Influence of Crash Box on Automotive Crashworthiness MIHAIL DANIEL IOZSA, DAN ALEXANDRU MICU, GHEORGHE FRĂȚILĂ, FLORIN- CRISTIAN ANTONACHE University POLITEHNICA of Bucharest 313 Splaiul Independentei st., 6th Sector, ROMANIA daniel_iozsa@yahoo.com; dan.micu@upb.ro; ghe_fratila@yahoo.com; antonache_florin@yahoo.com Abstract: In this paper, frontal impact behaviours of three car frontal parts with a rigid obstacle at rest is presented. The purpose is to analyze the best crashworthiness. The models have different crash boxes and are analyzed using Explicit Dynamics module of Ansys software. Shape and dimensions of the model were obtained from repeated simulations and constant improvements. Finite element mesh size for each part of the model varies, depending on its role. Velocity of the car model was computed by equalizing the kinetic energy of the modelled geometry with the kinetic energy of a considered automobile. The results present a comparison of deformations and stress, resulting an analyze of absorbed energies values during the impact. Key-Words: crash box, frontal impact, crashworthiness, Ansys, deformation, car structure 1 Introduction Crashworthiness is the ability of a structure to protect its occupants in the event of a crash. Frontal impact cars is one of the most often crash types. Automotive manufactures increasingly employ computer simulation, because physical vehicle crash-testing is highly expensive [1]. Currently, dynamic explicit integration is commonly used for the simulations like impact and collision.[] A D concept model of a detailed automotive bumper model was introduced and it was discretized by using lumped mass spring elements in [3]. The time efficiency and the good approximation of results proved its utility in crash analysis, confirming that early stages of product design can make use of the simplifications and rapid decisions can be taken for early improvements. It is useful to utilize mathematical optimization by altering the geometry and the material and structural properties of the bumper- beam and crashbox to improve the low speed performance[4]. When a vehicle impacts in less than 15 km/h velocity, the insurance companies require that the damage of the vehicle should be as small as possible. Section presents the steps necessary to simulate frontal impact. The first step consists in establishing a mathematical model to use in crash analyze of a car frontal part. Three models of crash boxes that belong to geometry of the impact energy management system are described in the second part presented in subsection.. Initial conditions of frontal impact simulations and meshing settings are presented in the last two subsections of section. Variations and comparisons of stress and plastic deformations of the all three models are analyzed in section 3. Simulating frontal impact.1 Study of mathematical models used on impact analyze of a car frontal part Simple or complex mathematic models can be used to study structure dynamics, depending on complexity of simulated phenomena, precision and/or computation rate. Figure 1 shows four of most usual mathematic models used to test bumper beams in impact computations. a. b. c. d. Fig. 1 Usual mathematical models used to test bumper beam in impact computations [5] The mathematic model with one damping element (c 1 ) and one elastic element (k 1 ) in serial communication is the most used (Fig 1.a). One damping element (c ) and one elastic element (k ) in ISBN: 978-960-474-403-9 49
parallel communication is another mathematical model (Fig 1.b). Complex structures or particular situations can be modelled using elastic elements (k 31 ) in parallel communication with a damping element (c 3 ) and an elastic element (k 3 ) in series communication (Fig 1.c), or with a damping element (4) in parallel communication with a spring element (k 41 ), both in series communication with a spring element (k 4 ) (Fig 1.d). An impact of an vehicle can be defined by four cases which are presented in Fig. a. b. c. d. Fig. Typical cases to study the impact of vehicles [5] The first case (Fig.a) is a frontal impact between a moving car and a rigid obstacle at rest. In this case the impact velocity (V e ) and impact energy (W e ) are those of the car: deformable barrier), both moving, is presented in Fig.d. V1 V V e = [km/h] (7) W1 W W e = [J] (8) W W 1 The mathematical model used is the one with elastic and damping elements in series communication (Fig 1.a) and the case to study is the impact of the rigid obstacle at rest by a moving car (case I)(Fig. a).. Modelling geometry of the impact energy management system Geometry modelling was performed using ANSYS, a structural analysis software, and the elements were defined by the surface type. Elements whose geometry is necessary to simulate a frontal impact are: an obstacle, a front bumper beam, crash boxes, flanges, front frame rail and a block representing the car. Figure 3 shows the components used to simulate the frontal impact. V e = V [km/h] (1) W e = W [J] () The second case (Fig.b) is a frontal impact between a moving car and a barrier equipped with a dampening impact energy (equivalent to a deformable barrier) at rest. To study this case the impact velocity (V e ) and the impact energy (W E ) is calculated using formulas: a. obstacle and bumper beam V e = V [km/h] (3) W e = W [J] (4) A frontal impact between a car and a rigid obstacle, both moving, is presented in the third case (Fig.c). Impact velocity (V e ) and impact energy (W e ) can be determined using the following formulas: V e = V 1 +V [km/h] (5) W1 W W e = [J] (6) W W 1 A frontal impact between a car and an obstacle provided with a damping system (equivalent to a b. crash boxes and flanges c. front frame rail and a block representing the car Fig. 3 Elements used to simulate the frontal impact ISBN: 978-960-474-403-9 50
Figure 4 shows the first model of the crash box integrated in the frame rail during the impact with the obstacle. The geometry was been modified by using different crash boxes. The cross-section profile and dimensions of the front cross beam were not been modified during the initial geometric model improvement. A top view of the three modelled geometric solutions for impact energy management system is presented in Figure 6. Fig. 4 Isometric view of first model of the crash box integrated in the frame rail during the impact with the obstacle Shape and dimensions of the model were obtained from repeated simulations and constant improvements. The objective is to obtain a better behavior if the structure is subjected to similar stresses to those that occur in a frontal impact. The model improvement in this phase was obtained by choosing the measure to increase the cross-section of the front frame rail and of crash boxes, by the relative disposition of the vehicle body block so that its center of gravity to be at an usual distance above the assembly and by choosing the front frame rail s curvature radius from the frontal part to the cockpit. The model was chronology developed from model M1, to model M and to model M3, as it can be noticed in Figure 5. Fig. 5 Isometric view of the three modelled geometric solutions for impact energy management system Fig. 6 Top view of the three modelled geometric solutions for impact energy management system Figure 7 presents an isometric view of the geometrical model solutions of crash boxes. Fig. 7 Isometric view of the geometric model solutions of crash boxes (removable ends of the front frame rail) Steels values of the physical parameters of materials were introduced in the analysis software library to model the impact energy management system materials (HSLAS S300MC and S50MC). The material models were saved separately with specific names to be assigned to each component separately. The steel model H.S.L.A.S. S50MC, named "Structural Steel NL 1" in the material library of the software is assigned to crash boxes and model HSLAS S300MC named "Structural Steel NL " is assigned to bumper beam, flanges and to frame rails. The "NL" suffix in the name of the steel refers to the fact that the materials have nonlinear material characteristics to simulate both material behaviours: plastic and elastic. This is necessary because during the simulation, the stress of the components exceed their yield strength. ISBN: 978-960-474-403-9 51
.3 Defining initial conditions to simulate frontal impact Particular conditions, such as rigid contacts with or without friction, fixed supports, pretensions, relative speeds etc., have to be imposed to the model components. These conditions are necessary because the results obtained from the dynamic simulation should behave as close to reality. Two static Bonded type contacts between the left front frame rail and the car and between the right front frame rail and car were established surfing in the "Model" part of the "Explicit Dynamics" module of Ansys software (Figure 8). Fig. 8 Rigid and static contacts established between the front frame rails and the car body box In Connections menu, Body Interactions field, a Frictional type of dynamic contact was established between the frontal cross member and the contact surface of the obstacle (Figure 9). Also, a fixed support was imposed on the outer surface of the obstacle plane farthest from automobile to represent the state of relative rest of the obstacle (Figure 8). The imposed velocity to car assembly was inferred from equalizing the kinetic energies of the modelled geometry and designed automobile as follows: mmod el Vmod el Ec mod el [J] (9) mauto Vauto Ecauto (10) where: E cmodel [J] - kinetic energy of the modelled geometry; E cauto [J] - kinetic energy of the automobile; m model [kg] - mass of the modelled geometry; m auto [kg] - mass of the automobile; V model [km/h] impact velocity of the modelled geometry corresponding to its kinetic energy; V auto [km/h] impact velocity of the automobile corresponding to its kinetic energy. Because: E cmodel = E cautol V mod el mauto Vauto [ km/ h] (11) m mod el According to European regulations regarding frontal impact test, the initial speed of the automotive before impact must be kept constant around 15 km / h ( 4,166 m / s). Fig. 9 Frictional type of dynamic contact between the frontal cross member and the contact surface of the obstacle In Explicit Dynamics module, Initial Conditions part, the initial linear and constant velocity, its direction and its orientation were established for components of both the car and impact energy system group (Figure 10)..4 Meshing geometric model using finite elements The finite element mesh size of each component of the model geometry can be chosen in the "Model" part, "Mesh" menu. Fig. 10 Initial velocity conditions of the simulation components Fig. 11 Meshing the assembly to simulate frontal impact ISBN: 978-960-474-403-9 5
Finite element mesh size for each part of the model varies depending on its role: for crash boxes a mesh as fine (10 mm), for cross member and flanges a larger mesh (15 mm), for front frame rail a large mesh (50 mm) and for car body block and obstacle a coarse mesh (100 mm) (Figure 1). "Generate mesh" button is used. A number of 4747 elements and 490 nodes resulted following the completion of the entire assembly meshing. Fig. 14 Stress variation of geometric model M1 during the impact simulation Fig. 15 Plastic deformation variation of geometric model M during the impact simulation Fig. 1 Finite element meshing of different sizes for each component of the model Table 1 The main parameters of each component used to simulate the frontal impact No 1 Criterion Thickness profile of the cross section [mm] Material Automobile Frame rails Flanges Cross member Crash Box Obstacle -.0.0 1.1 1.0 - Structural Steel Structural Structural Structural Steel NL Steel NL Steel NL Structural Steel NL 1 Structural Steel 3 Mass [kg] 847.80 4.895 0.448 3.148 0.36 56.75 4 Mesh size 100 50 15 15 10 100 5 Velocity [m/s] 4.190 15 km/h 0 Fig. 16 Stress variation of geometric model M during the impact simulation 3 Results The demountable crash boxes deflection should not do flaming but controlled by folding deformation using initiators such as ribs, holes, folds, cuts, different shapes of sections, elements with variable thickness and constant increase of sections and of inertia moments. After modelling the geometry and imposing the initial conditions the "Solve" button is used to run the simulation. The results can be read and save in the "Explicit Dynamics" module, "Solution" part. Fig. 17 Plastic deformation variation of geometric model M3 during the impact simulation Fig. 18 Stress variation of geometric model M3 during the impact simulation Fig. 13 Plastic deformation variation of geometric model M1 during the impact simulation Fig. 19 Stress and plastic deformation variations of geometric model M1 during the impact simulation ISBN: 978-960-474-403-9 53
Fig. 0 Stress and plastic deformation variations of geometric model M during the impact simulation Fig. 1 Stress and plastic deformation variations of geometric model M3 during the impact simulation 4 Conclusion Total plastic deformation growth during the impact, reaches a maximum value and remain quasiconstant around this value (saturate) for all three models. From this moment, it is considered that the impact energy is not consumed any more by the crash boxes, but the energy is sent to the front frame rail. The aim is to consume higher quantities of energy away from the passenger compartment in a short time interval. The amount of transmitted energy to other body parts and/or to passenger compartment should be minimized. It is observed that the model M has the highest strain in the shortest deformation time. A larger deformation implies a higher consumption of impact energy and a less time for this strain is an increased safety for car occupants. Stress is represented from blue to light blue on the surface of crash boxes, and maximum stress appear only in some points. That means the stress values are small. Fig. Comparison of plastic deformation variations of geometric models during the impact simulation Fig. 3 Comparison of stress variations of geometric models during the impact simulation References: [1] Micu, D.A., Straface, D., Farkas, L., Erdelyi, H., Iozsa, M.D., Mundo, D., Donders, S., A cosimulation approach for crash analysis, UPB Scientific Bulletin, Series D: Mechanical Engineering, 76 (), 014, pp. 189-198; [] Micu, D.A., Iozsa, M.D., Stan, C., Quasi-static simulation approaches on rollover impact of a bus structure, WSEAS, ACMOS, Brașov, June 6-8, 014; [3] Sîrbu, A.D.M., Research on improving crashworthiness of the frontal part of the automotive structure, PhD Thesis, POLITEHNICA University of Bucharest, Romania, 01; [4] Redhe, M., Nilsson, L., Bergman, F., Stander, N., Shape Optimization of Vehicle Crash-box using LS-OPT, 5 th European LS-DYNA Users Conference, Birmingham, 005; [5] Donald Malen, Fundamentals of Automobile Body Structure Ddesign, SAE International, 011. ISBN: 978-960-474-403-9 54