Improved vehicle crashworthiness design by control of the energy absorption for different collision situations

(1)

energy absorption for different collision situations

Citation for published version (APA):

Witteman, W. J. (1999). Improved vehicle crashworthiness design by control of the energy absorption for different collision situations. Technische Universiteit Eindhoven. https://doi.org/10.6100/IR518429

DOI:

10.6100/IR518429

Document status and date: Published: 01/01/1999

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne Take down policy

If you believe that this document breaches copyright please contact us at: openaccess@tue.nl

(2)

Improved Vehicle Crashworthiness Design

by Control of the Energy Absorption

(3)

CIP-DATA LIBRARY TECHNISCHE UNIVERSITEIT EINDHOVEN Witteman, Willibrordus J.

Improved Vehicle Crashworthiness Design by Control of the Energy Absorption for Different Collision Situations / by Willibrordus J. Witteman.

- Eindhoven : Technische Universiteit Eindhoven, 1999. Doctoral dissertation, Eindhoven University of Technology With literature list and summary in Dutch.

Proefschrift.

ISBN 90-386-0880-2 NUGI 834

Subject headings: vehicles; crashworthiness / longitudinal member / vehicle structure / numerical simulation

Trefwoorden: voertuigen; botsveiligheid / langsligger / autoconstructie / numerieke simulatie

(4)

Improved Vehicle Crashworthiness Design

by Control of the Energy Absorption

for Different Collision Situations

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de

Rector Magnificus, prof.dr. M. Rem, voor een commissie aangewezen door het College voor

Promoties in het openbaar te verdedigen op dinsdag 15 juni 1999 om 14.00 uur

door

Willibrordus Jacobus Witteman

(5)

prof.dr.ir. R.F.C. Kriens en

(6)

(7)

(8)

Summary

Increased traffic intensity, growing concern of the public and new stringent legislation, have made vehicle safety one of the major research areas in automotive engineering. Especially the unfavorable crash results (large deformation of the passenger compartment of many cars) occurring in more realistic crash tests, which deviate from the compulsory full overlap crash test against a concrete block, are reason to worry. In the case of a partial frontal overlap (offset) collision or an off-axis crash direction only part of the vehicle structure can be used for energy absorption. This leads to dangerous intrusions of the passenger compartment, because only one of the two longitudinal members is used for energy absorption.

This thesis describes the design of a new frontal vehicle structure that directs the asymmetric crash load of an offset collision as an axial load to the second unloaded longitudinal member. Only by using both longitudinal members and through a progressive folding pattern, enough energy can be absorbed in the front structure to prevent a deformation of the passenger compartment. To prevent a premature bending collapse, the new longitudinal members consist of two functional components: an inside square crushing column for a normal stable axial force level and a stiff outside sliding supporting structure that gives the necessary extra bending resistance. An integrated cable system transmits the force to the other longitudinal member. With this novel design concept, a vehicle has similar energy absorption in the front structure for the entire range of collision situations (full, offset, oblique). By means of numerical crash simulations, this concept has been optimized and evaluated. Results show that for an entire range of frontal collision situations similar deceleration curves can be obtained. However, to further reduce the injury level of the occupants, optimal crash decelerations for various crash velocities are necessary. To this aim, a method is described for numerical FEM dummy simulations to obtain optimized crash pulses for different velocities. The novel concept is very suitable to adapt the structural stiffness to these new deceleration pulses. To realize the optimal deceleration during the crash for each velocity, solutions have been presented based on controllable energy absorption by additional friction or based on controllable hydraulic flow restriction. With this total design, an optimal vehicle deceleration curve is possible for each velocity over the entire frontal collision spectrum, yielding the lowest levels of the occupant injury criteria.

(9)

(10)

Samenvatting

Toegenomen verkeersintensiteit, meer bewustwording van de consument en strengere wetgeving, maken veiligheid van auto’s tot belangrijk onderzoeksgebied in de automobielontwikkeling. Vooral de tegenvallende botsresultaten (grote vervormingen van het passagiersgedeelte van veel auto’s) in op de realiteit gelijkende botstesten (die afwijken van de verplichte volledige overlap botstest tegen een betonblok), zijn zorgwekkend. In het geval van een gedeeltelijke overlap (offset) botsing of een schuine botsrichting kan slechts een gedeelte van de voertuigconstructie gebruikt worden voor de energie absorptie. Dit lijdt tot gevaarlijke intrusies van het passagierscompartiment, omdat slechts één van de twee langsliggerbalken voor energie absorptie wordt gebruikt.

In dit proefschrift wordt een nieuwe frontale voertuigconstructie beschreven, die bij een offsetbotsing de krachten op slechts één getroffen langsligger doorleidt naar de andere ongetroffen langsligger en wel zo dat deze ook axiaal deformeert. Alleen op deze manier absorberen beide langsliggers met voortgaande plooivorming genoeg energie om een vervorming van het passagiersgedeelte te voorkomen. Om een vroegtijdige knik te voorkomen bestaan de nieuwe langsliggers uit twee functionele delen: een inwendige vierkante crashkoker voor een normaal stabiel axiaal krachtniveau en een uitwendige stijve ondersteunende schuifconstructie voor de noodzakelijke extra buigweerstand. Een geïntegreerd kabelsysteem zorgt voor de krachtoverbrenging naar de andere langsligger. Met dit nieuwe ontwerp heeft een auto een vergelijkbare energie absorptie in de frontale voertuigconstructie voor de hele range van botssituaties (volledig, offset, schuin).

Met behulp van numerieke botssimulaties is het nieuwe concept geoptimaliseerd en beproefd. De resultaten laten zien dat vergelijkbare vertragingscurven kunnen worden verkregen voor een heel spectrum van frontale botssituaties. Om echter het letselniveau van de inzittenden verder te reduceren, zijn de optimale botsvertragingen bij verschillende botssnelheden nodig. Hiervoor is een methode beschreven waarmee met numerieke FEM dummy simulaties geoptimaliseerde vertragingscurven bij verschillende snelheden zijn verkregen. Het nieuwe concept is zeer geschikt om de constructiestijfheid aan te passen voor deze nieuwe vertragingscurven. Om de optimale vertraging tijdens een botsing voor elke snelheid te realiseren, worden constructieve oplossingen voorgesteld gebaseerd op regelbare energie absorptie door extra wrijving of gebaseerd op regelbare hydraulische stroombegrenzers. Met dit totale ontwerp is een optimale voertuig vertragingscurve

(11)

voor elke botssnelheid voor het gehele frontale botsgebied mogelijk, wat tot de laagste niveaus van de letselwaarden leidt.

(12)

Chapter 1 General Introduction

Serious or fatal traffic accidents are considered as one of the most threatening dangers in daily life. It is an unexpected event that can change people’s life radically. In the Netherlands (15 million inhabitants) about 1200 people are killed annually and more as 12000 people are injured. About half of these are attributed to car occupants. See Table 1.1 in which accident statistics are shown since 1990. Despite increased mobility, these values are slightly on the decline. Although a safe driving style minimizes accident risk, car occupants are also exposed to unexpected road conditions and risky or drunken drive behavior of other road users (like a badly timed passing maneuver or a slipping car on the wrong side of the road). Especially, frontal accidents on country roads against other cars have a high fatality rate. Due to efforts to avoid frontal collisions, the car front is generally only partly involved and not always axially. In addition, the incompatibility between different vehicles yields more fatalities. These collision situations are until now not legally tested. Only a few car manufacturers are using such collision situations as safety design goal for a longer time.

Table 1.1.

Number of death people of different traffic participants in The Netherlands (SWOV 1998). 1990 1991 1992 1993 1994 1995 1996 1997 Car driver 456 425 425 428 437 465 414 399 Car passenger 246 205 201 187 177 192 161 148 Truck 8 19 14 10 15 16 15 11 Delivery van 45 39 36 28 61 41 44 57 Bus 2 0 7 2 0 1 1 3 Motorcycle/scooter 72 88 93 106 112 90 91 92 Motorbike 84 97 81 74 73 81 79 67 Small motorbike 14 16 24 18 25 37 28 21 Bicycle 304 238 251 244 269 267 233 242 Pedestrian 144 145 152 147 124 142 109 119 Other 1 9 1 8 5 2 5 4 Total 1376 1281 1285 1252 1298 1334 1180 1163

(17)

Increased traffic intensity, growing concern of the general public, and more stringent legislation have made vehicle safety one of the major research areas in automotive engineering. An area of particular concern is the early crashworthiness design of cars. Cars have to pass the compulsory crash test as issued by the authorities. However, this test does not guarantee that cars are safe in crash situations that deviate from the prescribed one (Witteman 1993). Hence, the entire collision spectrum within which the car must be safe must be investigated.

1.1. Description of the research issue

The improved frontal crashworthiness of cars necessitates totally new design concepts, which take into account that the majority of collisions occur with partial frontal overlap and under off-axis load directions. Realistic crash tests with partial overlap have shown that conventional longitudinal structures are not capable of absorbing all the energy in the car front without deforming the passenger compartment. See Figure 1.1 for an offset test against a rigid barrier and Figure 1.2 in which a car is used in a full overlap and in an offset test against a deformable barrier for comparison. It is clear to see that in case of a full overlap collision there is no intrusion of the passenger compartment, while in the offset test the passenger compartment of the same car collapses. The reason for this is that the structure of the longitudinal members is specifically designed for meeting the less severe requirements of the compulsory full overlap test, in which both longitudinals are loaded axially.

(18)

Figure 1.2. Example of identical cars with the same collision speed in a full (above)

and in an offset (under) crash (video fragment TRRL).

For improved frontal car safety it is necessary to design a structure that absorbs enough energy in each realistic crash situation. To protect the occupants, the passenger compartment should not be deformed and intrusion must be avoided too. To prevent excessive deceleration levels, the available deformation distance in front of the passenger compartment must be used completely for a predetermined crash velocity. This implies that in a given vehicle concept the structure must have a specific stiffness. Normally, the two main longitudinal members will absorb most of the crash energy with a progressive folding deformation of a steel column. The main problem is that in real car collisions these two longitudinal members often are not loaded in a synchronous fashion and also not loaded pure axially. The majority of collisions occur with partial frontal overlap, in which only one longitudinal is loaded, or under an off-axis load direction. This implies that most longitudinals fail under a premature bending collapse rather than a much more energy absorbing progressive folding pattern. This gives rise to two design conflicts. The first conflict is that the same amount of energy must be absorbed either with a single or with both

(19)

longitudinals. The second conflict is that the same amount of energy must be absorbed in the case of an off-axis impact angle as in the case of a normal axial impact. These problems can not be solved by just increasing the stiffness of the longitudinals in such a way that each longitudinal is capable of absorbing all of the energy, see the following reasons. To absorb enough energy, a stiff longitudinal is needed for the offset crash in which normally only one longitudinal is loaded. The same longitudinal must be more supple in case of a full overlap crash, since both longitudinals must not exceed the desired deceleration level (Witteman 1993). In addition, a stiff longitudinal is needed to absorb enough energy in an off-axis load direction resulting in a higher bending resistance to help transform off-axis loads into axial loads and to prevent a bending collapse. The same but more supple longitudinal is needed in the case of a normal axial load to avoid overly high deceleration forces. Another issue is the crash velocity. To absorb all the kinetic energy, which is proportional with the square of the velocity, the deformable structure length must have a specific stiffness. This stiffness results in an average mean force, which multiplied with the deformation shortening gives the absorbed energy. For an acceptable injury level of the occupants, the total deceleration level must be as low as possible, using the maximum available deformation length without deforming the passenger compartment. This means that for example in a 64 km/h crash compared with a 32 km/h crash, a four times longer deformation distance is needed for the same deceleration level. See Figure 1.3 in which the relation between impact energy, deformation length, force and crash velocity is plotted for different vehicle masses, stiffnesses and average deceleration levels. In this figure the example is plotted with the dashed lines. For a crash velocity of 32 km/h respectively 64 km/h the necessary deformation length for the same constant deceleration of 20 g is 20 cm respectively 80 cm. Since a deformation length is mostly restricted to 60 – 80 cm, it is not desirable to use only 15 – 20 cm with a relative high deceleration level for the 32 km/h crash. Another example is plotted with the dotted lines, how to ‘walk around’. There is started with a 56 km/h crash velocity and a vehicle mass of 1100 kg. This gives an impact energy of 133 kNm. In case of 80 cm available deformation length, the crash force is 166 kN and the average deceleration is 15.4 g. Although the stiffness normally increases during the crash and at higher crash speeds there is made use of the stiff engine; the only way to generate an optimal crash pulse at different collision speeds is variable structure stiffness. After detection of the crash velocity, the optimal stiffness of the longitudinal member should be realized.

(20)

1.66 32 0.2 0.8 56 1.33 15.4 64 Figure 1.3. “

(21)

1.2. Research objectives

The objective of the research project presented here, was to design a concept structure that substitutes the conventional energy absorbing longitudinal members in a frontal vehicle structure and yields optimized deceleration pulses for different crash velocities and overlap percentages. To this aim the structure must have a stiffness that can be varied in accordance to the specific crash situations.

The novel design presented in this thesis can cope with the following four crashworthiness problems:

1. In the case of a full overlap crash (both longitudinals and engine involved) as in the case of an offset crash (at 40 per cent overlap only one longitudinal directly involved) a similar amount of energy must be absorbed by the front structure. 2. In the case of an oblique load direction as in the case of an axial load direction a

similar amount of energy must be absorbed by the front structure.

3. With a not much longer deformation length, much more energy must be absorbed at high crash velocities (resulting in less fatal injuries) and a lower injury level must be obtained at lower crash velocities.

4. A deceleration pulse must be obtained which is optimal (lowest injury level) for the concerning collision speed and the chosen dummy restraint parameters.

1.3. Research strategy

Increased protection for the entire collision spectrum can be obtained by structures consisting of longitudinal members with an advanced geometric form, giving higher bending resistance without increasing the axial stiffness, in conjunction with a rigid connection between the front ends of these members. From several longitudinal square cross-sections, the influence of the width and thickness dimensions on the crash behavior is evaluated. The purpose of this study is to conceive an advanced geometric design for a longitudinal member optimized for a wide collision spectrum. The influence of various crash situations on the amount of energy absorbed by such longitudinal members will be discussed and representative crash tests are proposed. However, to reach a similar amount of energy absorption in case of an offset collision compared with a full overlap situation, additional measures are necessary. For this, the other longitudinal, which is not directly loaded, has to crumple axially as well. A cable system is introduced to perform axial shortening of the unloaded longitudinal with a tensile force to the rear. To prove the new concept, numerical simulations of a

(22)

frontal car structure have been carried out. In these simulations, deceleration levels in the same order of magnitude are found. Since it has been demonstrated that it is possible to design a vehicle structure which generates a crash pulse that is almost independent of the crash direction and overlap percentage, it is useful to do an independent search for optimal pulses at several crash velocities, because the found structure-based pulses are not obviously the optimal pulses for minimal injury to the occupants. Therefore, the reverse question is answered: which crash pulse gives the lowest injury levels with an already optimized restraint system, instead of finding the optimized restraint system for a given crash pulse. For this research, a method is described in which a numeric model of an interior and a FEM dummy has been used to find the levels of the injury criteria. To compare the results of different crash pulses, an overall severity index has been used. From a described research an optimal pulse has been found after several considered pulse variations at a crash speed of 56 km/h. This pulse, used as example, deviates much from a traditional pulse, which shows normally an increasing stiffness of the structure near the end of the crash, but gives as it seems much lower injuries. During the first part of the deformation length the deceleration level can be high, then a low deceleration interval is desired, and at the end (dummy is restrained by belt and airbag) the deceleration can be high again. Also for other crash velocities, pulses are mentioned with adapted pulse characteristics for optimal results. Finally, new ideas are given how to further customize the energy absorption for different crash velocities to reach the optimal pulse. Therefore, an intelligent structure must be built which realizes such optimal pulses as closely as possible. In case of a high passenger floor (like an MPV), new design concepts are discussed.

The research described in this thesis has a mean focus point on technically realizable design solutions, which are realistic but conscious not optimized initially for the weight and costs to find the highest technical potential of the proposed solutions.

1.4. Thesis outline

The contents of this thesis are structured as follows.

In Chapter 2, an inventory is given of parameters that determine useful crash situations. The influence of specific crash situations on the vehicle structure load is analyzed and arguments are mentioned which prove why at least two frontal impact tests are necessary for vehicle crashworthiness assessment. Two extreme crash situations are represented, which generally result in very different crash pulses and which cope with the just prescribed new legal test situations.

(23)

To find optimal cross-section geometry of the new longitudinal, in Chapter 3 several basic forms are compared on energy absorption at different collision situations. For a square cross-section, the influence of thickness and perimeter on crashworthiness aspects is analyzed. From these results, a new concept is represented, which combines a higher bending resistance with an unchanged axial force level. In addition, the most efficient way of triggering to generate a stable folding process starting at the front part of the longitudinal is discussed. The manufacturing of the new longitudinal and a quasi-static experiment on a part as verification is described. The folding process is as expected.

To absorb enough energy in offset collisions as well, an additional cable system is designed to operate with the new longitudinal structure. This system is described in Chapter 4. A numerical model of a vehicle front has been developed to simulate different crash situations. The results show that the system works (reported problems are mostly numerical) and that in each crash situation a similar deceleration curve can be obtained.

In Chapter 5, a method for numerical crash simulations with a dummy is described, which can be used to find an optimal deceleration pulse. Structural designs are presented with the aid of which optimal pulses can be obtained. Especially a stronger deceleration in the first part of the crash needs an intelligent solution, in which the additional energy absorption is dependent on the crash velocity. Design concepts based on friction or based on hydraulic flow restriction are presented.

(24)

Chapter 2 The Necessity of Improved Crashworthiness Design

2.1. Introduction

In the present chapter, the crash situation and the tests that are used to assess vehicle crashworthiness are reviewed. It will be shown that at least two frontal impact tests are necessary for frontal vehicle crashworthiness assessment.

In recent years a large number of frontal crash tests have been reported (Auto Motor und Sport 1992, Consumentengids 1993 1994) which were designed specifically to represent real crash situations. In these tests, many cars showed a crash behavior leading to unacceptable deformations of the passenger compartment yielding less survival space. Apparent reason for this is that insufficient design efforts were made to cope with the high and uneven mechanical loads occurring in such realistic collisions. The design was based on the until 1998 only compulsory full overlap test against a rigid wall. Improved frontal crashworthiness of cars necessitates additional design requirements, which take into account that the majority of collisions occur with partial frontal overlap, at oblique angles of incidence, and at velocities that deviate significantly from the regulated test speeds. In reality, the collision statistics ask for a design that is optimized for the entire range of frontal collision situations.

Ideally, a vehicle's crashworthiness should be validated by multiple crash tests. Evidently, this is economically unfeasible. To solve this problem the introduction of a set of at least two compulsory tests are proposed in this chapter. If these chosen tests represent extreme cases within the frontal collision spectrum, crashworthiness in intermediate situations may then be interpolated from the respective test results. In new car development projects, the car body is optimized to comply with the compulsory crash tests as issued by the authorities. Consequently, the production car will show a sufficient level of crashworthiness in collisions that match or closely resemble those simulated in the tests. However, in collision situations that deviate from those represented in the tests, its actual performance is uncertain. Defining the proper crash tests is very important since it largely influences the crashworthiness design and, ultimately, determines the passive safety performance of the car. Although this topic also concerns side impact crashes, the discussion in this thesis will be limited to frontal crashes.

(25)

2.2. Frontal crash parameters

Obviously, one must know the entire spectrum of all possible collision types in order to design a car that will be safe enough in any collision that may occur. Collision statistics reported by car manufacturers (Seiffert 1992, Justen 1993) and research institutes (Her Majesty 1991) and in the National Automotive Sampling System (NASS) (Stucki 1998) and the Fatality Analysis Reporting System (FARS) provide an extensive database of crash parameters. Important parameters are the collision speed, the obstacle type, the impact location and the impact direction. From this databases, suitable ranges can be defined which cover the collision situations most likely to occur. The following observations can be made:

1. At least 90 percent of all frontal collisions take place at speeds up to 56 km/h, see Figure 2.1.

2. Naturally, the number of different obstacle types is endless. However it appears that a major division can be made into three standard obstacle types, i.e. the rigid wall (simulating buildings or heavy trucks), the deformable barrier (simulating other vehicles), and the pole (simulating trees and pillars).

3. The majority of frontal collisions happen with frontal overlap percentages (the part of the bumper that makes contact with an obstacle) varying from 30 up to 100 per cent.

4. Frontal collisions are considered to occur in impact directions having angles of incidence with the longitudinal car axis varying from -30 degrees up to 30 degrees.

How each of these crash parameters influences the proper design of the vehicle's safety structure will be briefly discussed.

(26)

Cumulative frequency in %

Figure 2.1. Cumulative frequency of velocities in frontal collisions

(source: Justen 1993). _____100 % --- 50 % -.-.- 40 % ... 30 % Velocity [km/h] 2.2.1. Collision speed

For optimal frontal crash behavior, all kinetic energy should be dissipated by the front structure. The lowest deceleration level of the passenger compartment is obtained, if the available deformation length in front of the car is as long as possible, see Figure 2.2. In this figure the needed deformation length at an average deceleration level is given. The upper curve is based on a completely linear relation between force and deformation length; the lower curve is based on a constant force independent of the deformation length, which gives the highest total energy absorption. In reality, the relation between deceleration level and deformation length will lie between these two curves.

For a specific crash velocity, the optimal situation is achieved if the entire available deformation length is used without deforming the passenger compartment. This implies, that in a given vehicle concept the structure must have a specific stiffness which is determined by the relation between the crash energy at this velocity and the available deformation length. Higher velocities result in a higher level of kinetic energy, which cannot be fully dissipated by this front structure. Hence, the passenger

(27)

compartment has to deform, which means that the necessary survival space cannot be guaranteed. Lower velocities will not use the whole available front structure causing the forces acting upon the occupant to be higher than necessary.

In case of a deformable barrier, the barrier also absorbs energy and increases the total deformation length. So for a similar level of energy absorption in the vehicle structure, the crash velocity of the car against a deformable barrier must be higher as in case of a crash against a rigid wall.

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 20 40 60 80 100 120

deformation at a speed of 56 km/h against a rigid wall [m]

deceleration level [g]

F

S

S F

Figure 2.2. Average deceleration level as a function of deformation length.

2.2.2. Obstacle type

The stiffness of the obstacle has a large influence on the crash behavior of a car. The stiff parts in the frontal structure of a car are the two main longitudinal members (possibly combined with the front wheel suspension) and the engine. They are responsi-ble for absorbing large amounts of energy during a crash. A typical longitudinal member that collapses in a regular pattern can absorb about 25 per cent of the impact energy

(28)

forces originating from the engine deform the stiff firewall, about half of the impact energy will be absorbed by the remaining front structure (Hobbs 1991). One must note that this is only true in the case of a collision against a rigid wall, a special situation in which very high load forces can be directed into the stiff parts of the car structure. In the case of a deformable barrier, the barrier will generally not be capable of generating such high loads. Hence, the stiff parts cannot deform from the very start of the impact. This results in large deformations in the supple parts of the structure. Consequently the front structure will absorb less energy. This leads to intrusion and deformation of the passenger compartment. Despite the fact that the decelerations are lower, the amount of crash energy that must be absorbed stays the same. The so-called dynamic stiffness of the structure (resistance against a rapid change of kinetic energy) becomes lower, which causes the deformation to occur near the passenger compartment rather than at the front end of the car (Hobbs 1993). Figure 2.3 shows a rough estimation of energy absorption distributed on the frontal structure during a crash with 56 km/h against a rigid barrier (De Santis 1996, Leeuwen 1997).

firewall front panel engine longitudinals _____________________________________________________________ 5 % 5 % 5 % 5 % 5 % 5 % 20 % 10 % 10 % 7.5 % 7.5 % 7.5 % 7.5 % Second half First half

Figure 2.3. Estimated energy absorption percentages in the frontal structure.

2.2.3. Collision place and direction

The overlap percentage determines which parts of the frontal structure of the car are hit and contribute to the energy absorption. In case of a full overlap against a rigid wall, the two stiff longitudinal members and the motion of the engine can absorb most of the

(29)

energy. During the first half of the crash duration, mainly the longitudinal members will be loaded. In the second half, the engine will be loaded as well. In crashes with partial overlap percentages of 70 per cent and lower, not all available stiff structures are used for energy absorption (Justen 1993), see Table 2.1.

Table 2.1.

Relative energy absorption for several frontal crash overlaps against a rigid wall (see Figure 2.3).

frontal overlap percentage

stiff parts in the structure

part of total energy absorption

first half of crash duration

part of total energy absorption second half of crash

duration 70 - 100% 2 longitudinals + surrounding structure + engine / firewall ≈ 50 % ≈ 50 % 40 - 70% 1 longitudinal + surrounding structure + engine / firewall ≈ 25 % ≈ 35 % 30 - 40% 1 longitudinal + surrounding structure ≈ 25 % ≈ 15 %

Research (Ragland 1991) has shown that in a car to car crash the percentages in the first column of Table 2.1 are even higher. Reason for this is that the car front has a non-uniform stiffness distribution, up to an overlap percentage of 50 per cent only one longitudinal member is absorbing energy, while the second member and the engine are not involved.

A crash against a stiff pole can be regarded as a crash with a small overlap against a rigid wall. Only one stiff part e.g. one of the longitudinals or the engine will be hit. Changing the crash direction in a test against a rigid wall from zero to 30 degrees leads to a so called glance-off (the car grazes the wall and changes direction to move further) if no anti slide (vertical strips on the rigid wall where the car structure hooks on) is used (Justen 1993). The crash load on the longitudinal will be lower. With the use of an anti slide, the car will turn with the front towards the wall and the bending

(30)

enough bending resistance. In general the 30 degrees impact presents a benign environment for a restrained occupant and did not reflect the impact conditions which lead to fatalities and serious injuries in the real world (Hackney 1985).

2.3. Analysis of useful crash situations

Accident analyses (Hobbs 1993) have shown that two-thirds of the collisions in which car occupants have been injured are frontal impacts, of which two-thirds occur in car to car accidents. The chances for collisions occurring with a full frontal overlap, or with an overlap percentage up to third, or with an overlap percentage from one-third up to two-one-thirds are comparable. This means that two-one-thirds of the frontal collisions use only part of the front structure. Also two-thirds of the collisions take place with impact directions under normal angle of incidence with the longitudinal car axis.

Defining the proper design requirement for a considered crash velocity is very important. Crashes at speeds above this design velocity lead to a higher number of fatal injuries. On the other hand, crashes at lower speeds lead to a higher number of minor injuries. Setting the design speed at 56 km/h is sufficient for more than 90 per cent of the frontal collisions. This speed is compatible with current restraint systems. A test speed of 48 or 50 km/h is too low, since one-third of the offset crashes having 40 or 50 per cent overlap occur at higher speeds, see Figure 2.1.

Summarizing, a suitable design crash situation will be crash at 56 km/h under normal angle of incidence against a deformable barrier with partial overlap.

Accident investigations have also shown that the major cause of serious and fatal injuries in frontal car crashes is the intrusion into the passenger compartment (Hobbs 1993). High seat belt forces on the occupant lead to minor injuries, while contact with the car interior or penetrating parts leads to major injuries. As a consequence, it is better to design a car body that is a lit too stiff (leading to increased deceleration forces) than a car body that is a lit too supple (leading to decreased deceleration forces but with too much intrusions).

Optimal safety implies on the one hand a sufficiently low deceleration level and on the other hand no intrusion of the passenger compartment under a wide range of crash situations. This is a design dilemma. The structure must be neither too stiff nor too supple. Referring to Table 2.1 it is clear that a full frontal overlap gives maximum energy absorption by the stiff parts. In a crash having an overlap of only 30 - 40 per cent,

(31)

merely half of the regular amount of energy is absorbed during the first half of the crash. During the second half, the amount is even worse, about 30 per cent of the regular value. Combining these percentages yields a total energy absorption of about 40 per cent of the regular value. This implies that the total stiffness of the structure available in this partial overlap crash is only 40 per cent of the stiffness regularly available in the full overlap crash. It must be remarked that a high stiffness is especially important near the

passenger compartment. This 2

½

times difference in stiffness affects the crash

behavior of a car substantially. For the same available deformation length the deceleration level of a car in a 30 - 40 per cent overlap crash is approximately 2

½

times as low as the deceleration level during a full overlap crash.

Suppose a car, designed for a full overlap crash, where the available deformation length at a specific collision speed would be fully utilized, becomes involved in a crash with a 30-40 per cent overlap with the same speed and the same available deformation length. Then, 60 per cent of the crash energy will not absorbed by the front structure. This remaining energy must be absorbed by the passenger compartment. This is clearly unacceptable.

Table 2.2 shows the mentioned differences. It also shows the differences for a reversed comparison: a car that is designed for a specific crash velocity in a 30 - 40 per cent overlap crash that becomes involved in a collision with full overlap.

Table 2.2.

Approximate relative deceleration levels in two different crash situations with different design goals.

Designed for Deceleration level at full

overlap crash

Deceleration level at 30 - 40% overlap crash

Full overlap Reference = 100 % ≈ 40 %

30 - 40% overlap ≈ 250 % 100 %

It will be clear that a 2

½

times higher deceleration level leads to much higher forces on the occupant. In the reversed case of a deceleration level which is 2

½

times lower than programmed, the car structure will not be able to absorb enough crash energy. This leads to high intrusion into the passenger compartment and probably to fatal casualties. An optimal crashworthiness requires a weighted design of the body structure for these

(32)

For the choice of a specific partial overlap percentage, the difference in consequences between a 30 and a 40 per cent overlap crash for the occupant must be considered. A 40 per cent overlap crash generates a higher deceleration level and more intrusion through the firewall in front of the occupant then a 30 per cent overlap crash. A 30 per cent overlap crash only leads to a higher intrusion level in the side structure (sill and door) and not in front of the occupant.

In conclusion two extreme different design crash situations can be found, i.e. the full overlap crash and the 40 per cent overlap crash.

2.4. Optimal crash pulses

The stiffness of the available front structure determines the deceleration pulse during a crash. This pulse should have a certain shape, ensuring minimal risk for the occupant. During a heavy collision, there are three important phases:

1. Crash initiation phase. In this phase, the sensor triggering for the belt pretensioner and the airbag must take place. For optimal sensor triggering, the front end of the car should be sufficiently stiff to generate within a short time interval a velocity change that lies above the trigger value of about 6 km/h.

2. Airbag deployment phase. In this phase, the airbag is inflated and the occupant tightens the belts while moving forwards with a relative velocity with respect to the car. This relative velocity should be sufficiently low, because in practice many injuries are the result of reaching a still inflating airbag or hitting the fully inflated airbag with a relatively high velocity. The deceleration of the car should be sufficiently low in this phase, implying that the stiffness must be relatively low.

3. Occupant contact phase. In this phase, the occupant has hit the airbag and there is a stiff contact between the occupant and the car. In this phase high decelerations may occur because the occupant will not be subjected to further shock loads caused by contacts with the interior. The frontal car structure should be stiff enough to decelerate substantially in the remaining time.

Research (Brantman 1991) has shown that for optimal occupant safety in a collision with 48 km/h impact velocity, the first phase lasts between 10 and 30 ms, the second phase lasts 35 ms and the last phase fills up the remaining time to a total of maximal 90 ms. In the first and second phase, the optimal relative velocity values are 8 km/h each. Figure 2.4 shows a crash pulse as a function of time against a rigid wall with

(33)

full overlap optimized for low injury values (HIC and Chest-G, see Section 5.2). This curve is achievable for a large number of current cars (Brantman 1991).

Figure 2.4. Achievable optimal crash pulse at 48 km/h against a rigid wall.

Figure 2.5 shows the three phases of a collision with impact velocity of 56 km/h in a velocity-deformation graph, calculated with the preceding graph (already optimized crash) but adjusted for the higher velocity. Because higher velocities do not significantly change the time duration (Faerber 1991), the same crash initiation time of 15 ms and an airbag deployment time of 35 ms are assumed. The crash duration is 90 ms with a total deformation length of 78 cm.

(34)

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 deformation length [cm] 0 5 10 15 20 25 30 35 40 45 50 55 60 ve lo c it y [ k m /h ] 9g 21g 24g

phase 1 phase 2 phase 3

Figure 2.5. Deceleration level during an optimal frontal deformation at 56 km/h.

Figure 2.6 shows two additional characteristics on both sides of the graph of Figure 2.5. The upper graph is an example of an offset (40 per cent overlap) crash and the lower graph is an example of a full overlap crash with the same vehicle. The two extreme characteristics are positioned as close as possible around the graph of the optimal (lowest injury values) crash situation. In this case the vehicle has a weighted design, the average deceleration of two extreme crash situations is optimal. In fact, a vehicle must be made stiffer as the optimal stiffness for a full overlap crash, to be sure it has enough stiffness in case of an offset collision. The deceleration levels represented by the upper graph are in the first two phases 50 per cent of the deceleration levels of the lower graph (14 g versus 28 g and 6 g versus 12 g) and in the third phase only 30 per cent (11 g versus 36 g). The reason is that the first two phases are comparable with the first crash duration phase mentioned in Table 2.1. The third phase is comparable with the second half of the crash duration. The deceleration level in each phase must change on the same deformation length. If the available deformation length of the stiff car structure in the lower graph is used, in this example at 64 cm, the deceleration level of the 40 per cent overlap crash should not be 30 per cent of the full overlap deceleration level any longer. It can be 36 g,

(35)

which is as high as acceptable in the full overlap crash (Brantman 1991). This choice results in an 18 cm longer deformation length in case of a 40 per cent overlap crash. This remaining deformable structure (not used in a full overlap collision) must be very stiff to absorb much energy in a very short length. The total deformation length in this example is 82 cm, which may be acceptable for a 40 per cent overlap crash of middle class cars at 56 km/h. 0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0 4 5 5 0 5 5 6 0 6 5 7 0 7 5 8 0 8 5

deformation length [cm]

0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0 4 5 5 0 5 5 6 0

v

e

lo

c

ity

[k

m

/h]

Optimal

40%

100%

6g

9g

12g

14g

28g

21g

11g

36g

24g 36g offs et full

Figure 2.6. Deceleration level during a frontal deformation in three cases.

2.5. Representative crash tests as a design goal

The above-mentioned extreme collision situations represent two different cases: A. A stiff front structure yields high deceleration forces.

B. A supple front structure yields a high chance on intrusion into the passenger compartment.

(36)

The differences can be even larger if case A is a full overlap collision against a rigid wall and case B is a 40 per cent overlap collision against a deformable barrier. In case of a crash against a deformable barrier, it is difficult to line up the stiff parts with the supple parts. Hence, the passenger compartment may be loaded before the stiff parts have absorbed enough energy.

These extreme cases A and B form a very challenging design goal because a compromise (a weighted design) must be found. Figure 2.6 can be used as a design guide to define the stiffness of the front structure. In fact, all other less critical collision situations must lie between the extreme curves. Hence, the deceleration level of a certain collision can be found by interpolating (i.e. summarizing the involved vehicle parts) between the mentioned graphs and will be situated even closer to the optimal deceleration level.

The only way to achieve the situation that all new cars have an acceptable crash behavior in different but realistic frontal crash situations is to define two compulsory tests that represent the two mentioned extreme cases as is done by the European Union since October 1998, see next section. If the test results for the most extreme collision situations are acceptable, the other less critical situations will naturally show good results because in the two compulsory tests the difference in the available structure stiffnesses is maximal. Case A can be considered as a deceleration test, important for belts and airbags. Case B can be considered as a structural test to ensure that intrusion is avoided and sufficient protection is given against irregular deformation. If the speed for the full overlap test against the rigid wall is set to 56 km/h, the test will be even more realistic. The speed for the other test of 40 per cent overlap against the deformable barrier must be a little bit higher to compensate for the energy absorption in the barrier.

2.6. Overview of actual and expected legal test requirements

Prescribed research has contributed to the ideas and proposals for a new additional crash test in the European Union (TÜV Rheinland 1992). In Europe a new additional test requirement called EU Directive 96/79 EC is developed and is effective since October of 1998 for new types and models of vehicles, and October of 2003 for all new vehicles. It is an offset collision, involving only 40 per cent of the frontal structure of the vehicle, into a fixed deformable barrier with 56 km/h. A general summary of the current test requirements in the United States is given in Table 2.3 and for the European Union in Table 2.4. Also the additional more severely requirements of the NHTSA’s (National Highway Traffic Safety Administration) New Car Assessment

(37)

Program (NCAP) consumer information program are mentioned. Consumer groups have with their crash test publications a large influence on the vehicle customers by answering the question which car must be bought for more safety and therefore a large influence on car manufacturers to increase the vehicle safety. The European consumer organizations use the Euro NCAP requirements.

Japan and Australia have safety requirements that are similar to requirement FMVSS No. 208 of the United States. This standard is most effective in preventing head, femur and chest injuries and fatalities. However, it does not directly address lower limb and neck injuries and it does not produce the vehicle intrusion observed in many real world crashes. The EU directive 96/79 EC has additional test dummy injury response criteria, particularly for the neck and lower limb. In the EU 96/79 EC offset test the lower extremities are loaded more. NHTSA and also Japan and Australia are currently assessing the additional safety benefits of adopting a supplemental regulation similar to the EU 96/79 EC standard.

Table 2.3.

Frontal crash test requirement in the United States.

Requirement FMVSS No. 208

impact speed 48 km/h (NCAP 56 km/h)

impact object (obstacle) fixed rigid barrier

vehicle place and direction full frontal perpendicular and (not for

NCAP) angles between +/- 30 degrees

dummy type and conditions unrestrained and belt restrained (NCAP),

50th percentile Hybrid III adult male

injury criteria HIC 1000

chest deceleration 60 g chest deflection 50 mm femur force 10000 N

(38)

Table 2.4.

Frontal crash test requirements in the European Union.

Requirement 74/297 EC 96/79 EC

impact speed 50 km/h 56 km/h

( Euro NCAP 64 km/h)

impact object (obstacle) fixed rigid barrier fixed deformable barrier

vehicle place and direction full frontal perpendicular 40 % overlap of the vehicle

width directly in line with barrier face

dummy type and conditions no dummies belt restrained,

50th percentile Hybrid III adult male

injury criteria or structure criteria

steering wheel intrusion horizontal and vertical

direction 127 mm

HIC 1000

chest deceleration 60 g chest deflection 50 mm femur force 10000 N additional criteria on chest

(viscous), the neck, the knee, lower leg bending, foot/ankle compression and

intrusion of compartment

The 50th percentile (mid-sized) male test dummy represents the mean of an adult

male as specified for the total age group. Extending the test requirements to address the effects of occupant size on injuries could be expected. Especially the small 5th percentile adult female dummies may be more at risk by interaction with the airbag because they are in closer proximity than larger occupants (Park 1998).

Another important issue is the compatibility of vehicles. There could be adverse effects on vehicle fleet compatibility after structural changes. A vehicle which has a stiffer or more aggressive front structure for his own increased frontal safety could be more dangerous for another car, especially if that other car is involved in a side impact crash. Also the use of the same fixed deformable barrier for light and heavy cars could lead to less compatibility in crashes between small and large cars. The amount of energy absorbed by the barrier is for a light car a larger proportion of the total crash energy as for a heavy car. To achieve a level of performance comparable to a small car, the front structure of the large car must be designed to crush more or to crush at a higher force level to absorb the additional energy. It is possible that a

(39)

small car becomes softer because a lot of its energy was absorbed by the barrier. This is another reason why a second rigid barrier test is important. The increased crash velocity by Euro-NCAP from 56 km/h to 64 km/h has also a negative influence on the compatibility. This velocity increase yields a 30 per cent higher amount of crash energy. That means that for the same deformation length the force level and thus the stiffness of all cars has to grow with 30 per cent. This effect increases the absolute difference in force levels between light and heavy cars, which deteriorates the compatibility. Otherwise the test velocity must be higher as where collision statistics ask for, because for a comparable vehicle deformation as in a car to car crash the initial kinetic energy must be higher to compensate the absorbed energy in the barrier. Another interesting test for the compatibility problem is a test with a moving deformable barrier. Such a test simulates much better collisions between cars and could improve the fleet compatibility. In this case the smaller vehicle is subjected to a harsher crash environment due to the higher energy absorption and a higher velocity change yielding a stiffer structure. On the other hand the large car would be subjected to a less severe crash environment in terms of velocity change, so a softer front structure gives a temperate crash pulse.

There could also be expected in the European Union a requirement on the compatibility with pedestrians. The large difference in mass between a vehicle and a pedestrian requires a soft front structure (bonnet, headlights, bumper) for easy deformation to lower the acceleration forces on the human body. Therefore the vehicle front must be made larger to cover the necessary stiff vehicle components.

2.7. Conclusions

For the basic crashworthiness of vehicles two extreme frontal collision situations have been defined to be used as representative tests (and since October 1998 also compulsory tests in the European Union) for the most severe cases within the actual collision spectrum, showing a very different yet realistic crash behavior. If a car shows acceptable results in these two tests, it can be expected that the car has also acceptable crashworthiness in other frontal collision situations. Although there are much more collision configurations (e.g. 30 degrees collisions, compatibility between cars or with pedestrians, other dummy sizes), the mentioned two extreme crash situations have the largest impact on the vehicle structure design and this essential design problem must first be solved before further optimization with additional tests for less demanding crash situations is useful.

(40)

Chapter 3 Numerical Design of Stable Energy Absorbing Longitudinal Members

3.1. Introduction

A car has to pass the compulsory crash test as issued by the authorities. However, this regulatory test cannot guarantee that a car will be safe in all likely crash situations. Hence, the total collision spectrum within which the car must be safe must be investigated.

For the design of a car structure that has to sustain a frontal collision, multiple aspects must be considered, i.e. collision velocity, crash direction, overlap percentage, and obstacle type. The energy dissipation in a frontal crash normally occurs by deformation of the longitudinal members. These must absorb a large part of the kinetic energy. In the present chapter, the design spectrum for a longitudinal member is further analyzed and a set of design requirements is formulated. With the aid of Finite Element Method (FEM) models of several longitudinal cross-sections, numerical simulations have been executed to evaluate the influence of the design parameters on the crash behavior. Based upon the results of these simulations, an advanced geometric form for a longitudinal member is presented.

3.2. Simulation parameters as design requirements

To enable the geometric design of the longitudinal members, it must first be clear what kind of loads can occur in real frontal collisions. These loads are determined by the parameters describing the frontal crash and their predominant values (see Section 2.2). Besides the collision speed, the obstacle type, the impact location and direction, also the vehicle mass has a large influence on the crash behavior. An average car with two occupants and luggage has a mass of approximately 1100 kg.

The goal for these simulation studies is to find a single geometric profile optimally suited for all the mentioned parameters rather than a profile that is optimized for a single specific crash situation.

(41)

To limit the amount of simulations, the following simplifications were chosen in relation to the above mentioned parameters:

1. Two collision speeds, viz. 28 and 56 km/h.

2. Obstacle type: only rigid walls, although a deformable barrier is more realistic, it has a disadvantage in comparing energy absorption’s of different geometry’s.

3. Two extreme overlap percentages, to be sure a crash load will lead to acceptable energy absorption in extreme crash configurations:

a. Full overlap (two longitudinal members and the engine are loaded)

b. 40 per cent overlap (only one longitudinal member is loaded) (Witteman 1993). 4. Three impact directions: 0, 15 and 30 degrees.

5. Two vehicle masses: 550 and 1100 kg. The effective mass for a single longitudinal member depends on the specific impact location:

In the extreme crash configuration of a full overlap collision, the effective mass changes from half of the entire vehicle mass (at the start of the crash the load is distributed onto two longitudinal members) into less than one third of the entire vehicle mass (the engine can take a considerable part of the load when it hits the barrier (Hobbs 1991)). To assume a constant cross-section of the longitudinal member, the simulations are confined to the first 350 mm from the front end. Under this condition a much simpler comparison between the various concepts is possible. At this specific length, the influence of the engine need not be considered. Hence, the effective mass is equivalent to half of the entire vehicle mass, i.e. 550 kg.

In the extreme crash situation of 40 per cent overlap, the effective mass equals the entire vehicle mass, i.e. 1100 kg (mostly the engine does not hit the barrier with high frontal peak loads in this situation (Justen 1993)).

The mentioned values were used as input for the crash simulations: one longitudinal member against a rigid wall with an added mass of respectively 550 and 1100 kg, at a speed of 28 and 56 km/h and with a load direction of 0, 15 and 30 degrees.

An optimized geometric profile must have the highest energy absorption per unit of mass for the total range of mentioned crash situations, with acceptable energy absorption for any individual crash situation. This is possible if one can guarantee a stable folding pattern for each parameter combination. Therefore, a good combination of profile thickness and perimeter for a specific shape is necessary. The following basic cross-sectional shapes have been studied: square, rectangle, circle, hexagon and

(42)

octagon. To enable objective comparisons, the mass per unit of length is kept constant. Before the research is described, first a short overview is given of research done by others on crushing columns.

3.3. Research overview of crushing columns

In the past there has been done a lot of research on crushing columns. Especially fundamental theoretical research on the mechanics of thin-walled structures is done by Wierzbicki (1983,1989), Abramowicz (1989) and Jones (1983), where relationships for the column width, wall thickness and shape geometry on the energy absorption’s for different folding modes are derived. Wang (1992) and Yuan (1992) made analyses only on circular tubes, which could have different folding modes. For rectangular tubes experiments and geometrical folding analyses are done by Kim (1996). The influence of the used material of the column on the energy absorption is also interesting. Especially the strain rate dependency of steel for dynamic collisions (Beermann 1982, Behler 1991, Markiewicz 1996) or the crash behavior of aluminum tubes (Belingardi 1994) or comparison of aluminum and steel (Albertini 1996, Magee 1978, Wheeler 1998) are interesting. Comparison between different cross-section geometry’s is mostly based on experiments with only a few different shapes (Groth 1991, Mahmood 1981). For a total overview of different cross-sections it is difficult to compare results of several publications, because the conditions like profile dimensions and material type are different. Also the shape of rectangular profiles is difficult to compare, in most publications additional flanges or stiffeners are used (Wheeler 1998, Kormi 1995) which have a less stable folding behavior as a basic rectangular shape. Giess (1998) describes a numerical simulation with a square profile where the wall thickness is varied over the cross-section yielding an optimized buckling load. Other research is about experiments on the way of welding box sections (or with adhesives), where the distance between the spot welds has influence on the folding behavior (Nishino 1992, Eichhorn 1984, Barbat 1995) or the influence of triggering (Krauss 1994, Yamaguchi 1985). In general most research is based on only an axial load, while more realistic load cases are with an angle of incidence (Crutzen 1996).

From this literature overview it is clear that a practical research with usable design rules on all the mentioned basic shapes and in perspective of realistic crashworthiness requirements (variation of mass, velocity, shape and load direction) is not available. Especially for oblique load directions too little results have been published, and as already mentioned, it is difficult to compare research results based on different

(43)

properties (material, thickness, load conditions, shape) for making an overview with all desired parameters where profiles can be compared under the same circumstances.

3.4. Simulation results

The material selected for the five mentioned profiles was FeP03 (Euro), a commonly used steel (for specifications see Table 3.11). Over the length of the longitudinal member, the profile thickness was kept constant at 2.0 mm. This value is realistic and generally gives a stable folding pattern. The dimensions of the profiles were chosen to have the same perimeter resulting in a constant mass per unit of length, see Table 3.1. The undeformed length of each profile is 350 mm.

Table 3.1.

Dimensions of five profiles.

profile shape perimeter [mm]

A square 300 = 75 + 75 + 75 + 75

B rectangle 300 = 60 + 90 + 60 + 90

C circle 300 = π × 95.5

D hexagon 300 = 50 + 50 + 50 + 50 + 50 + 50

E octagon 300 = 45 + 30 + 45 + 30 + 45 + 30 + 45 + 30

The results of the simulations (Baaten 1994), which concern a collision with a 56 km/h impact speed, a mass of 1100 kg, and a normal angle of incidence are shown in Figure 3.1. The simulation method used in this research is described in Appendix A. Both extremities of the column have an undeformable plate (rigid body), this is more realistic (normally they have a connection with stiffer vehicle components) and better for mutual comparison and to exclude different end effects. The energy absorption is plotted as function of the deformation length (rather than a function of time) because this facilitates the comparison of different structural design concepts. With the deformation length is meant the shortening of the profile. The simulation was terminated at the moment that the load shows a large increase as a result of reaching the maximum

(44)

Figure 3.1. Energy absorption of five different profiles with a load direction of 0

degrees.

Based upon these five simulations, it can be concluded that a square and a rectangular profile have significantly lower energy absorption than the other three profiles. The octagonal profile absorbs slightly more energy during the deformation than the circular and the hexagonal profiles, which absorb nearly the same amount of energy. The circular profile yields the longest possible deformation length and, hence, is capable of absorbing slightly more energy in its final deformation. This highest energy absorption for a circular profile is in agreement with the observations by other authors (Beermann 1982, Belingardi 1994, Groth 1991). Note that there were no geometrical imperfections used and the folds have to fit between the undeformable extremities.

(45)

Figure 3.2. Five profiles with a different cross-section, deformed with a load direction

(46)

The following simulations again concern a collision with a 56 km/h impact velocity and a vehicle mass of 1100 kg but here with an angle of incidence of 30 degrees. With an oblique load direction, numerical simulations of the cross-section of the rectangular profile are carried out in two orientations: standing (profile code B1) and lying

(profile code B2). It can be expected that in lying position the higher bending

resistance of the profile results in a better energy absorption. To avoid simulations with too much unrealistic bending, because a rigid fixation of the profile is assumed, the calculations were terminated after a deformation length of half of the original length. This is illustrated in Figure 3.3. In Figure 3.4 the simulation results are presented, the energy absorption as function of a maximum deformation length of 175 mm. These six deformed profiles are presented in Figure 3.5.

(47)

Figure 3.4. Energy absorption of six different profiles with a load direction of 30

degrees.

B1

B2

From these simulation results, it can be observed again that the profiles: circular, hexagonal and octagonal perform much better than the square and rectangular profiles. The octagon has the highest energy absorption. Comparison of Figure 3.1 with Figure 3.4 shows that at the common deformation length of 175 mm the absolute amount of absorbed energy with a load direction of 30 degrees averages about 65 per cent of the energy absorption with a load direction of zero degrees. This is also evident from Table 3.2 and Figure 3.6 in which the percentages represent the energy absorption with respect to the energy absorption with a load direction of zero degrees for each profile. Only the rectangle in lying orientation looks less sensitive for variation of the load direction; the energy absorption decreases to 76 per cent with a load direction of 30 degrees.

(48)

SQUARE (A) RECTANGLE STANDING (B1) CIRCLE (C) HEXAGON (D) OCTAGON (E) RECTANGLE LYING (B2)

Figure 3.5. Six profiles with a different cross-section, deformed with a load direction of

(49)

Table 3.2.

Energy absorption of six profiles with three different load directions and equal deformation length.

profile shape energy absorption

[Nm] 0 degrees 175 mm deformation energy absorption [Nm] 15 degrees 175 mm deformation energy absorption [Nm] 30 degrees 175 mm deformation A square 14161 100 % 10939 77 % 8490 60 % B1 rectangle 12300 100 % 10715 87 % 7904 64 % B2 rectangle see B1 11486 93 % 9323 76 % C circle 19082 100 % 15919 83 % 12783 67 % D hexagon 19703 100 % 16505 84 % 13111 67 % E octagon 22147 100 % 16151 73 % 14437 65 %

It is noted that if the simulations for all profiles had been executed to include a longer deformation length the difference in energy absorption between normal and oblique would be even larger because no further buckling of the material is possible under this oblique direction.

The group of simulations at the same velocity, with the same mass but with a load direction of 15 degrees gives, as expected, an energy absorption value between the values presented in the Figures 3.1 and 3.4, see Table 3.2 and Figure 3.6. In this situation, larger differences between the various profiles of the decrease of absorbed energy are shown with respect to the energy absorption with a direction of zero

Improved vehicle crashworthiness design by control of the energy absorption for different collision situations

energy absorption for different collision situations

Improved Vehicle Crashworthiness Design

by Control of the Energy Absorption

Improved Vehicle Crashworthiness Design

by Control of the Energy Absorption

for Different Collision Situations

PROEFSCHRIFT

Summary

Samenvatting

Table of contents

Chapter 1

General Introduction

Chapter 2

The Necessity of Improved Crashworthiness Design

½

½

½

½

deformation length [cm]

v

e

lo

c

ity

[k

m

/h]

Optimal

40%

100%

9g

12g

14g

28g

21g

36g

Chapter 3

Numerical Design of Stable Energy Absorbing Longitudinal Members

B1

B2