6th BETA CAE International Conference

ANALYSIS OF A PRESSED COMPOSITE AUTOMOTIVE TAILGATE

USING ANSA & μETA
Andy Ngai, Mark Arnold
Penso, UK

KEYWORDS – Composites, Design Optimization, Failure, Finite Element Analysis, Laminate

ABSTRACT – Composite materials utilised within the automotive industry have increased in
recent times, due to the demand to reduce weight and vehicle C02 emissions. Penso were
commissioned by a European automotive OEM to design and manufacture a prototype
continuous fibre reinforced composite tailgate. The main objective for the composite design
was to minimise mass, whilst meeting all strength and stiffness targets of the production
SMC tailgate. Three main parts of the tailgate considered for light weighting were the tailgate
inner panel, tailgate outer panel and spoiler mechanism carrier. The laminate tool function
within the pre-processor ANSA was utilised for ease of building, handling and modifying the
composite model. A laminate design was developed using non-linear static FE analysis.
Results were plotted using the post-processor μETA, with the Tsai-Wu failure criterion and a
user defined inter-laminar failure criterion used as design metrics. The laminate design was
then assessed by Penso’s composite manufacturing team to develop overlap joint locations
and preliminary ply shapes appropriate for processing utilising Penso’s pressed composite
technology. Ply shapes and draped material directions were updated in the FE model for
further refinement and prediction of finalised mass/performance. The final composite design
resulted in a combined mass save of over 65% whilst having comparable structural
performance to the SMC production design.

TECHNICAL PAPER -

1. INTRODUCTION

Penso were commissioned to design and manufacture a continuous fibre reinforced

composite tailgate. This would be a lightweight alternative to the current production tailgate
which is manufactured from Sheet Moulding Compound (SMC), a discontinuous glass-fibre
reinforced polymer material.

2. TAILGATE DESIGN

The production tailgate comprises of a bonded sub-assembly of the inner and outer SMC
panels, with a steel spoiler mechanism carrier bolted to the outer panel, as illustrated in
Figure 1. Metallic reinforcing mounting plates for the hinges, gas struts and latch are bonded
to the inside of the inner panel, with the rear screen glass bonded to the outer panel. The
individual masses of the three main panels are 5.7 kg, 3.2 kg and 1.1 kg for the tailgate SMC
inner panel, tailgate SMC outer panel and the steel spoiler mechanism carrier respectively,
providing a total mass for the three panels of 10.05kg.
The three main parts of the tailgate that were considered for light weighting; were the tailgate
inner panel, tailgate outer panel and the spoiler mechanism carrier. The requirements of the
new composite design had to meet all existing strength and stiffness performance targets,
whilst also preserving existing A-surface geometry and hardware. The load cases used to
assess the structural performance targets against of the tailgate were torsion, cantilevered
bending, latch load and margin & flushness.
6th BETA CAE International Conference

Figure 1. Tailgate parts considered for light weighting.

The composite parts comprised a continuous fibre reinforced polymer material, using pre-
impregnated (prepreg) epoxy resin carbon fibre reinforcement. The composite parts were
laminated from a single prepreg composite sheet material. The carbon fibre material used
was the Cytec 2020-40%-6KHS-2x2T-400, which is a 2x2 Twill woven fabric, supplied by
Cytec Industries. Static coupon tests were performed for this prepreg to obtain mechanical
properties in longitudinal (warp) and lateral (weft) fibre directions. Validation of these material
models were carried out through simulation of coupon testing performed according to various
ASTM test methods (Arnold et al, 2014).
Penso’s pressed composite technology will be utilised to manufacture the composite parts of
the prototype tailgate, this is a process in which the prepreg composite material is placed into
an open matched metal mould cavity. The tool is then closed under pressure to encapsulate
the fibre and resin compound within the mould cavity. Heat and pressure are maintained until
the composite material has cured.

3. LAMINATE OPTIMISATION

3.1. Finite element model

ANSA pre-processing software was used to import CAD geometry and converted into a
surface suitable for shell meshing. The mesh was created on the A-surface which
corresponded to the preform tool surface so that the laminate plies stacked from A to B
surface. Each panel was meshed with first order conventional S4 and S3R shell elements,
having mesh size of approximately 5mm. These elements were utilised for their accurate
solutions for in-plane bending problems.
The laminates tool within ANSA was utilised for the ease of building, handling and modifying
of all the layers of the composite model, which exploits a special ANSA laminate property.
This ANSA laminate property simplifies the number of layers the user has to manage during
pre and post-processing e.g. the inner panel comprised 1 laminate having 49 layers
compared to the 244 composite shell section properties having a total of 1351 plies when
written out to the Abaqus/Standard FE solver. The 244 composite shell section properties
6th BETA CAE International Conference

that exist, when ANSA comments are not read in, are shown in Figure 2. With fewer plies to
manage there is less room for error during pre and post-processing. As draping simulation
was not performed during the development of the prototype tailgate, the material orientation
was defined within ANSA by projecting a vector of the fore/aft direction onto every element.

Figure 2. Composite shell properties on the inner panel.

Abaqus/Standard 6.12-2 was used for the non-linear static finite element analysis of the
tailgate. The carbon fibre panels were modelled with laminate properties defined using
*SHELL SECTION, COMPOSITE. For the orthotropic carbon fibre skins an *ELASTIC,
TYPE=LAMINA material card was used with *FAIL STRESS to define the strength terms.
The modulus of elasticity for warp and weft directions were both set to the minimum of these
two values obtained from tensile test data, however it was noted that these values differed
from those derived by compression tests. The tensile/compressive strengths in warp and
weft directions were both reduced to the open hole tensile/compressive strength to
incorporate hole and defect tolerance, a method proposed by Niu (1992). The resulting knock
down factors were 0.75 for tensile strength and 0.45 for the compressive strength.
To complete the tailgate model, the adhesive was modelled using solid elements, referencing
isotropic elastic materials properties, connected to neighbouring panels using the distributing
coupling constraints. The metallic mounting plates were modelled as shell elements
referencing non-linear elastic-plastic material properties, whilst the rear glass was modelled
with shell elements referencing isotropic elastic material properties.
3.2. Deriving an optimised laminate design
An optimised layup for the composite panels, which disregarded joints between all-over plies,
was derived to comply with the structural performance targets, whilst also minimising mass.
The optimised layup was derived through the use of analysis tools within the μETA post-
processor. The μETA post-processor was used to extract results for the non-linear static
analysis of the composite tailgate for comparison with structural performance targets. The
CompositePost Toolbar within μETA was used to process the composite materials within the
analysis. The toolbar offers a set of integrated specialized functionalities which is within a
single user interface for ease of use. Useful functionalities available to the user are the
possibility to Identify layer id of max failure, Max/min of all layers, Identify critical direction, 2d
DNA plot for failure criterion results for multiple elements and a 2D plot of failure criterion and
stress tensor per layer can also be plotted.
Within the 2d plot window there are three plots which assist with the analysis of the
composite material parts showing detailed information regarding each element. In Figure 3,
detailed information regarding the element which had the highest failure criterion is shown for
an earlier preliminarily layup; the layup at this element is [45/0/0/45].
6th BETA CAE International Conference

Figure 3. 2D plot vs thickness of preliminarily layup.

The left part of the plot illustrates the distribution of the layers along the total thickness of the
element. The middle part of the plot contains three curves corresponding to the stress tensor
components results along the thickness. The right part of the plot contains a curve
corresponding to the failure criterion result along the thickness, where failure criterion is a
composite design metric used to determine if ply failure will occur within the laminate. This
failure criterion plot is the most useful plot to understand and reduce the failure criterion
value. In Figure 3 the plies which exceed or are close to the critical value of 0.67 are plies 1
and 4 with a value of 0.84 and 0.66 respectively. To reduce this value, the material of the 45°
plies, 1 and 4, were changed from the 240gsm material to the 400gsm, which increased the
thickness of the ply from 0.275mm to 0.441mm. The failure criterion for ply 2 and 3 were
seen to be relatively low with a value of 0.29 and 0.23 respectively, therefore an opportunity
to remove one of these 0° plies was seen possible to reduce mass without compromising the
strength excessively.

Figure 4. 2D plot vs thickness of optimised layup.

6th BETA CAE International Conference

Figure 4 shows a 2d plot vs thickness for the same element as was shown in Figure 3,
incorporating the change in thickness of the 45° plies to 0.441mm thick and removing a
single 0° ply giving a layup of [45/0/45]. The failure criterion value has now decreased to
0.39, 0.08 and 0.37 for ply 1, 2 and 3 respectively, which are all below the critical value of
0.67. This method shows that the element with the highest failure criterion value complies
with the critical value, therefore the composite parts complies with the structural performance
target and the layup was deemed optimised.
3.3. Optimised laminate design
To reduce the possibility of distortion during the manufacturing process the composite panels
were constrained to have a minimum thickness of two all-over plies and a symmetric
balanced laminate.
The optimised laminate for the outer panel consists of two all-over plies at 45° relative to the
material orientation axes, with an extra local reinforcement ply at 0° positioned where the
spoiler mechanism carrier is attached. The optimised laminate for the inner panel comprises
three all-over plies at 45°, 0° and 45°, with three extra local reinforcement plies at 45°, 0° and
45° enclosed within the main laminate. These reinforcement plies are positioned at the four
corners of the panel to improve global stiffness, and also locally at the latch, hinge and gas
strut mounting region to improve the strength within these areas. Figure 5 shows a thickness
plot of the optimised layup for the inner and outer tailgate panels. The laminate for the spoiler
mechanism carrier simply consists of three all-over plies at 0°, 45° and 0.

Figure 5. Optimised laminate design for tailgate inner and outer panel.
3.4. Finite element results for optimised laminate design
For post processing, Tsai-Wu Failure Measure and a user defined Inter-Laminar Failure
Criterion were used as composite design metrics to determine whether ply failure or
delamination would occur. μETA post-processor was utilised to plot Tsai-Wu Failure Measure
results. The Tsai-Wu Failure Measure values are plotted at the in-plane integration points at
the mid-section point per ply. The in-plane integration point results were used since these are
the locations where results are typically more accurate (Cook et al, 1989). It was also
believed that by extrapolating the results at the integration points to the corners may
introduce additional uncertainties in failure measures. For the user-defined Inter-Laminar
Failure Index, output values were taken at surface section points of each ply. Nodal
displacements were taken at certain pre-defined nodes, depending on the load case, to
ensure compliance with stiffness targets.
3.4.1. Failure index
For the post processing of the laminate, the Tsai-Wu Failure Measure was used. The Tsai-
Wu Failure Measure is used with composite materials under loading. A value greater than 1
6th BETA CAE International Conference

indicates failure. The Tsai-Wu Failure Measure is proportional to applied load, unlike other
available failure criterion’s such as Failure index and Strength Ratio, with the maximum value
indicate high stress regions, this makes it more intuitive for the user to post process results.
A safety factor of 1.5 was considered for strength assessment to account for load variability,
(Niu, 1992). Rather than incorporate this safety factor into the material strength terms, this
was instead incorporated during post processing by ensuring the maximum Tsai-Wu Failure
Measure value was <0.67. This was also believed to provide some robustness against
variation in material allowables and uncertainty in FEA methods.
Figure 6 shows a Tsai-Wu Failure Measure plot for the torsion load case. For this load case,
a vertical Z-direction load of 267N is applied at the right hand bump stop, while the left hand
bump stop is constrained in z-direction and the hinges constrained in all DOF. The max Tsai-
Wu Failure Measure is 0.36, indicating ply failure will not occur.

Figure 6. Tsai-Wu failure measure plot for torsion load case.

3.4.2. Interlaminar shear
Interlaminar shear stress, an indication if delamination would occur, is analysed through the
use of the output variable transverse shear stress. Transverse shear stress (TSHR) was
requested in Abaqus/Standard to output results for transverse shear stress in the 13-
component, τ13, and transverse shear stress in the 23-component, τ23, for each ply boundary
for every element. The resultant transverse shear stress was calculated from τ13 and τ23 and
this can then be used in a user defined equation shown in Equation 1 to calculate the Inter-
Laminar Failure Index (Bednarcyk et al, 2008). The allowable Inter-laminar shear strength
Fsu13 and Fsu23 was assumed to be the same for the woven fabric.

(1)
Figure 7 shows a resultant transverse shear stress plot for the torsion load case. The max
resultant transverse shear stress was found to be 9.79 MPa, which gave an Inter-Laminar
Failure Index of 0.20, indicating delamination would not occur.
6th BETA CAE International Conference

Figure 7. Resultant transverse shear stress plot for torsion load case.
3.4.3. Adhesive
Major Principal Stress and Max Shear Stress along with the corresponding allowable
stresses for tensile and shear, were used to calculate a failure index to determine if tensile
and shear failure would occur within the adhesive. This method, when combined with the
linear elastic material model, gives a conservative result. The stresses found within the
analysis were very low, therefore a more detailed assessment for the adhesive was
considered unnecessary.
3.4.4. Displacement
Figure 8 shows a displacement plot for the torsion load case. By using the z-deflection at the
unconstrained right hand bump stop and the distance between the two bump stops, the twist
angle can be derived using trigonometry. The derived twist angle was 1.63 degrees.

Figure 8. Displacement plot for torsion load case.

6th BETA CAE International Conference

3.4.5. Results summary

Table 1 shows the results of the optimised composite tailgate for the torsion load case,
discussed in sections 3.3.1, 3.3.2 and 3.3.4, along with the results for the other three tailgate
load cases. From Table 1 it can be seen that the stiffness results comply with the targets set,
apart from the torsion load case where the twist angle slightly surpasses the 1.6 degree
target by 2%. However, this value was considered to be acceptable in the expectation that
the additional overlaps into the final tailgate model will improve the torsional performance.
The strength results also comply with Penso’s target of <0.67 for Tsai-Wu Failure Measure
and Inter-laminar shear failure index. It was found that the composite tailgate laminate design
was driven mainly by the stiffness requirements, particularly for the torsion load case, which
is a important reason the tailgate outer and inner panel laminates are biased to plies oriented
at 45 degree.
Deflection Deflection Max Tsai-Wu Max Interlaminar
Load case
Target Result Failure Measure Shear FI
Latch Load < 1 mm 0.11 mm 0.14 0.05
Torsion < 1.6 degrees 1.63 degrees 0.36 0.20
Margin & Flushness < 2 mm 1.29 mm 0.53 0.28
Cantilevered Bending < 0.85 mm 0.02 mm 0.24 0.12

Table 1. Optimised tailgate load case results.

4. DESIGN FOR MANUFACTURE

The optimised laminate design was evaluated by Penso’s composite manufacturing team to
develop the preliminary ply shapes and overlap joint locations. That are suitable for manual
draping and applicable to Penso’s pressed composite technology. A preform tool was
manufactured to the preliminary A-surface geometry and used for layup trials. Due to the
complex geometry of the tailgate inner and tailgate outer panels, each of the all-over plies
was subsequently divided into approximately twelve separate overlapped pieces to enable
draping of the prepreg. The overlap distance between any two plies was typically 10mm, to
ensure a smooth A-Surface the step in laminate thickness was located on the B-Surface.
Attention to the detailed joint architecture was essential as the use of butt joints alone does
not allow the transfer of loads from one ply to another, unlike overlap joints. Therefore,
overlap joints were considered to be more desirable. Figure 9 shows an example of a
staggered overlap joint used. The lower ply of an overlap joint is butt jointed to the upper ply
of the previous overlap joint. This method of staggering the overlaps reduces the thickness
variation across the joint, which makes things easier during the manufacturing stages.

Figure 9. Staggered overlap joint between three plies.

Once the ply layup and overlap locations had been developed for manufacture, the ply
boundaries documented by the laminator in their draped configuration were passed to
Penso’s analysis team along with detailed information concerning fibre distortion. These ply
boundaries, overlap locations and material directions were then incorporated into the FE
model for prediction of refined mass/performance data and the generation of layup manuals.
Material directions gained from layup trials were compared to those in the analysis model
and modifications were made where large inconsistences existed between the two sets of
data. As the parts were to be manufactured using layup books generated directly from the FE
models the ply shapes needed to be accurately defined.
Figures 10 and 11 show thickness plots for the updated FE model of the tailgate outer and
inner panels respectively, incorporating the ply boundaries and overlap locations determined
by the manufacturing team.
6th BETA CAE International Conference

Figure 10. Thickness plot of final tailgate outer panel including overlap joints.

Figure 11. Thickness plot of final tailgate inner panel including overlap joints.
The load cases for the updated tailgate model which included overlap joints were rerun. It
was observed that due to the thicker regions added to represent the overlaps the
performance of the structure increased marginally, as shown in Table 2. As classical
laminate shell theory is being used to represent the laminate, the performance is expected to
increase due to the increase in thickness at these overlap regions. However this does not
accurately represent the joint, as the reduction in strength due to the discontinuity of the plies
is not considered.
Deflection Deflection Max Tsai-Wu Max Interlaminar
Load case
Target Result Failure Measure Shear FI

Latch Load < 1 mm 0.10 mm 0.12 0.04

Torsion < 1.6 degrees 1.30 degrees 0.25 0.14
Margin & Flushness < 2 mm 1.83 mm 0.48 0.20
Cantilevered Bending < 0.85 mm 0.02 mm 0.21 0.10

Table 2. Final tailgate load case results.

6th BETA CAE International Conference

By incorporating the overlaps within the model, it is possible to determine a more accurate
mass prediction for the composite parts. The mass increase due to the overlaps was
approximately 10% per panel. Table 3 shows a mass breakdown for the three composite
panels, it can be seen that the mass of these panels were 2.3kg, 0.9kg and 0.2kg for the
tailgate inner, tailgate outer and spoiler mechanism carrier respectively. This equates to a
total mass of 3.4kg, which is a 6.6kg or 66% mass reduction compared to the current
production tailgate.

Production SMC Mass Prototype CFRP Mass

Tailgate Inner 5.7 kg 2.3 kg

Tailgate Outer 3.2 kg 0.9 kg
Spoiler Mechanism Carrier 1.1 kg 0.2 kg
Total 10.0 kg 3.4 kg

Table 3. Final tailgate mass.

By modelling the layup from the A-Surface and taking into account the ply shapes and ply
thickness, the B-Surface could be generated for the composite parts of the tailgate. The A
and B-surfaces of the composite parts can then be used to create tool geometry for
manufacturing.

5. CONCLUSION

The main objective of this project was to demonstrate that by using a composite material on
the production tailgate weight reduction could be achieved, whilst also achieving comparable
results to the targets/baseline. Results from the FE analysis indicate that all strength and
stiffness targets were achieved for the composite tailgate, whilst having equivalent or
improved structural performance to that of the current production tailgate. Total mass of the
prototype carbon fibre tailgate panels was 3.4kg, providing a mass saving of 6.6kg or 66%,
compared to the corresponding SMC/steel panels. However, the overall mass saving of the
whole tailgate assembly was greater than 66%. This is due to the removal of extra
components such as the anti-pinch strips, addition of a new lightweight rear screen glass and
also incorporating the interior trim into the carbon fibre parts.
Further improvements which could have been made to improve the design are by
redesigning the geometry of the tailgate so that it is specific for carbon fibre use. The section
depth could be reduced, rather than use carry over geometry designed for lower strength and
stiffness materials. This would improve the draping of the material and also reduce the
complexity during the pre-forming phase, which will improve the optical quality of the finished
composite parts. Location of the ply overlap joints were positioned to optimise the global
stiffness, whereas preferably they should be located in less visible areas to provide the best
aesthetics. Metallic mounting plates, which have been carried over from the original
production design, such as at the hinge, latch and gas strut mounting areas, could be
replaced by metallic inserts to reduce the assembly part count and total mass.
Further work has been carried out since, to improve the process of Penso’s pressed
composite technology. Various draping simulation softwares have been assessed to
determine how accurately they are able to predict ply orientation/distortion due to draping
over complex geometry. The accurate material orientations would ensure the analysis
models accurately reflect the design intent data and would evaluate ply producibility during
the design phase to avoid any potential difficulties which may occur during manufacturing.
Robustness studies have been carried out to produce a practical method for quantifying the
variability of continuous fibre-reinforced composite structures to uncertainty in ply orientation.
Continuous progress is being carried out with Penso’s laminators to develop the layup of the
tailgate composite panels to reduce laminating time, yet still achieve equivalent structure
performance.
6th BETA CAE International Conference

6. ACKNOWLEDGEMENTS

This project was co-funded by Innovate UK as part of the IDP7 Low Carbon Vehicles
programme, in which Penso were a lead manufacturing partner.

7. REFERENCES

(1) Arnold, M., Kilby, C. & Ngai, A. (2014). Validation of Laminated Composite Shell
Elements and Material Models within Various FEA Solvers. Proceedings of NAFEMS
UK 2014 Conference, Oxford, June 2014.
(2) Bednarcyk, B.A., Aboudi, J., Yarrington, P.W., Collier, C.S., (2008). Simplified Shear
Solution for Determination of the Shear Stress Distribution in a Composite Panel from
the Applied Shear Resultant. Proceedings of 49th AIAA/ASME/ASCE/AHS/ASC
Structures, Structural Dynamics, and Materials Conference, Schaumburg, IL, April
2008.
(3) Cook R.D., Malkus D.S. & Plesha M.E., (1989). Concepts and Applications of Finite
Element Analysis, Third Edition. Chichester: John Wiley & Sons.
(4) Niu, M. (1992). Composite Airframe Structures: Hong Kong Conmilit Press Ltd.
(5) Abaqus Users Manual, Version 6.12-1, Dassault Systémes Simulia Corp.,
Providence, RI.