Influence of the different vehicle subframe

configurations in the PDB assessment

Master’s Thesis in the Automotive Engineering International Master’s Program

AMIR REZA RIAZI

DARIUS SIMKUS

Department of Applied Mechanics

Division of Vehicle Safety
CHALMERS UNIVERSITY OF TECHNOLOGY
Göteborg, Sweden 2011
Master’s Thesis 2011:28
 MASTER’S THESIS 2011:28

Influence of the different vehicle subframe

configurations in the PDB assessment
Master’s Thesis in the Automotive Engineering International Master’s Program

AMIR REZA RIAZI

DARIUS SIMKUS

Department of Applied Mechanics

Division of Vehicle Safety
CHALMERS UNIVERSITY OF TECHNOLOGY
Göteborg, Sweden 2011
Influence of the different vehicle subframe configurations in the PDB assessment

Master’s Thesis in the Automotive Engineering International Master’s Program

AMIR REZA RIAZI
DARIUS SIMKUS
© AMIR REZA RIAZI, DARIUS SIMKUS, 2011

Master’s Thesis 2011:28

ISSN 1652-8557
Department of Applied Mechanics
Division of Vehicle Safety
Chalmers University of Technology
SE-412 96 Göteborg
Sweden
Telephone: + 46 (0)31-772 1000

Cover:
Crash simulation of the simplified Ford Taurus 2001 to the PDB barrier

Chalmers reproservice / Department of Applied Mechanics

Göteborg, Sweden 2011
Influence of the different vehicle subframe configurations in the PDB assessment
Master’s Thesis in the Automotive Engineering International Master’s Program
AMIR REZA RIAZI
DARIUS SIMKUS
Department of Applied Mechanics
Division of Vehicle Safety
Chalmers University of Technology

ABSTRACT
There is no consideration for crash compatibility of passenger vehicles in safety
regulations. The EU Project FIMCAR is investigating different frontal crash tests that
can assess a vehicle’s frontal crash performance for both self and partner protection.
Existing candidates need further development in establishing an objective
measurement from the test data. The PDB (Progressive Deformable Barrier) is one of
the candidates with ability to detect load distribution of a vehicle frontal structure in a
crash. Deformation of the PDB barrier surface is used to evaluate vehicle
performance. The proposed PDB barrier and evaluation process needs further
investigation before acceptance for vehicle regulatory or consumer testing.
The PDB’s ability to detect different front end structural configurations of a vehicle
was evaluated by simulations. A finite element model of the vehicle (2001 Ford
Taurus) and the PDB were used. The vehicle performance benefits of different sub-
frame configurations were indentified with car-to-car simulation results that were
used as the reference for car-to-PDB simulations. A new protection criterion of a
partner vehicle in crash was also developed using available PDB test results.
The performance differences of various sub-frame configurations were detected
through the car-to-car simulations. Initial car-to-PDB simulation results were not able
to detect these differences because of issues with components of the PDB. For this
reason, the PDB model was modified to improve lower structural interaction
detection. The modified PDB showed better results.
The results of the physical PDB and the FE PDB simulations were evaluated with the
developed criteria. The new PDB criteria worked well with scanned physical PDB
deformation faces, but not with the unmodified PDB simulation results. This
discrepancy may be due to the removal of covering structures in the vehicle FE model
that created a more aggressive car structure than a true production car.
The PDB was able to detect different sub-frame configurations. The car-car results
showed that the longer sub-frame configuration increases self protection because
structural interaction starts earlier in the crash. Due to the limitations of the PDB and
vehicle models used in the project, the proposed criteria did not properly assess the
different subframe configurations. Further work is needed to confirm the criterion is
robust to apply in frontal crash compatibility evaluation.
Key words: Compatibility, PDB, Self-protection, Structural Interaction, Fork Effect,
Finite Element

I
Contents
ABSTRACT I
CONTENTS II
PREFACE IV
NOTATIONS V

1  INTRODUCTION 1 
1.1  Background 1 
1.2  Compatibility 1 
1.3  PDB 3 
1.4  PDB measures 5 
1.5  Literature review 7 
1.6  Objective 8 
1.7  Limitations 8 

2  METHODOLOGY 10 
2.1  FE models 10 
2.1.1  Vehicle model 10 
2.1.2  Vehicle model modifications 10 
2.1.3  PDB model 12 
2.2  Simulation matrix 13 
2.3  Vehicle Performance measures 14 
2.3.1  Intrusion measurements 14 
2.3.2  Acceleration pulse 14 
2.3.3  Deformation modes 14 

3  RESULTS AND DISCUSSION 15 

3.1  Car to car 15 
3.2  Car to standard PDB model 18 
3.3  Car to modified PDB model 23 
3.4  Crash test comparison 28 

4  PDB CRITERIA 34 

5  SUMMARY AND CONCLUSIONS 42 

6  RECOMMENDATIONS 43 

7  REFERENCES 44 

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

II
APPENDICES 46 
Appendix A. PDB Model Validation 46 
Appendix B. MATLAB Codes 47 

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

III
Preface
This thesis project was carried out between January 2011 and June 2011 at SAFER
(Vehicle Safety Division in Chalmers University of Technology). The modifications
of the FE models were made by ANSA and LS-PrePost. The FE simulations were
made by LS-DYNA solver on the Chalmers super computer service BEDA. This
project is part of the project FIMCAR (Frontal Impact and Compatibility Assessment
Research) conducted by European Union.
We would like to show our high gratitude to our supervisor Dr Robert Thomson.
Thanks for all his support and guidance. Also we would like to thanks Linus
Wågström and Aleksandra Krusper for their technical support throughout all the
stages of the project. We finally want to thank Dr Johan Davidsson for his leading
advice and comments about our work and report.
Göteborg June 2011
Amir Reza Riazi
Darius Simkus

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

IV
Notations

EX Car model with extended sub-frame configuration

BA Car model with basic sub-frame configuration
SH Car model with shortened sub-frame configuration
WO Car model with without sub-frame cross beam
C2C Car to car crash scenario
C2PDB Car to PDB crash scenario
C2PDB1 Car to standard PDB crash scenario
C2PDB2 Car to modified PDB crash scenario
FWDB Full Width Deformable Barrier
FWRB Full Width Rigid Barrier
ODB Offset Deformable Barrier
PDB Progressive Deformable Barrier
FIMCAR Frontal Impact and Compatibility Assessment Research
NCAC National Crash Analysis Center
VC-COMPAT Improvement of vehicle crash compatibility through
the development of crash test procedures

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

V
1 Introduction
1.1 Background
Annually, a huge number of people are being killed on the roads of the European
Union. There were about 54300 deaths in 2001 which decreased by about 36 percent
until 2009 (based on the statistics from European Commission of Road Safety) [1].
This positive trend shows promising future for road traffic safety and more work is
needed to help improve the trend. In the future, active safety would play a significant
role in vehicle safety, however passive safety will remain important for a long period.
In the passive safety area of the vehicles, crash compatibility is one of the most
important parameters. Each vehicle manufacturer has its own strategy to improve
crash compatibility of their cars and there are no considerations in legislation.
Different ideas have been proposed for evaluation of crash compatibility.
The PDB (Progressive Deformable Barrier) concept was developed to represent
vehicle frontal structures similar to a partner vehicle as a tool to assess partner
protection. The stiffness of the barrier increases progressively in the longitudinal
direction and it has two stiffness configurations vertically. The barrier allows
detecting the front end structure and load paths of the vehicle. Also, the PDB
represents the vertical car force distribution, because the vehicles usually have
stronger lower front load path than the upper ones [2].
1.2 Compatibility
Structural interaction is achieved when energy absorbing structures in a vehicle are
efficiently deformed by structures in the collision partner. The best condition is for all
of the main frontal structures of a vehicle like longitudinals and sub-frame to contact
their pairs from the partner car in the impact. In this scenario, the main load paths
(Figure 1.1) of the colliding cars would absorb the impact energy. The load paths of a
vehicle, like the Ford Taurus, are the sub-frame, lower rails and upper rails in the
front. They are direct loads through the sills and A-pillars to transfer the crash load to
the compartment.
Accident analysis in FIMCAR identified fork effect and small overlap in frontal
collisions. If the impact happens with fork effect (horizontal misalignment of load
paths evolved during collision, Figure 1.2) or a small overlap, there is a risk that
strong parts interact with soft parts of the partner vehicle and cause big intrusion to
the compartment.

A-pillar Upper rails

Lower rails

Sill

Figure 1.1 Load paths of vehicle in frontal crash Sub-frame arm

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

1
 Load Paths

Figure 1.2 Fork effect

FIMCAR also identified problems when the vehicles are not aligned vertically. This
causes over/underriding. Similar compatibility issues occur where stiff structures
overcrush softer structures in the partner vehicle.
Crash compatibility of vehicles refers to how well structure of two vehicles
counterpart in frontal collision and amount of damage to the colliding vehicles. For
crash compatibility adjusted force level, structural interaction, optimized passenger
compartment strength are indispensable.
Mass is an important parameter in compatibility of the colliding vehicles, because
according to the law of conservation of momentum in an impact, the lighter vehicle in
front-to-front crash would experience bigger velocity changes than the heavier
vehicle. It means that the passenger of the lighter vehicle would experience higher
accelerations. The mass of the vehicles are not possible to be adjusted because of the
different masses in current vehicles on the street and the demand of the market for
different vehicle types.
The difference between the frontal structural stiffnesses would cause incompatibility
in the crash. The vehicle with a less stiff front structure would experience more front
deformation and the stiffer vehicle would have less deformation in the front structure.
Thus the softer vehicle is always more vulnerable to compartment intrusions and
passenger injuries. The global stiffness of front structures in a vehicle is highly
dependent to the structural interaction and distribution of the forces between load
paths of vehicles involved in a crash. If the load paths are interconnected with all
framing elements of the longitudinal and vertical force resisting vehicle front
structure, the contact forces can be well distributed in the vehicle. The ideal extension
of this property is for the loadpaths to be evenly, or homogeneously, distributed
throughout the vehicle.
As a more distributed deformation of a vehicle is considered as better car front
structure design, the assessment of homogenous vehicle deformation should be
studied. Thus to assess compatibility, a test method to evaluate homogenous vehicle

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

2
deformation should be developed. One promising candidate is the Progressive
Deformable Barrier (PDB).
1.3 PDB
The PDB barrier is shown in Figure 1.3. It is 797 mm deep, 1000 mm wide and 702
mm high. It is composed of the front deformable core, 250 mm, the progressive
deformable core, 450 mm and the back deformable core, 90 mm depth.

Figure 1.3 PDB barrier dimensions

The PDB is composed of three deformable cores, four plates, cladding, blind rivets
and epoxy resin as shown in Figure 1.4 and Figure 2.6

Figure 1.4 PDB components [3]

1 – Back plate,
2 – Back deformable core,
3 – Two intermediate plates,
4 – Progressive deformable core,
5 – Front deformable core,

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

3
6 – Contact plate,
7 – Outer cladding,
8 – Blind rivets,
9 – Epoxy resin.

The first 250 mm deep crushing strength area has a constant crush load. The second
450 mm deep progressive crushing strength area has a progressive crush load. The
third crushing strength area is progressive and has the same depth as the second one,
but is weaker. The fourth, back crushing strength area, is 90 mm deep and keeps
constant load, it is implemented to avoid bottoming out of the barrier [4]. These
crushing strength areas and strength specifications represent the strength of a vehicle
front structure and are shown in Figure 1.5

Figure 1.5 PDB cores and crush strength [4]

The PDB model should make equally severe crash test for light and heavy passenger
vehicles as shown in Figure 1.6. EES – energy equivalent speed is the energy
absorbed by PDB excluded from total vehicle kinetic energy before car to PDB
collision. This energy is recalculated to a speed and represents the energy amount
absorbed by the vehicle – the crash severity for the vehicle itself.

Figure 1.6 EES with different vehicle weight configurations [5]

One of the abilities of the PDB barrier is to detect structural interaction of the vehicle.
PDB structure are deformed by the frontal structures of the impacting vehicle, thus,

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

4
the PDB deformation analysis gives data about the vehicle structural architecture and
should thereby evaluate structural interaction.
1.4 PDB measures
PDB analysis was the most important part of the project. The PDB structure is
deformed during crash by a vehicle, the deformation amount and variation of the
barrier front surface is analyzed as the distribution of the barrier deformation is
affected by the front end structure of the vehicle.
The latest PDB assessment criteria was developed by FIMCAR. A front end of a
vehicle has three different stiffness areas vertically: from underneath to lower rails,
from lower rails to upper rails, above upper rails (Figure 1.1). For this reason, to
analyze the crushed PDB model, the front surface was divided to 3 vertical areas
(Lower, Middle and Upper) as described in the “Off-set Test Procedure” [5]
presentation at FIMCAR workshop January 19th 2011 (Figure 1.7). An example of a
PDB deformation pattern is shown in Figure 1.8.

Figure 1.7 Average position of lower and subframe issue from VC-COMPAT WP15
(left) and consideration of three vertical areas (Right)

Figure 1.8 A sample PDB deformation

The PDB deformation caused by a vehicle is scored by longitudinal deformation

criteria in these vertical areas, Figure 1.9.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

5
Figure 1.9 Longitudinal deformation criteria limits [5]
Deformation of the PDB was measured with different methods. Percentiles,
maximum, minimum, mean and most common value of the deformation measured
from the crushed PDB face were determined for analyzing the compatibility criteria.
The aim was to have the amount of deformation of the PDB in each area, independent
of the shape of the deformations and irregularities in the PDB crushed face.
A homogeneity criteria [5] is also proposed. Homogeneity is defined as the quality of
deformed surface being uniform. If the surface is smooth, it will be considered as
homogenous, if the surface is rough, it will be considered as inhomogeneous.
Homogeneity of the crushed PDB surface was always the point of interest, because
the aim of replacing the current ODB barrier to PDB was to detect the structural
interaction of the vehicle and PDB is supposed to detect it. There were proposals for
how evaluate the homogeneity of the PDB surface, but some are not used anymore
and there are controversies about some of them.
A candidate to assess homogeneity is using total variations which is a mathematical
concept used to distinguish between the homogenous and inhomogeneous surfaces.
Total variation shows the deformation variations of each cell to its neighbor. The sum
of all of the deformation variations in the PDB surface could evaluate the
homogeneity of the deformations. A surface could be called homogenous when it is
more similar to a flat surface than a wavy surface.

Figure 1.10 Visualization of the variations in a surface [6]

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

6
1.5 Some previous analysis have been done on the PDB
model by considering the whole surface of the PDB,
but that caused the mixing the interaction of the front
vehicle structures in the different vertical levels [7].
Literature review
A review of the subject of vehicle crash compatibility and PDB assessment has been
performed. The study was focused on exploring the research on physical tests and FE
simulations. Physical tests were more reliable than the simulations because they were
based on the real models and simulations had more variables and analysis data. There
were mainly five resources, EEVC WG15 (Enhanced European Vehicle-safety
Committee Working Group 15) [8], VC-COMPAT (Supported by European Union)
[3], ACEA-EUCAR [9], NHTSA (National Highway Transportation Safety
Administration) in US and DSCR (Directorate for Road Traffic and Safety) in France
[10] and Nissan Motor Company [11].
During several projects at Chalmers, the crash compatibility of the vehicles was
investigated with FE simulations with different crash scenarios. Avramov and Rachid
[12] performed extensive car to car crash simulations with different impact angles and
vertical and horizontal overlaps. Park, et al. [13] conducted simulations of car to car
and car to rigid barrier. Vehicle models with different sub-frame lengths were used to
evaluate the different vehicle front structure in the vehicle crash compatibility. Wu
and Chhim [14] modified and introduced extra load paths in the front structure of the
vehicle and performed crash simulation scenarios for evaluating the idea. The aim of
all of the above projects was to investigate the crash compatibility of the vehicles in
different configurations of the position and structure interaction which would cover
more real life accident situations.
PDB is a new barrier proposed by France to be used instead of the current ODB
barrier in Regulation No. 94 [4] as an update for frontal impact legislations. The
criteria for assessing the crash compatibility of cars described by Delannoy, et al.
[15].The EEVC WG15 members performed research project [16] to analyze and
improve the car crash compatibility and frontal impact. Two approaches for assessing
the crash compatibility were evaluated by the partners of the project, using FWDB
and ODB test together or FWRB and PDB test together. Several physical tests and FE
simulations were performed to evaluate the assessment approaches.
Physical crash tests been performed for evaluating the PDB abilities for assessing the
crash compatibility. Meyerson et al. [17] conducted several car-to-PDB and car-to-car
crash tests. Delannoy et al. [18] performed crash test comparisons of the ODB and the
PDB barriers in crash compatibility and self protection. The ability of barriers has
been discussed in comparison. They concluded that even with the lower generated
deceleration of the PDB barrier; the test procedure could represent the real world
accident because intrusion and acceleration is combined in the test. Tatsu et al. [11]
from Nissan motor of Japan, performed several crash test for comparing the ODB and
PDB barriers with different types of vehicles. The mentioned researches investigated
the performance of the PDB barrier with physical tests for crash compatibility
evaluation using PDB deformation, intrusion and dummy injury measures.
Park et al. [7] conducted FE simulation studies of the PDB with 5 NCAC vehicle
models. The proposed criterion for crash compatibility by Delannoy [15] was applied

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

7
to the results of the simulations to evaluate the performance of the criteria for
different types of vehicles. FE simulations were performed by the German industry
(VW) for criticizing the PDB [16] by making the main load path of the vehicle rigid.
The research showed that the PDB would absorb more energy if the vehicle’s front
structure becomes more rigid. The French industry (PSA) responded to the criticism
by simulation studies. They showed that it is not a proper strategy to use the energy
absorption capacity of the PDB because it would decrease the self protection and
increase the mass of the vehicle.
The literature review showed there were several positive results to improve and assess
the crash compatibility of cars and evaluation of the controversy methods. The
research used physical tests and FE simulations. The PDB barrier is a candidate for
revising the legislation. Based on the physical tests and simulations, using PDB in the
legislation would be beneficial if an appropriate and validated evaluation method
come in the legislation also. Lack of a proper criteria for assessing the crash
compatibility of cars with the help of the PDB barrier, led us to conduct FE simulation
of a vehicle with different sub-frame configurations and evaluate the results with a
new proposed crash compatibility criteria.
1.6 Objective
The objective of the project is to investigate the abilities of the PDB barrier for
distinguish front end structure performance of different vehicles through the crash
test. Also, the ability of Car to PDB results should be evaluated as a method to test
both self protection and partner protection in one test. For achieving this goal, car to
car and car to PDB crash simulation scenarios were performed. The car models had
different sub-frame configurations to reveal the sensitivity of the PDB to different
structure interactions. The comparison of measurements (acceleration pulse,
deformation modes, and firewall intrusions) between car-to-car and car-to-PDB crash
simulations would show how PDB could assess the crash compatibility.
A criterion for evaluation of the vehicle crash compatibility by using the PDB barrier
is needed to be applied on the results. It should show how the different structural
interactions could be examined by the PDB barrier with an organized procedure.
1.7 Limitations
The PDB FE model provided by FIMCAR was a simplified model. Solid elements
were used for modeling the honeycomb layers in the PDB model instead of the shell
type, for decreasing the simulation time.
The vehicle model used in the project did not have any physical crash test to the PDB
barrier. Thus, the validation of the crash simulations of the vehicle model to the PDB
model was not possible. Other physical PDB validation methods used instead, for
example the tubular impactor test which is explained in the Appendix A. PDB Model
Validation. The vehicle model was validated to the frontal physical US NCAP test
and it was mentioned that the model could be used for all impact scenarios.
The PDB performance and ability to detect structural interaction of the front
structures of the vehicle were investigated, but development and improvement of the
PDB was not within scope of the project.
The project investigation has been done for analyzing just the front structures of the
vehicle, because PDB is supposed to evaluate the interaction of the frontal structures.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

8
The crash dummies were not used because the vehicle to the PDB structural
interaction and the car body performance were more the focus and using the dummy
in the crash simulation would add to the complexity.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

9
2 Methodology
The methodology of the project is described in this chapter. The FE model of the
vehicle and PDB are also explained.
2.1 FE models
There were two types of models used in this project, finite element models of vehicle
and progressive deformable barrier (PDB). FE models are described in this section.
2.1.1 Vehicle model
A Ford Taurus 2001 FE model was selected, because it can represent a European mid-
size vehicle and the model is available on the web site of the National Crash Analysis
Center (NCAC). The model contains 778 parts and is divided to 1.057 million
elements; it is shown in Figure 2.1 [6].

Figure 2.1 Ford Taurus FE model

The NCAC model was changed to a simplified version developed at Chalmers
University of Technology (by Rachid, Ebbinger, Park, Avramov, Krusper) Figure 2.4.
Less time was needed for running crash simulation with the simplified Ford Taurus.
2.1.2 Vehicle model modifications
An extra beam was introduced into the firewall of the simplified Ford Taurus model
(Figure 2.2). This modification stiffened vehicle’s crash response, because the motion
of the longitudinals is restricted upwards and force is transferred to the A-pillars and
sills instead of the firewall.

Figure 2.2 Extra beam in the firewall of the Taurus model [13]

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

10
The crash pulse of the simplified model represents US-NCAP crash pulse of the
modified vehicle models are shown in Figure 2.3. The figure shows the corridor of the
acceleration pulse of midsize vehicles in US-NCAP test (56 km/h to the rigid barrier).
Taurus_Henrik and Taurus_RA were two other modifications of the Taurus model
during previous research at Chalmers and Taurus_CK is the Taurus model that has
been used in the current research.

Figure 2.3 US-NCAP crash pulses

The low priority parts like front bumper cover and hood were excluded from the front
end of the model in the simplified model thus this model has more exposed frontal
structures for interacting with PDB model. All components from the rear end to the B-
Pillars were made rigid. The simplified Ford Taurus is shown in the Figure 2.4.

Figure 2.4 Simplified Chalmers Ford Taurus

The weight of the simplified vehicle is around 1390 kg. A failure criterion for the sub-
frame mountings to the compartment floors was defined. The sub-frame would be
released when load peak reaches 50 kN [13].
Sub-frame Configurations
The Ford Taurus sub-frame carries the engine and lower frontal suspension mounting
points as shown in Figure 2.5 (1). Four different sub-frame configurations were used:
shortened (2), basic (3), extended (4) and without sub-frame (5) as shown in Figure

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

11
 1 2

5 4

Sub-frame arm

Figure 2.5 Different Taurus sub-frame configurations

2.1.3 PDB model

In this section the PDB model is described. This model is used for car to barrier crash
tests simulations. The FE model of PDB is shown in Figure 2.6 was developed by
General Motors Europe and released on February 2011. Solid elements used instead
of shell elements for defining the honeycomb cores in this PDB model.
The MAT_MODIFIED_HONEYCOMB material model is used for aluminum
crushable honeycomb foam material with anisotropic behavior [19]. Thus aluminum
honeycomb cores are modeled as solid elements with aluminum honeycomb solid
material.
CONTACT_AUTOMATIC_SINGLE_SURFACE_ID card, the interaction,
recommended by FE of PDB developer, between disjoint parts, is used during
simulations.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

12
Figure 2.6 The FE model of the PDB

2.2 Simulation matrix

Dynamic crash analysis is necessary for better understanding of the crash
compatibility of the vehicles. In order to investigate the effect of the different sub-
frame configurations in crash compatibility and ability of the PDB barrier for
representing that effect, a simulation test matrix arranged in Table 2.1. The bullet
vehicle which is the vehicle model with 4 sub-frame configurations would crash to the
target which is either the PDB model or the vehicle with the basic sub-frame
configuration.
Car to PDB simulations have been conducted with 60 km/h and 50% horizontal
overlap. This is the speed and overlap required by the PDB protocol by Delannoy [15]
Car to car tests have been done with the speed of 56 km/hr which is from the accident
study by PENDANT [16] that showed that 85 % of all types of injuries was with the
velocity of 56 km/hr. The horizontal overlap was 50% and is based on the car to PDB
crash test.

Table 2.1 Simulation Matrix

Bullet  Target  Velocity(km/hr)  Abbreviated Simulation Name 
SH  PDB  60  C2PDB1_SH 
BA  PDB  60  C2PDB1_BA 
EX  PDB  60  C2PDB1_EX 
WO  PDB  60  C2PDB1_WO 
SH  Basic  56  C2C_SH 
BA  Basic  56  C2C_BA 
EX  Basic  56  C2C_EX 
WO  Basic  56  C2C_WO 

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

13
2.3 Vehicle Performance measures
In this section, the measurements that have been used for the analysis in the project
are described.
2.3.1 Intrusion measurements
The amount and shape of intrusions to the firewall and driver’s footwell was
measured for investigating self protection. The measurement points are shown in the
Figure 2.7. The intrusions have been measured related to the local coordinates in the
backseat of the vehicle model. Because the rear part of the vehicle was rigid and
mounting the accelerometer far from the front, in the rigid area, allowed having the
crash pulse of the occupant compartment recording during crash. This method of
measurement allows for monitoring the progress of intrusion through the whole crash
time.

Figure 2.7 Measurement intrusion points at firewall [20]

2.3.2 Acceleration pulse
The acceleration pulse of the crash was measured from the accelerometers on both
sides of the backseat. The average acceleration between the two accelerometers used
for analysis. The average acceleration multiplied by the mass of the vehicle model
was used as the crash force. This acceleration pulse was useful for comparison of car
to PDB with car to car and analyzing the behavior of PDB.
2.3.3 Deformation modes
The behavior of deformation of the front structure and other structures of the vehicle
model during different crashes (shown in Table 2.1) was analyzed. This analysis of
the deformed bumper beams shape helped to assess the ability of the PDB to represent
the interaction of another car in the crash.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

14
3 Results and discussion
In this chapter, the results of the crash simulations are presented and discussed. Car-
to-car simulations are described in the beginning and Car-to-PDB simulations
afterwards. Some issues about the PDB model would be described and some
modifications to the PDB model in order to fix the issues are proposed and simulated.
3.1 Car to car
The investigation of the C2C crash simulations shows different front structure
interactions. These interactions are presented in Figure 3.1. The EX model (Taurus
with extended sub-frame) sub-frame structure starts to interact earlier during the
collision, for this reason, the deformation is higher for the sub-frame arm. The BA
(Taurus with basic sub-frame) and SH (Taurus with shortened sub-frame) versions
have less sub-frame arm deformation, but as could be seen in Figure 3.20 the bumper
beam is deformed more in these two cases in comparison to EX version. WO model
(Taurus without a sub-frame) does not have sub-frame interactions, thus the bumper
beam deformation was the biggest.
Sub-frame arm Sub-frame

Bumper beam

Figure 3.1 C2C crash simulations with different sub-frame interactions (the right
vehicles have BA sub-frame configuration)

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

15
Figure 3.2 shows the deformation of the bumper beam and longitudinals. The bumper
beam in the EX model has the smallest fork effect amount and in the WO model the
biggest fork effect amount could be noticed because amount of intrusion of the
longitudinal is higher into the bumper beam of the target vehicle. Because the lower
load path was extended in the EX model and was loaded earlier in the crash, more
energy was absorbed by the sub-frame parts and less energy absorbed by lower rails.
For this reason, the EX had the smallest fork effect.
Lower rail

Figure 3.2 C2C crash simulations with different sub-frame interactions and
hidden sub-frame (the right vehicles have BA sub-frame configuration)
Figure 3.3 shows that the amount of displacement of the vehicle is related to the
length of the sub-frame. In the EX model we have the lowest amount of displacement
and when the length of sub-frame is decreasing, the displacement amount of the
vehicle would increase and we have the maximum displacement in the WO model.
Also, we could notice that the sub-frame interaction starts between about 400-600 mm
of the vehicle displacement in the Figure 3.3, marked by the circle. We can notice that
the latest structure interaction is in the WO vehicle at around 560 mm. Increasing the
length of sub-frame led earlier sub-frame interactions. The EX sub-frame started
interacting with the other car around 100 mm earlier than SH.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

16
Figure 3.3 Acceleration comparison between car-to-car crash simulations with
different subframe configurations
Figure 3.4 shows the same trend as in Figure 3.3 for the bullet cars (cars with varying
sub-frame configurations). The increasing length of the sub-frame caused lower
intrusions to the firewall. WO had the highest intrusions, especially in the upper area.
The longer sub-frame had a better self protection. The target vehicle in crash to the
bullet vehicle with WO configuration showed the best self-protection (Figure 3.4 right
side).

Figure 3.4 C2C firewall intrusions to the bullet cars (left) and target cars (right)
The average firewall intrusions of the target and bullet vehicles are shown in Table
3.1. The longest sub-frame configuration had the lowest average intrusion in bullet,
target and total vehicles firewalls during crash. The WO had the biggest average
firewall intrusion in the bullet vehicle as well the average difference between bullet
and target vehicles.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

17
Table 3.1 Average firewall intrusion

Sub- Average Average target Total firewall Difference in firewall

frame bullet firewall firewall intrusion intrusion
intrusion, mm intrusion, mm (bullet + target), [bullet – target], mm
mm

BA 202 179 380 23

SH 183 177 360 6

EX 134 161 294 27

WO 230 122 353 108

Discussion
The crash simulations of the C2C were evaluated and described in this section. The
deformation modes, acceleration pulses and amount of the firewall intrusion have
been compared. The results showed that the longer sub-frame had less fork effect
(Figure 3.1 and Figure 3.2), as well as overall vehicle displacement (Figure 3.3) and
firewall intrusion (Figure 3.4). Furthermore, the structural interaction of the vehicles
started for smaller displacements (Figure 3.3). This trend should be compared with the
car to PDB crash test results. The SH, BA and EX had low average difference in
firewall intrusion, the WO had the biggest average difference in firewall intrusion
between the bullet and target vehicle, so it was the least desirable configuration. The
EX configuration had the lowest average total firewall intrusion. The EX case had the
best performance in self, partner and total average firewall intrusion performance.
3.2 Car to standard PDB model
Car to standard PDB crash test simulations were conducted with four different sub-
frame configurations. Figure 3.5 shows the front structure interactions with PDB from
bottom view. In the upper row the pictures represent sub-frame interaction (outer
cladding sheet is removed from view for better visualization of honeycomb layers
deformation). The lower row of pictures shows the same bumper cross beam
deformation (PDB is removed from view for better visualization).
In all cases the sub-frame arm had interactions with the PDB face, while the sub-
frame crossbeam did not interact with the front honeycomb layer and slid under it.
The sub-frame then started to interact with the middle honeycomb layer. For this
reason, the sub-frame interaction area in PDB was unreasonable. The bumper cross
beam had a similar amount and shape of deformation; we can see this trend in Figure
3.20, Section 3.4. This trend occurred because the sub-frame crossbeam did not
interact with PDB. Thus the different sub-frame configurations did not influence the
crossbeam deformation and the bumper cross beam was loaded by similar loads
through the collisions.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

18
Figure 3.5 Front structure deformation during collision with standard PDB
Figure 3.6 shows deformation of the PDB with four different vehicle configurations.
The investigation areas of the PDB are shown by the black lines (as described in the
section 1.4). The lower rails and front bumper beam had interaction with the middle
area of standard PDB, while the sub-frame structure interacted with the lower area of
the PDB.
The sub-frame cross beam of the EX model slid under the front and middle
honeycomb layer. As a result of the sub-frame sliding under the barrier, the lower area
of PDB was deformed upwards, but not in the longitudinal direction. The effect of the
sub-frame interaction could not be noticed in the lower area of the PDB deformation
(Figure 3.5 and Figure 3.6 EX). The WO model did not have the interaction with the
lower are of PDB, because of the absence of the sub-frame cross beam. The BA and
EX sub-frame had a small interaction with the middle honeycomb layer which could
be noticed in the Figure 3.5 and Figure 3.6.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

19
 EX BA

SH WO

Figure 3.6 Standard PDB deformation comparison

Figure 3.8 shows the 99th percentile of the deformation of the three different areas in
standard PDB. The amount of 99th percentile for the middle area was about the same
for the different sub-frame configurations. The 99th percentile increased for the lower
area and decreased for upper area of the PDB by increasing the length of the sub-
frame, because the longer sub-frame absorbed more energy and caused more
deformation in the lower area. The lower area of the EX model had the same amount
of 99th percentile as the WO model and that was because of the insufficient interaction
of the sub-frame cross beam with the lower area of the PDB. The 99th percentile was
smaller for the longer sub-frame configuration in the upper area, because the longer
sub-frame arm prevented lower rails to be pulled up, while the WO model did not
have the stabilizing effect of the sub-frame cross beam (Figure 3.7).

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

20
Figure 3.7 Deformation comparison of the lower rails (the left view-WO, the right-EX
sub-frame configuration)

Figure 3.8 99th percentile deformation comparison of PDB areas

Figure 3.9 shows the crash acceleration for the different vehicle configurations versus
displacement. The elongation of the sub-frame caused smaller overall displacement of
the vehicle. The trend, that the EX configuration had the smallest and the WO the
biggest overall displacement, is noticeable. This trend for the car to PDB crash test
(Figure 3.9) was similar to the C2C cases (Figure 3.3).
The sub-frame interaction area in the beginning of the crash (from about 400 mm of
vehicle displacement, in Figure 3.3 and Figure 3.9 marked by a circle) was not as
distinguished as the C2C cases, but the overall trend was the same through the
different sub-frame configurations. The longer sub-frame had earlier interaction with
PDB.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

21
Figure 3.9 Acceleration comparison between of the car to standard PDB model
crash simulations with different sub-frame configurations
Figure 3.10 represents the firewall intrusion of the vehicles with different sub-frame
configurations after the collision to the PDB. The overall amount of the firewall
intrusion is smaller compared to the C2C cases (Figure 3.4). The C2C cases show the
clear trend of the intrusions by increasing the length of sub-frames, while the car to
PDB impacts do not show the distinguished amount of the firewall intrusions.

Figure 3.10 Car to PDB firewall intrusions

Discussion
In this section crash tests of the car to PDB were described. The PDB model had two
issues. First, the interaction of the PDB and sub-frame parts was not proper. In the
beginning of the crash the lower edge of the PDB was lifted up by the outer cladding
sheet. For this reason, the sub-frame did not interact with the front honeycomb and
slid under it. Also, the SH sub-frame configuration had the better interaction with the

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

22
middle honeycomb layers than the EX model, because the lower honeycomb layers of
the SH model were raised up less than in the EX case.
The thick outer cladding sheet and the contact plate prevented detection of the vehicle
structural interactions properly. One of the purposes of using the PDB instead of the
ODB barrier is to detect structural interaction. The covering surfaces of the current
PDB distribute the forces over the honeycomb and local deformations are prevented.
As a result, a few modifications will be applied to the current PDB model. These
modifications would be trials to show, how the above described issues could be
solved.
The firewall intrusions of the C2PDB were not similar to neither bullet neither target
C2C case, because of the different structures of the colliding objects. The firewall
intrusions of the C2C_WO (target vehicle) and C2PDB cases were similar, because
the sub-frame was not loaded locally during these collisions cases.

3.3 Car to modified PDB model

The issues of the PDB model were described in the end of the previous chapter. A few
modifications were applied to the current PDB model for trying to solve the described
issues.
Modifications
1. Removing the outer cladding sheet
The thick outer cladding sheet was removed. The lower edge of the outer cladding
sheet was pulling up the front honeycomb layer (Figure 3.11). For this reason it was
preventing the proper interaction of the sub-frame cross beam and the front
honeycomb layer. The sub-frame cross beam was sliding under the front honeycomb
layer. The outer cladding sheet was removed to prevent the pulling up effect of the
honeycomb.

Figure 3.11 The outer cladding sheet pulling up effect

2. Modifying the contact plate
The contact plate and the outer cladding sheet were collaborating to distribute the
forces of the impacting structures; the contact plate was useful for scanning the PDB
deformation. For this reason, the mechanical properties of the contact plate were
changed to represent a neutral plate. The Young’s modulus and yield strength of the

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

23
contact plate were decreased by 1000 times. As a result, the contact plate had no
influence on the deformation behavior and just was used for scanning the PDB
deformation.
3. Modifying the epoxy resin
Figure 3.12 shows the insufficient strength of epoxy resin. For this reason, the
mechanical properties of the epoxy resin used between the plates and the honeycombs
in the PDB model were investigated. Epoxy resin H9440 is recommended by the
French proposal of PDB [4] and based on the datasheet from the Axson Company, the
epoxy resin has 21 MPa of shear strength and 30 MPa of tensile strength. But in some
areas of the PDB model and specially in the front layers the strength of 8 kPa was
used for the epoxy resin (between contact plate and front deformable core and
between intermediate plate and front deformable core), therefore the described epoxy
strength used in the PDB model was changed from 8 kPa to 21 MPa.

Figure 3.12 Epoxy resin failure

Results of the modified PDB model
Figure 3.13 represents the front structure interaction of the Taurus with different
configurations to the modified PDB. The interaction of the sub-frame could be noticed
in the bottom view shown in the upper row of the picture and interactions of the lower
rails of the vehicle are presented in the lower row of the picture (PDB model is
removed from view for better visualization).
The sub-frame cross beam interacted as expected with the front deformable core in all
crash test configurations. The sub-frame cross beam did not slide under the PDB. The
interaction of the sub-frame cross beam started from the front deformable core and not
from the middle deformable core like the C2PDB1 cases (Figure 3.5).
Generally, the bumper cross beam had more significant fork effect than in the
C2PDB1 cases.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

24
Figure 3.13 Front structure deformation during collision with modified PDB
Figure 3.14 shows the deformation of the modified PDB during the crash with
different vehicle configurations. A comparison to Figure 3.6 shows that the modified
PDB is able to detect more local structural interaction. The interaction of the sub-
frame cross beam was better seen in the lower area of the PDB deformation. In the
C2PDB1 and in the EX model (Figure 3.6), this interaction is not noticeable in the
lower area and this problem is solved in the modified version of the PDB model.

Figure 3.14 Modified PDB deformation comparison

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

25
Figure 3.15 shows the amount of 99th percentile of deformation in PDB front surfaces.
The deformation trend (longer sub-frame cause bigger deformation in the lower area
and smaller deformation in the upper area) for the lower and the upper areas was
similar to C2PDB1 cases (Figure 3.8). The EX model had around 370 mm in the
standard PDB and around 510 mm of 99th percentile in the lower area of the modified
PDB. The 99th percentile deformation increased in the lower area, because of the
sufficient sub-frame and PDB interaction. As a result the modified PDB is able to
detect the interaction of different sub-frame lengths in the lower area.
The 99th percentile deformation increased in the middle area of the C2PDB2 EX case,
because the EX lower rails were stiffened by the extended sub-frame arms and more
local deformation was caused (Figure 3.14, marked by the circle). While in the
C2PDB1 EX case local loads were distributed (Figure 3.6) by the cladding sheet and
contact plate and stiffer structure was not detected by the standard PDB.
The 99th percentile deformation increased by around 50 mm in the EX upper area
compared to BA, because the modified PDB was less stable without the cladding
sheet and upper outboard corner of the barrier moved down around 60 mm more in
the EX than BA case. Thus, the area, marked by the square, (Figure 3.14 BA and EX)
partly moved down to the middle area and 99th percentile decreased in the upper area
(Figure 3.15).

Figure 3.15 99th percentile deformation comparison of modified PDB areas

Figure 3.16 shows the acceleration pulse of the crash simulation of different vehicle
configurations to the modified PDB. In this picture it could be noticed that the
interaction area of the sub-frame (from about displacement of 400 mm) has more
distinguished shape for the different sub-frame configurations. Generally, the vehicle
with the longer sub-frame had higher acceleration at the beginning of the sub-frame
interaction (from 400 mm displacement). The trend of the overall displacement was
similar to the C2PDB1 (Figure 3.9) case, but more distinguished. However, the WO
and SH models had the same overall displacement. The C2PDB2 acceleration trend
was more distinguished and similar to the C2C (Figure 3.3) case than C2PDB1.The
acceleration increases slower through C2PDB collisions compared to C2C, because
the structure of PDB is more homogenous than the front of the vehicle.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

26
Figure 3.16 Acceleration comparison of the car to modified PDB model crash
simulations with different subframe configurations
Figure 3.17 represents the firewall intrusions of the vehicle with different sub-frame
configurations. The C2PDB2 firewall intrusions were more distinguished compared
with the C2PDB1 case. Thus, the SH and WO sub-frame configuration were
penalized more than the EX and BA models in the upper area. Though, the firewall
intrusion does not have any specific trend in the lower area. Generally the overall
amount of firewall intrusion increased in the C2PDB2 in comparison to C2PDB1
crash tests.

Figure 3.17 C2PDB2 firewall intrusions

Discussion
In the C2PDB1simulations the PDB had issues, described in the discussion of the
section 3.2, and for this reason the FE PDB model was modified. The modifications of
the PDB showed promising results. The sub-frame interaction issue was solved by

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

27
Generally, the amounts of the firewall intrusions of the C2PDB2 were bigger than the
C2PDB1 and more similar to C2C cases. The trend between the different sub-frame
configurations were not clear because the difference between structure interaction in
the C2C and C2PDB. In the C2C cases, the front structures of the vehicles interacted
locally with each other, while in the C2PDB cases, the vehicle front structure
interacted to the homogenous honeycomb layers.
The acceleration pulses in C2PDB2 cases were generally similar to the C2PDB1, but
the differences between the acceleration in the beginning of the crash for different
sub-frame configurations were bigger for the C2PDB2 cases.

3.4 Crash test comparison

In this section the PDB ability to detect different vehicles structural interaction for
crash compatibility evaluation will be compared.
PDB deformation comparison
Figure 3.18 shows the vehicle structural interaction to the PDB and modified PDB.
The local structural interaction is more amplified in the modified PDB than standard
PDB face. The lower structures were not detected by the standard PDB, because the
front deformable core was pulled up by the outer cladding sheet. Thus, the modified
PDB represented more data about the front end structure of the vehicle, for this reason
the protection of the partner vehicle could be estimated better.

Figure 3.18 C2PDB1 and C2PDB2 deformation comparison (from the left:
C2PDB1, C2PDB1 without outer cladding in the view, C2PDB2)
Figure 3.19 shows that the issue of interaction of the sub-frame cross-beam is solved
in the modified PDB. The lower honeycomb layers are deformed by the sub-frame
cross beam. Consequently the lower structures of the vehicle were detected properly
by the modified PDB model.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

28
Figure 3.19 Sub-frame cross beam interaction to PDB comparison (from the left:
C2PDB1, C2PDB2)
Figure 3.20 shows the shape of the front bumper beam after crash. The shape of the
bumper beam deformation for all of the vehicle configurations in each case is similar.
It could be noticed that the C2PDB2 cases have more similar fork effect to the C2C
cases.

Figure 3.20 Deformation shape of the front bumper beam (from the left: C2C,
C2PDB1, C2PDB2)
Acceleration pulse
The following 4 figures (Figure 3.21, Figure 3.22, Figure 3.23, Figure 3.24) show the
acceleration pulse for each sub-frame configuration in each crash test scenario. The
C2PDB have similar shapes of acceleration pulse during the crash and would start
without any specific peak in the middle of the crash and would continue to the end of
the crash to the maximum acceleration point by a constant slope.
The car to car cases with EX and BA models are starting with a high peak in the sub-
frame interaction area (from about 35ms). The car to car cases with SH and WO
models are starting to have higher acceleration later than EX and BA models (but
without any specific acceleration peak).

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

29
Figure 3.21 Acceleration comparison between car-to-car crash simulations and car
to different PDB model configurations (Car with extended sub-frame)

Figure 3.22 Acceleration comparison between car-to-car crash simulation and car
to different PDB model configurations (Car with basic subframe)
The slope of acceleration in the C2C_SH and C2C_WO is more similar to the C2PDB
cases, because of the less significant structural interaction.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

30
Figure 3.23 Acceleration comparison between car-to-car crash simulation and car
to different PDB model configurations (Car with shortened subframe)

Figure 3.24 Acceleration comparison between car-to-car crash simulation and car
to different PDB model configurations (Car without subframe)
The first acceleration peak in C2C cases is starting with sub-frame interactions
(around 35ms) and the second peak is the interaction of wheels to the sub-frame of the
partner vehicle (around 55ms). The sub-frame interaction acceleration peak is higher
for EX and BA (around 35 ms) because they had longer sub-frames and the peak was
lower for SH and WO (around 40 ms) because of less engagement of their sub-frame
parts for absorbing energy. The acceleration peak in the C2PDB cases is observed

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

31
around 90ms which is in the final moments of the crash and that means the overall
collapsing and buckling of the crash energy absorbing structures are completed.
Energy calculation comparison
Figure 3.25 and Figure 3.26 show the amount of the absorbed energy by PDB. Three
methods were used to extract this data: the PDB software developed by UTAC and
Universite de Provence, the MATLAB script, provided by VTI and LS-PrePost. The
MATLAB script and the PDB software estimates similar amount of absorbed energy.
The estimation is based on the deformation of the PDB front plate scanned after
deformation. The LS-PrePost data is based on the change of the PDB internal energy,
thus the barrier volume and shape after a deformation. In this case the EES was not
consistent with the other software. The main reason was the negative sliding energy
caused by negative volume of solid elements. This energy was 3% of the total energy
during the SH with C2PDB1 simulation. During the C2PDB2 simulations, the
negative sliding energy varied between 5-8 %, except during the EX case – 49 %. For
this reason, the EES trend based on the C2PDB2 LS-PrePost data is not reliable.
In Figure 3.25 the C2PDB1 data shows that the vehicle with longer sub-frame
absorbed more energy, except the EX case. In this case the PDB was deformed from
underneath, thus the MATLAB script and the PDB software estimated bigger energy
absorption by the PDB, but the LS-PrePost (except SH) to C2PDB1 as well as
C2PDB2 simulations data show more reasonable trend, because the PDB front was
not lifted up from underneath during C2PDB2 simulations.

Figure 3.25 EES calculation by the different software for the C2PDB1 simulations
The C2PDB2 data shows that the more clear relation between sub-frame length and
absorbed energy by the vehicle, Figure 3.26. The EES estimation by LS-PrePost failed
due to the described sliding energy issue.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

32
Figure 3.26 EES calculation by the different software for the C2PDB2 simulations
The data shows that longer sub-frames absorb more energy, the difference between
WO and EX version was up to 8% of total energy. Hence, the longer sub-frame
caused better structural interaction and more energy absorption.
The LS-Dyna absorbed energy data estimation for C2PDB2 EX case was affected by
negative volume of deformed solid elements. This energy issue did not influence the
deformation and acceleration data, because the C2PDB2 results were inline with the
C2PDB1results in terms of trends in the acceleration and deformation values. Thus,
we estimate that the sliding energy was affected only.
Discussion
In this section the results of all crash tests were compared. The modified PDB had
better interactions with the lower vehicle structures and was more sensitive for the
local loads, because of the removed cladding sheet and weakened contact plate. In the
C2PDB2 more fork effect was observed in bumper cross beam in comparison to the
C2PDB1 and was more similar to C2C cases in general shape of the bumper beam.
The EES estimation of C2PDB2 by LS-PrePost failed due to FE model behavior, but
the deformation amount of the EX vehicle and PDB trend was as expected (based on
C2C and C2PDB1 cases). Hence, the sliding energy became negative and increased
total energy amount by 48%, but did not affect the deformation of the vehicle and the
PDB. After overall analysis we could conclude that the more energy was absorbed by
the PDB through the vehicle with longer sub-frame configuration collision to the
barrier cases than through the vehicle with shorter sub-frame collision to PDB cases.
The PDB was not able to act like another vehicle and have structural interaction.
Besides, the PDB detected the pressure of the vehicle structure through the collisions,
thus the deformation amount of PDB was related to the interaction area and the load
(loads are related to stiffness of the vehicle structure), while the deformation amount
was related to the structure stiffness through the C2C cases. It absorbed the crash
energy in a controlled way.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

33
4 PDB criteria
The PDB barrier is candidate for assessing the crash compatibility of the vehicles in
comparison to the current ODB barrier. Evaluation method of crash compatibility
with the help of the PDB is very critical for the condition that the PDB would be used
for the regulation of frontal collision
A crash compatibility criterion has been developed during this project. The
assessment criterion was tested on the results of crash test of 36 vehicles to the PDB
barrier, the crash data obtained from FIMCAR project and a few of them are
published in the website of the FIMCAR. This method of crash compatibility
evaluation is described in this chapter

Analysis of the different areas

The different areas of the PDB were described in Section 1.4. Figure 4.1 shows the
PDB deformation of a sample vehicle. The middle area of the PDB shows a large hole
that has more than 600 mm of deformation and means that the vehicle has very strong
lower rails and had improper structural interaction with the PDB barrier in the middle
area. From this picture it could be concluded that the partner vehicle in the front to
front impact to this vehicle would have big intrusion in the same area, if the partner
car does not have a strong structure in the same position. This conclusion is true only
if all of the impact partners to this vehicle have homogenous front structure like the
whole structure of the PDB barrier. However the structural interaction of the front to
front impact of the vehicles is complicated and is specific to each case, the PDB
barrier has promising abilities to evaluate it to some extent.

Figure 4.1 PDB deformation of a sample vehicle

The assessment of crash compatibility of the vehicles is wide and complicated, this
assessment has been implemented by using a few parameters and a simple procedure.
Assessment introduction
The middle area of the PDB is considered as the main core of the evaluation. The
interaction of the lower rails, which are handling the biggest energy absorption of the
impact, because those are the main load paths in the front structures of the vehicles, of
the vehicles would be detected in the middle area of the PDB. The evaluation of this
area would build the main score of the vehicle in the criteria. The lower and upper
areas of the PDB would be evaluated and in the case of improper structural
interactions, the main score would be penalized.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

34
The homogeneity of the deformation is important for just the lower and middle areas
of the PDB because the main impact load carrying structures of the vehicles like
lower rails and sub-frames would interact with those areas. The only important
parameter for the upper area of the PDB is the amount of the deformation and not
homogeneity of the deformation. If a structure of any vehicle interacts to the upper
area of the PDB and make big deformations to that area, the vehicle would have
higher risks of overriding the partner vehicle in the front to front impact (because
there would be a risk of having lower rails in the high height), and direct interaction to
upper part of the passenger body in the partner vehicle in side impact. Thus, the
amount of deformation is crucial in the upper area of the PDB even with a
homogenous surface of deformation.
The lower and middle areas of the PDB have been evaluated with the calculation of
total variations of the PDB deformations which is representing the homogeneity of the
deformed surface of the PDB and the amount of total deformation in each area. For
balancing the effect of the main front structures, the relation between the amount of
deformation in the lower and middle area of the PDB is considered.
Further explanation of special parameters for evaluating the PDB deformations is
introduced and the assessment based on them as found in the following section.
Percentiles of the deformations
Percentiles of the deformations are used as a statistical value for checking different
amounts of deformation in the PDB face. It is the statistical percentage of deformation
is data that below its value. For example, if the value of the 50th percentile of the
deformations is 300mm, it means that, 50 percent of deformations are smaller than
300mm.
Reference depth of deformation
Detecting the holes in the PDB deformation and comparing the amount of the
deformation of each area to the other area of the PDB needs a reference depth of PDB
deformation. One method to find a proper reference depth is to use the most common
range of deformation by calculating the distribution of deformation in each depth.
Figure 4.2 shows the distribution of the grid cells of the PDB middle area for the
vehicle from the Figure 4.1. Different ranges of the depth of deformation have been
tested in order to find the one that could represent the most common depth of
deformation and not the holes and other improper features of the PDB deformation.
This range of depth is a moving window that is assessed over the data in Figure 4.2.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

35
Figure 4.2 Distribution of cells through PDB depth for a sample vehicle
The best range of depth was 240 mm and it means for which 240 mm window of
depth the maximum distribution of the PDB deformation occurred. The middle depth
of the most common depth of range for the middle area of the sample PDB described
above was 188 mm (obtained by analyzing the PDB). Figure 4.3 shows the most
common range of depth for the middle PDB area of the sample vehicle. The range is
starting from 68 until 308 mm of deformation. The depth that is indicating the middle
depth of the mentioned range of depth is called the reference depth in this project. It
could be noticed that the selected area didn’t include the hole.

Figure 4.3 The most common range of depth for a sample car
Assessment procedure
The first step in the assessment is to evaluate the middle area of the PDB deformation
for the main score and subtract the results of the other areas of deformation from this
core score. The idea for evaluation of the middle area is to combine the total variation,

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

36
99th percentile of deformation and the reference depth for the middle area. Formula
3.1 is showing how the evaluation of the middle area has been done.
(3.1)
,

:
:
:
:
5: 5
99: 99

The term "max , 99 " selects the maximum value between the
deviations of biggest deformations from the reference depth and the reference depth
itself and is sensitive to the holes in the surfaces. The reason for multiplying this term
to “TV” is that “TV” increases with the number of holes and is not sensitive enough
to small smooth holes. The terms containing the reference depth amplify the effect of
the narrow deep holes. The term " 5 " gives the differences between the
reference depth and the small deformations and is sensitive to small overlap of the
structures in the area. The score was increased to an objective value by weight factor,
which is 1×106. The theoretical scale of the middle area 0; ∞ , where
better performing vehicles have higher scores.
Figure 4.4 shows the PDB deformation of 8 sample vehicles from the FIMCAR and
the results of the calculation of the formula above are represented in Figure 4.5. It
could be noticed that the PDB faces with holes in the middle area received lower
scores than the PDBs with homogenous deformation in the middle area.

Figure 4.4 Deformation PDB samples from FIMCAR

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

37
Figure 4.5 Score of the middle area of the PDB for the sample vehicles
The balance between the amounts of deformation of the lower and the upper area of
the PDB is important for the assessment because the big difference should increase
the risk of the structural interaction misalignment and under/overriding during the
front to front impact. For investigating the mentioned balance, the relation between
the reference depth of the middle and lower area is helpful. Figure 4.6 shows the ratio
between the reference depth of the middle and lower area for the above eight PDB
samples. It could be noticed that the models with ratios closer to one, have more
similar amount of deformation for the lower and upper areas (without considering the
holes).

Figure 4.6 Ratio between the reference depth of the middle and lower area
If the ratios bigger than 1.5 between middle and lower area, the “TV” of the lower
area are subtracted from the core score, the final score of the lower and middle area
would be the result (3.2):

(3.2)

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

38
The theoretical scale of the lower and middle area score is difficult to define. Based
on physical PDB data variation, expected values for this scoring system are
. The maximum score represents the best performance of a
frontal vehicle structure.
The calculation results of the above formula could be seen in the Figure 4.7. It could
be noticed that the vehicle “d” has the lowest score because it has big structural
interaction in the middle area and almost no interaction in the lower area. The vehicle
“e” went higher in the ranking because it had good balance of interaction between the
lower and middle area.

Figure 4.7 Score of the lower and middle area of the PDB for the sample vehicles
More evaluation of the PDB deformation could be performed specially for the upper
area. The upper area of the PDB has strange deformations in the upper edge of the
barrier (ex. Vehicle b, e and f) because the upper part of the PDB is pulled down.
Thus, the evaluation of the upper part of PDB is uncertain in the provided sample
PDBs. The procedure of evaluation could be a combination of amount deformation to
prevent large deformations in the upper area (regardless of homogeneity).
Evaluation of the results
The Figure 4.8 shows the score for the lower and middle area of the PDB deformation
for C2PDB1 and C2PDB2 crash scenarios. In the standard PDB deformations, The
BA has the lowest score because it had the best interaction of the sub-frame to the
lower area of the PDB. The exterior, like the front bumper cover were excluded from
the vehicle model, thus the frontal structures have an aggressive interaction to the
PDB. This is the reason for having lower scores for more interaction of the sub-frame
cross beam to the PDB. The EX model has similar score to the WO and BA because
the sub-frame in EX did not have proper interaction (Figure 3.6).

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

39
 l

Figure 4.8 Score of the PDB deformations from results

In the modified PDB configuration scores, increasing the length of a sub-frame
decreased the evaluation score. The sub-frame cross beam had proper interaction to
the lower area of the PDB and elongation of the sub-frame caused deeper
deformations and bigger total variation of the deformation in the lower area, for this
reason score of the lower and middle area decreased. It could be noticed that the WO
model has the highest score in comparison to other models with sub-frame cross
beam.
Discussion
The PDB barrier was capable of evaluating the crash compatibility of vehicles, the
assessment of the evaluation is more important than the PDB itself.
The criteria explained in this chapter were based on the following factors:
- Over/under running effect, detected by term.
- Fork effect, detected by term.
- Small and not enough overlap, detected by ratio between the reference depth
of the middle and lower area (ratio ML).
The formulas and methods were examined on 36 PDB samples from the FIMCAR
project and explained based on 8 selected models of those models. The criteria were
applied on the PDB deformations from the results of the project also.
By extending the subframe parts, lower scores were achieved for the PDB assessment
on the simulation results of C2PDB2 despite longer sub-frame showed better self
protection in C2C simulations. The vehicle models had aggressive front ends with
locally stiff structure and excluded exterior parts and the PDB has uniform strength in
its honeycomb structure which is sensitive to pressure. Thus each of load paths
interacted to PDB individually and sub-frame cross beam had smaller area and caused
more local pressure to the lower area and as a consequence the total variation of the
lower area increased. The longer sub-frame intruded more into the lower part of the
PDB locally and increased the total variation more, thus the score for the extended

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

40
subframe was the lowest. In C2C cases, the front structures were in perfect match in
the vertical direction and longer sub-frames showed better performance.
Investigation of the research of Park [13] showed that misalignment of the interacting
front structures would cause more intrusion to the firewalls of the colliding cars with
all different sub-frame configurations. As a result, the PDB should give information
about the worse case for interaction of the front structures of the examined vehicle,
rather than in the case of matching front structures of the colliding vehicles.
The current criteria has good abilities to compare different PDB models in the factors
described above, but some thresholds should be defined for the failure of the cars in
the examination and some for maximum scores.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

41
5 Summary and conclusions
Crash test based legal requirements for vehicle compatibility do not exist. The French
proposal is to change Regulation No. 94 [4] using the PDB (progressive deformable
barrier) for frontal collisions. This study reviews the ability of the PDB to assess the
collision compatibility between vehicles. For this purpose, the simulations of car-to-
car crash with different sub-frame configurations were done as well as car to the PDB
cases. The simulation results were compared and the correlation between car to PDB
and car to car crashes was found.
The study revealed the PDB’s ability to detect lower structural changes in the front
end. The deformed PDB provides useful data to investigate the aggressivity of
vehicles, but the PDB barrier does not have a validated system for the evaluation of
the partner vehicle protection. Thus, the PDB criterion was developed.
The car to car simulation results showed that the longer sub-frame configuration
decreased fork effect and firewall intrusions and the frontal structures of the vehicles
started interacting earlier during C2C simulation. While the C2PDB1 with EX crash
test signified improper interaction of the PDB and the sub-frame. The interaction of
the PDB and the EX sub-frame encountered issues during the C2PDB1 simulations.
The front lower edge of the PDB front deformable core was lifted up by the cladding
sheet and the contact between the sub-frame and PDB did not occur. Furthermore, the
local forces of the structural interaction was distributed by the thick cladding sheet
and contact plate. For this reason, the PDB was modified.
After modifications to the PDB model, the C2PDB2 simulations showed that the sub-
frame interaction issue was solved and the PDB front deformable core deformations
were more local. Also, the vehicle firewall intrusions were bigger and more similar to
C2C compared to C2PDB1 results.
During the C2C, C2PDB1 and C2PDB2 result comparison, it was observed that the
PDB did not behave like a vehicle because the structural design of the front of a
vehicle is inhomogeneous while the PDB structure is uniform. Deformation shape of
the PDB front is caused by the pressure applied through the front end structures of
the vehicle. Hence, the PDB structure was deformed locally by the vehicle structure.
For this reason the deformed front surface of the PDB was used for crash
compatibility criteria. The PDB criteria took into account under running, overrunning,
fork effect and small overlap aspects.
The longer sub-frame configurations caused a better self protection during C2C crash
tests. The standard FE PDB version was not able to detect some configurations of the
lower vehicle structures, this issue was solved by modifying the barrier. After the
modifications, the PDB detects lower structures. Because of higher sensitivity of the
modified PDB to the local deformations and aggressive front structures of the vehicle
model, the longer sub-frame configurations earned lower scores in the PDB criteria.
The physical car to PDB crash tests deformation results of the front PDB surface
showed that the barrier is able to detect different vehicle structures. The
aggressiveness of a car can be evaluated by the compatibility criteria. The developed
criteria assessed PDB deformations in a reasonable order and proved that the
compatibility of a car can be evaluated by the analysis of the deformed PDB and the
vehicle. Thus the PDB has ability to detect the structural interaction, but the procedure
that the PDB deformation would be evaluated is crucial.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

42
6 Recommendations
During the crash test simulations of the original Taurus vehicle model to PDB, the
deformations of the barrier were more distributed. The local deformations of the PDB
were more concentrated during the C2PDB1 simulation. This trend happened because
the exterior parts of the original model distributed forces more than the simplified
model. For this reason the simplified vehicle model presented more severe PDB
model deformation than the standard vehicle. The EX to PDB simulation revealed the
structural interaction issue in the lower area. Therefore the PDB should be validated
in ultimate cases as well, even if the case does not represent existent vehicle
configuration.
The standard FE PDB was not able to detect the lower structural interaction and had a
negative sliding energy issue during the EX to C2PDB1 simulation. The modified FE
PDB model has detected the lower structures, but the negative sliding energy issue
was magnified. Thus, the FE PDB model should be developed further. One of the
approaches could be the development of PDB model based on shell elements instead
of the solid elements to present aluminium honeycomb crushable foam specifications.
Another solution could be to develop the FE PDB model with the shell elements in the
front deformable core, while other cores would remain with solid elements. The
hybrid configuration requires less simulations time than FE PDB model with all shell
elements, in this case the front PDB face and the sub-frame structural interaction issue
also could be solved.
The final FE PDB version should be validated by the rigid tubular impactor and a
detailed car to the PDB simulation case. These results should be compared and
validated with corresponding physical tests.
The proposed PDB criterion evaluates aggressiveness of a vehicle. Thus, the criteria
would be used in the new regulation and should be well defined, because the
evaluation results will be compared and vehicles would be rated based on it.
Therefore, the rating should be robust, otherwise there is a risk of unreasonable scores
for some vehicles. Hence, the proposed criteria should be verified by more of PDB
crash tests. As a result the implementation of a faulty criterion would be prevented.
The PDB barrier is sensitive to pressure which means, structures with smaller front
surface would make big intrusions in the PDB surface compared to structures with
bigger front surface (with the same stiffness). Thus the front structure of the vehicle
should have the responsibility of distributing crash forces among the different load
paths of vehicle front structure. This could be achieved by increasing the number of
load paths in the front structure and joining them together for more force distributing
effect in the interacting height of the lower and middle area of the PDB, because side
collision scenarios also should be considered.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

43
7 References
[1] http://ec.europa.eu/transport/road_safety/pdf/observatory/trends_figures.pdf
[2] http://www.pdb-barrier.com/files/SAE2004-20O4B-160pdf.pdf
[3] Davies, H. C., TRL Limited (2003-2006): Experimental development of improved
vehicle compatibility, Final technical report, VC – COMPAT

[4] Proposal submitted by France, 2007. Regulation No. 94 (Frontal collision)

Proposal for draft amendments Proposal submitted by France. Economic
Commission for Europe inland Transport Committee World Forum for
Harmonization of Vehicle Regulations Working Party on Passive Safety.

[5] Development of assessment criteria for off-set test, FIMCAR, January 2011
workshop (www.fimcar.eu)

[6] http://www.ncac.gwu.edu/vml/models.html

[7] Park, C.K. et al., Simulation of Progressive Deformable Barrier (PDB) Tests.
dynalook.com.

[8] EEVC WG 15 (2005): EEVC Approach to the Improvement of Crash

Compatibility between Passenger Cars, Proceedings of the ESV Conference, 05-0155
[9] Zobel, R. Schwarz, T. Thomas G. (2005): Towards a beneficial, scientifically
meaningful, and applicable compatibility-testing, AECA, 05-0052, Germany
[10] Summers, S. Hollowell, T. Prasad, A., NHTSA’s program for vehicle
compatibility, NHTSA, 307, USA

[11] Tatsu, K. et al., 2009. An evaluation of PDB test results for partner protection
and self protection. Enhanced Safety of Vehicles.[vp]. 15-18 Jun.

[12] Aramov, N. Rachid, K., 2008, Car Crash Compatibility, A FE Parametric Study,
Master’s Thesis. Chalmers University of Technology, 2008:27, Göteborg

[13] Park, C.K. et al., 2009. The Influence of Subframe Geometry on a Vehicle’s
Frontal Crash Response. Enhanced Safety of Vehicles.[vp]. 15-18 Jun, pp.1-11.

[14] Wu, T., Chhim, M., 2009, Frontal Crash Compatibility, Influence of load path
configuration, Master’s Thesis. Chalmers University of Technology, 2009:08,
Göteborg

[15] Delannoy, P. et al., 2004. New barrier test and assessment protocol to control
compatibility. In SAE World congress 2004.

[16] EEVC WG 15, 2007, European Enhanced Vehicle-safety Committee, Working

Group 15 Car Crash Compatibility and Frontal Impact Final Report to Steering
Committee. EEVC.

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

44
[17] Meyerson, S., Delannoy, P. & Robert, G., 2009. Evaluation of advanced
compatibility frontal structures using the Progressive Deformable Barrier (PDB).
Enhanced Safety of Vehicles.[vp]. 15-18 Jun.

[18] Delannoy, P., Martin, T. & Castaing, P., 2005. Comparative evaluation of frontal
offset tests to control self and partner protection. Enhanced Safety of
Vehicles.[vp]. 6-9 Jun, pp.1-12.

[19] LS-DYNA Keyword User’s Manual, May 2007, Version 971, LSTC.

[20] Krusper, A., Thomson, R., Energy-absorbing FUPDs and their interactions with
fronts of passenger cars, 2010, International Journal of Crashworthiness

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

45
Appendices
Appendix A. PDB Model Validation
Tubular impactor crash to PDB for the conformity of the PDB model and its data and
comparison of that with the physical tests. The test procedure is described in the
proposal for the Regulation No.94 [4].

Force vs displacement

600
0,5; 500
500

400
Force (kN)

0,2; 300 0,5; 400

300 0,05; 250

200
0,2; 200
0,05; 150
100

0
0,00 0,10 0,20 0,30 0,40 0,50 0,60
Displacement (m)
PDB-XT 1 PDB-XT 2 CAE model

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

46
Appendix B. MATLAB Codes
1. Homogeneity calculation code

%% Compute homogeneity of lower, middle and upper area

% Calculates the homogeneity for lower area

Lower=deformation(start_index_z:low_end_index,start_index_y:end_index_y);

HL = 0;
for i = 1:length(Lower(:,1))-1
for j = 1:length(Lower(1,:))-1
a = abs(Lower(i,j+1)-Lower(i,j)) + abs(Lower(i+1,j)-Lower(i,j));
HL = HL + a;
end
end

HL = HL / ( 2 * (length(Lower(:,1))-1) * (length(Lower(1,:))-1) );
HL = 1 / HL;

% Calculates the homogeneity for Middle area

Middle=deformation(low_end_index+1:mid_end_index,start_index_y:end_index_y);

HM = 0;
for i = 1:length(Middle(:,1))-1
for j = 1:length(Middle(1,:))-1
a = abs(Middle(i,j+1)-Middle(i,j)) + abs(Middle(i+1,j)-Middle(i,j));
HM = HM + a;
end
end

HM = HM / ( 2 * (length(Middle(:,1))-1) * (length(Middle(1,:))-1) );
HM = 1 / HM;

% Calculates the homogeneity for Upper area

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

47
Upper=deformation(mid_end_index+1:end_index_z,start_index_y:end_index_y);

HU = 0;
for i = 1:length(Upper(:,1))-1
for j = 1:length(Upper(1,:))-1
a = abs(Upper(i,j+1)-Upper(i,j)) + abs(Upper(i+1,j)-Upper(i,j));
HU = HU + a;
end
end

HU = HU / ( 2 * (length(Upper(:,1))-1) * (length(Upper(1,:))-1) );
HU = 1 / HU;

2. Reference depth calculation code

%% Hole detection of lower, middle and upper area

%% Counts the number of grids in each depth div in lower area

Lower=deformation(start_index_z:low_end_index,start_index_y:end_index_y);

CountLow = zeros(1,n);
for i = 1:length(Lower(:,1))
for j = 1:length(Lower(1,:))
k = ceil( Lower(i,j) / DivSize);
CountLow(k) = CountLow(k) + 1;
end
end

% figure(20)
% t = DivSize:DivSize:PDBdepth;
% plot(t,CountLow)

% Finds the most common depth of deformation for lower area

for i = 1:n-m+1
CountLow2(i) = sum(CountLow(i:i+m-1));
end

[d f] = max(CountLow2);

DL = 0;
for i = 1:m
DL = DL + (f+i-1)*DivSize *(CountLow(f+i-1) / CountLow2(f));
end
% DL

HL = 0;
for i = 1:length(Lower(:,1))
for j = 1:length(Lower(1,:))
a = abs(DL-Lower(i,j))^p;
HL = HL + a;
end
end

HL = HL / (length(Lower(:,1)) * length(Lower(1,:)));

HL = 100/((HL)^(1/p));

%% Counts the number of grids in each depth div in middle area

Middle=deformation(low_end_index+1:mid_end_index,start_index_y:end_index_y);

CountMiddle = zeros(1,n);
for i = 1:length(Middle(:,1))
for j = 1:length(Middle(1,:))

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

48
 k = ceil( Middle(i,j) / DivSize);
CountMiddle(k) = CountMiddle(k) + 1;
end
end

% Finds the most common depth of deformation for Middle area

for i = 1:n-m+1
CountMiddle2(i) = sum(CountMiddle(i:i+m-1));
end

[d f] = max(CountMiddle2);

% figure(21)
% t = DivSize:DivSize:PDBdepth;
% plot(t,CountMiddle)

DM = 0;
for i = 1:m
DM = DM + (f+i-1)*DivSize *(CountMiddle(f+i-1) / CountMiddle2(f));
end
% DM

HM = 0;
for i = 1:length(Middle(:,1))
for j = 1:length(Middle(1,:))
a = abs(DM-Middle(i,j))^p;
HM = HM + a;
end
end

HM = HM / (length(Middle(:,1)) * length(Middle(1,:)));

HM = 100/((HM)^(1/p));

%% Counts the number of grids in each depth div in upper area

Upper=deformation(mid_end_index+1:end_index_z,start_index_y:end_index_y);

CountUpper = zeros(1,n);
for i = 1:length(Upper(:,1))
for j = 1:length(Upper(1,:))
if Upper(i,j) <= 0
Upper(i,j) = 0.1;
end
if Upper(i,j) >= PDBdepth
Upper(i,j) = PDBdepth-0.1;
end
k = ceil( Upper(i,j) / DivSize);
CountUpper(k) = CountUpper(k) + 1;
end
end

% Finds the most common depth of deformation for upper area

for i = 1:n-m+1
CountUpper2(i) = sum(CountUpper(i:i+m-1));
end

[d f] = max(CountUpper2);

DU = 0;
for i = 1:m
DU = DU + (f+i-1)*DivSize *(CountUpper(f+i-1) / CountUpper2(f));
end
DU

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

49
HU = 0;
for i = 1:length(Upper(:,1))
for j = 1:length(Upper(1,:))
a = abs(DU-Upper(i,j))^p;
HU = HU + a;
end
end

HU = HU / (length(Upper(:,1)) * length(Upper(1,:)));

HU = 100/((HU)^(1/p));

3. Firewall nodes data extractor from the nodout file (Based on the idea from
Krusper, A.)

function Firewall_graph(filename,Time,ShiftNodeID,LineShape,Thickness)

M = textread(filename,'%s');

NodeDatabase=NodeDB;
NodeDatabase(:,1)=NodeDatabase(:,1)+ShiftNodeID;
Nodes = NodeDatabase(:,1);

%% finding desired time's row in the file

i = 2;
while ~(strcmp(M(i), Time) && strcmp(M(i-1), 'time') )
i = i + 1;
end
d = i;

%% Shifting Nodes related to node number 3183940

NodeDatabase(:,2) = NodeDatabase(:,2) - NodeDatabase(4,2);

NodeDatabase(:,3) = NodeDatabase(:,3) - NodeDatabase(4,3);
NodeDatabase(:,4) = NodeDatabase(:,4) - NodeDatabase(4,4);

%% Extracting displacements from the desired time

for i= 1:length(Nodes)
k = 1;
while ~(strcmp( M(d+k), num2str(Nodes(i)) ))
k = k + 1;
end
NodeDisplacemant(i,1) = Nodes(i);
NodeDisplacemant(i,2) = str2double(M(d+k+1));
NodeDisplacemant(i,3) = str2double(M(d+k+2));
NodeDisplacemant(i,4) = str2double(M(d+k+3));

NodeCoorAfterDeform(i,1) = NodeDatabase (i,1);

NodeCoorAfterDeform(i,2) = NodeDatabase (i,2) + NodeDisplacemant(i,2);
NodeCoorAfterDeform(i,3) = NodeDatabase (i,3);
NodeCoorAfterDeform(i,4) = NodeDatabase (i,4);
end

CHALMERS, Applied Mechanics, Master’s Thesis 2011:28

50