress and
o Tsai-Wu
• Element type: SC8R
• Through-thickness mesh density:
o facesheets: 2 elements
o core: 4 elements

Results

Table 5 shows the load level at which each type of failure event is predicted using the three different
failure criteria. Figure 9 displays the load-vertical displacement curves for the load application point at
0°. Note that the max-stress and Tsai-Wu failure criteria only provide information on failure initiation for
the homogenized composite material, hence these criteria predict that matrix failure and fiber failure both
initiate at the same load level (50-51%). Helius:MCT predicts that the matrix constituent failure initiates
at a load of 42% and fiber constituent failure initiates at a load of 43%.

Table 5: Failure events for composite conic

# Elements # Elements
Model Criterion Element Type Matrix Failure1 Fiber Failure1 Global Failure1,2
(facesheet) (core)
4
Max Stress SC8R 2 4 51% 51%4 n/a3
stressBasedFailure 4 4
Tsai-Wu SC8R 2 4 50% 50% n/a3
SC8R Helius:MCT™ SC8R 2 4 42% 43% 68%

Notes: 1: Load percentage is on a 0-100% scale

2: Global failure is defined as a large discontinuity in the load-vertical displacement curve for
the 0° application point of the load head (Figure 1)

3: Global failure is unable to be determined, see text for details

4: Composite laminate failure

Figure 9 displays the load-displacement curves for the load application point at 0°. Before global failure
is achieved (68%), both methods produce similar predictions for overall structural stiffness. In using the
linear elastic material failure criteria that are available in Abaqus, material stiffness degradation is not
accounted for; therefore, the failure analysis can only proceed up to the point where localized failure
initiation is predicted. The capability of Helius to degrade material stiffness at integration points that

Page 18 of 22
have failed (either matrix of fiber failure) lends the unique capability of accurately capturing the
progression of local failure initiation into eventual global failure. Along with this, the user can accurately
describe the post-fail behavior of the structure that is not available with the Abaqus failure criteria.

Failure initiation:
50% - Tsai-Wu criterion
51% - Max Stress criterion

Figure 9: Vertical load-displacement curve for the 0° actuator point

Page 19 of 22
E1.4.2 Comparison of Helius:MCT with Abaqus’ progressive damage model

For composite materials, Abaqus provides a progressive damage model that uses the Hashin criteria to
predict the initiation of four different constituent failure modes and uses damage evolution equations to
predict the post-initiation stiffness degradation that results from the evolution of each of the four
constituent damage modes. This section provides a comparison between the progressive failure model
used in Helius:MCT and the progressive damage model provided by Abaqus.

There are several fundamental differences between the progressive failure model used in Helius:MCT and
the progressive damage model provided by Abaqus. These differences are discussed below.

1. In the Abaqus progressive damage model, the material stiffness is gradually reduced as
deformation continues to accumulate after the initiation criterion is met. In contrast, Helius:MCT
imposes an instantaneous stiffness reduction that is determined by the particular constituent that
failed. This type of instantaneous, discrete stiffness reduction normally poses severe convergence
difficulties for finite element codes; however, Helius:MCT is specifically developed to efficiently
handle this type of behavior and exhibits very robust convergence behavior.
2. In the Abaqus progressive damage model, damage initiation and damage evolution of the material
constituents (fiber and matrix) are predicted based on the composite average states of stress and
strain. In contrast, Helius:MCT uses the constituent average stress states to predict failure in
each of the individual material constituents.
3. The Abaqus progressive damage model predicts damage initiation and damage evolution based
solely on the in-plane stress and strain components, ignoring the contribution of the transverse
stress and strain components. In contrast, Helius:MCT predict constituent failure using the full 3-
D constituent average stress state. However, even if one ignored the contribution of the
transverse stress components, the functional form of the in-plane stress contribution is still
different in Helius:MCT and the Abaqus progressive damage model.
4. As damage evolves, the Abaqus progressive damage model only accounts for stiffness reduction
in the in-plane stiffnesses (E11, E22, G12), leaving the transverse stiffnesses (E33, G13, G23)
unchanged. In contrast, as material constituent failure occurs, Helius:MCT explicitly accounts
for stiffness reduction in both the in-plane and transverse stiffnesses.
5. The Abaqus progressive damage model can only be used in conjunction with 2-D continuum
elements and shell elements. Helius:MCT can be used in conjunction with 2-D continuum
elements, shell elements, and 3-D continuum elements.

In this section, the Abaqus progressive damage model and Helius:MCT are each used to simulate the
failure response of the composite conic structure. Due to the differences listed earlier in items 4 and 5
above, continuum shell elements (SC8R) are used in both models. This choice eliminates the differences
listed in items 4 and 5; consequently, any differences in the predicted failure response of the structure are
due entire to the differences listed in items 1-3 above.

In addition to the fundamental mathematical differences listed above, the Abaqus progressive damage
model poses addition difficulty to the user in that it is rather difficult and confusing to define. The
process of defining the Abaqus damage initiation (Hashin) criteria is straightforward and requires only
industry standard strength measurements (same as Helius:MCT):

• Longitudinal tensile and compressive lamina strength (+S11 and –S11)

• Transverse tensile and compressive lamina strength (+S22 and –S22)

Page 20 of 22
 • Longitudinal and transverse shear strength of the lamina (S12 and S23)

However, the process of defining the Abaqus damage evolution relationships is difficult and confusing.
The four parameters required for the damage evolution relationships represent the amount of energy that
is dissipated in each of the four constituent failure modes:

• Energy dissipated during damage for fiber tension (Gcft)

• Energy dissipated during damage for fiber compression (Gcfc)
• Energy dissipated during damage for matrix tension (Gcmt)
• Energy dissipated during damage for matrix compression (Gcmc)

These energy dissipation constants are not readily available for most unidirectional composites (including
AS4-3501-6). Furthermore, the convergence performance of the Abaqus finite element code is very
sensitive to the numerical values chosen for these energy dissipation constants; therefore, even if the user
has access to experimentally measured energy dissipation values, it is likely that the user will have to
adjust these values in an effort to improve the convergence behavior of the finite element solution. In
contrast, Helius:MCT does not require these energy dissipation values and Helius:MCT actually improves
the convergence behavior of the finite element solution rather than degrading the convergence behavior.

The progressive failure response of the composite conic structure was simulated using different models
(listed below). Both models were identical except for the failure criteria that were employed. The
SC8R.inp model used Helius:MCT to predict constituent material failure and stiffness reduction, and the
Hashin.inp model used the Abaqus progressive damage model to predict constituent material failure and
stiffness reduction.

Model(s)

SC8R.inp
• Failure criteria: Helius:MCT
• Element type: SC8R
• Through-thickness mesh density:
o facesheets: 2 elements
o core: 4 elements

Hashin.inp
• Failure criteria: Hashin
• Element type: SC8R
• Through-thickness mesh density:
o facesheets: 2 elements
o core: 4 elements

Results

For this particular example problem, a converged solution could not be obtained using the Abaqus
progressive damage model after the first localized fiber failure occurred. Consequently, the results
reported here do not include a global structural failure load predicted with the Abaqus progressive
damage model. Table 6 shows the load level at which each type of failure event is predicted using the
two different progressive failure models.

Page 21 of 22
 Table 6: Failure events for composite conic

# Elements # Elements
Matrix Failure1 Fiber Failure1
1,2
Model Criterion Element Type Global Failure
(facesheet) (core)
3
Hashin Hashin SC8R 2 4 50% 71% n/a
SC8R Helius:MCT™ SC8R 2 4 42% 43% 68%

Notes: 1: Load percentage is on a 0-100% scale

2: Global failure defined as large discontinuity in load-vertical displacement curve for 0°

application point of load head

3: Global failure is unable to be determined, see text for details

As seen in Table 6, the Hashin criterion predicts that localized matrix failure occurs at a higher load level
(50%) than predicted by Helius:MCT (42%). This difference is due entirely to the differences in the
matrix constituent failure criteria used by Helius:MCT and the Hashin model; e.g., constituent average
stress versus composite average stress and the specific functional form of the stress-based failure criteria
used in each model.

Helius:MCT predicts that the first localized fiber constituent failure occurs at a load level of 43%.
However, the Hashin criterion predicts that the first localized fiber constituent failure does not occur until
the load level has reached 71%. The large difference in the load at fiber failure initiation is due primarily
to two issues. First Helius:MCT uses the fiber average stress state, while the Hashin criterion uses the
homogenized composite average stress state. Second, Helius:MCT predicts an instantaneous stiffness
reduction in conjunction with localized matrix failure; therefore, load is re-distributed to the fibers much
more rapidly than predicted by the Abaqus damage evolution model which uses a gradual stiffness
reduction.

Helius:MCT predicts that global structure failure occurs at a load level of 68%. Interestingly, the Abaqus
damage evolution model did not even predict the initiation of localized fiber failure until a load level of
71%. It should be emphasized that after the initiation of localized fiber failure, a converged solution
could no longer be obtained with the Abaqus damage evolution model, thus a global structural failure
load could not be determined.

Page 22 of 22