finite element analysis of a transit bus - wit press · preprocessor software. the finite element...

Finite element analysis of a transit bus

H. Li, R. Nimbalkar, L. Kwasniewski & J. Wekezer FAMU-FSU College of Engineering, Tallahassee, FL, USA

Abstract Most of the bus safety standards in the USA are not applicable to cutaway buses for which a production process is split into two stages. First, the chassis and cab section are assembled by automobile manufactures. Then the vehicle is shipped to another company, where bus body and additional equipment are installed. Lack of strict structural standards for transit bus body builders stimulates the need for crashworthiness and safety evaluation for this category of vehicles. Such an assessment process is needed and important since transit buses are often used to transport disabled passengers. Although a full scale crash test is considered the most reliable source of information regarding structural integrity, crashworthiness and safety of motor vehicles, the high cost of such tests and difficulties in collecting data result in an increasing interest in the analytical and computational methods, which allow for extensive safety studies once the finite element model is validated. This study focused on a selected transit bus, the Ford Eldorado Aerotech 240. Due to the lack of design data the reverse engineering process was used to acquire the geometric data of the bus. The finite element (FE) model was developed based on the geometry obtained by disassembling and digitizing all major parts of the actual bus. The FE model consists of 73,600 finite elements, has 174 defined properties (groups of elements with the same features) and 23 material models. All parts are connected using different multi point constraints and special links with failure to model actual types of structural connections such as bolts and spot welds. LS-DYNA non-linear, explicit, 3-D, dynamic FE computer code was used to simulate behavior of the FE model under different impact scenarios, such as front impact and side impact of two buses at various velocities.

Structures Under Shock and Impact VIII, N. Jones & C. A. Brebbia (Editors)© 2004 WIT Press, www.witpress.com, ISBN 1-85312-706-XStructures Under Shock and Impact VIII, N. Jones & C. A. Brebbia (Editors)© 2004 WIT Press, www.witpress.com, ISBN 1-85312-706-X

1 Introduction

Transit buses are often made by at least two major auto manufacturers. While the first makes the chassis and cab section, the second installs bus body and assembles additional equipment. This process lacks a comprehensive safety assessment and is not regulated (in contrast with school buses) by strict federal safety standards. Such an assessment process is needed and important since transit buses are often used to transport disabled passengers. Two methods are available for crashworthiness and safety evaluation: full scale testing and finite element analysis (FEA). Although a full scale test is considered the most reliable source of information regarding crashworthiness and safety of vehicles, high cost of such tests and difficulties of collecting sufficient data result in increasing interest in finite element method. This efficient and cost effective tool [1] enables extensive investigations once the finite element model is validated. Advance of computer technology makes it feasible to accurately simulate the structural response during a collision using a detailed model with a large number of elements within hours or even minutes [2]. LS-DYNA finite element code is the leading analytical tool in the automotive industry worldwide due to its wide array of element types and material models, its different contact-impact algorithms, and multi point constraints [3] [4]. This paper presents some experiences, findings and conclusions from the research conducted to evaluate the crashworthiness of a selected transit bus through computer impact analysis (virtual crash testing).

2 Procedures

Crashworthiness and safety of Ford Eldorado Aerotech 240 transit bus were examined during this research. Since blueprints and design data of the bus are usually not available, the process called reverse engineering [5] had to be adopted to acquire geometric data and to develop the FE model for the computational mechanics analysis. The actual bus was carefully disassembled into individual parts to allow for accurate geometric data acquisition. All structural components were taped, scanned, digitized, and mapped into computer. In addition to geometric entities such as surfaces, curves and points, material properties like thickness, material type, and weight were also collected. Subsequently, the scanned geometry for each part was imported into preprocessor software. The finite element model of each structural part was developed from the scanned geometric data using graphical preprocessor MSC PATRAN. All parts were assembled to one complete finite element model of the bus. Analysis was conducted by using LS-DYNA and the results were visualized using post-processor LS-POST.


398 Structures Under Shock and Impact VIII

3 Vehicle tear down and digitizing

A FaroArm digitizing equipment and the accompanying software AnthroCAM were the major tools used to obtain the geometric data of the bus components. It allowed for digitizing of points, scanning curves and construction of polylines [6]. Before tear down process began, the global coordinate system was established by setting up about 750 reference points and recording their coordinates using FaroArm (Figure 1). Three to ten reference points were digitized for each structural part. These points were used to establish the position of the part in the global coordinate system after it was removed from the bus.

Figure 1: Defined coordinate system and reference points. All structural components were identified, labeled and removed from the bus and subsequently scanned. Figure 2 presents the disassembled front part of the bus and the scanned curves.

Figure 2: Front part of the bus and the scanned curves. After the non-structural components were removed, the structure of the bus was thoroughly studied. This was helpful to build the FE models of bus components and the entire bus model. Disassembly resulted in full exposure of the connections of the major components. Joints among the structural parts, such as hinges, rivets, welds, bolts, and rubber pads were identified and were appropriately modeled on the computer, using multipoint constraints (MPCs), spot welds, node merging and node tying. Some deterioration of components and


Structures Under Shock and Impact VIII 399

manufacture flaws were found after the disassembly. These defects could adversely affect the performance of the bus, and in some cases, should be accounted for in the FE model, by reducing material properties.

4 Development of FE model

The scanned geometric data were imported into MSC/PATRAN [7], a graphical preprocessor, in which FE meshes were constructed and modified. The numerous options of PATRAN made it convenient and efficient to generate FE mesh from actual, complex geometric shapes and to refine an existing coarse mesh whenever needed. Decision regarding: element formulations, material models, material characteristics, contact algorithms, MPCs and connections, loading and boundary conditions, solution parameters and others, was made to complete the model [8] [9] and to set up crash scenarios. Some laboratory tests were conducted for selected structural components and material samples to identify material parameters and connection characteristics. The models of components were assembled together by several methods, one of which was MPCs. By defining the relationship between dependent degrees-of-freedom and the response of independent degrees-of-freedom, MPC provided the opportunity to model bolts, screws and welds with failure [3]. One example of MPCs was the modeling of spot weld in the bus cage [10], a common feature in automotive manufacturing process. Different MPCs between rigid rod and beam elements were also used to model connections in front and rear suspension systems and to model hinges in connections of all doors. Contact describing the interaction between the adjoining parts occurs frequently in transient dynamics systems [4]. Different contact algorithms, such as self contact, master-slave contact, were applied to the model. Since contact algorithms are computationally demanding and can increase run time dramatically, they were defined only for limited groups of elements being involved in interaction. 23 material types were identified for the structural components of the actual bus structure. The bus body was built using two layers of composite material with additional honeycomb layer placed in between them. Composite layers were modeled using shell elements and honeycomb was represented by solid elements. The final FE model of the bus consisted of 174 parts, 23 material models, and 73,595 elements. Summary of the final FE model of the bus is provided in Table 1. A large number of the vehicle components were modeled with shell elements since most of the structural parts of the bus were made of metal and composite sheets. A fully integrated quad element number 16, available in LS-DYNA, was selected for analysis as the most reliable element formulation, based on several numerical tests [10]. The final FE model of the bus is presented in Figure 3.



Table 1: Summary of the final FE model of the bus.

No. Entity Number

1 Number of parts (LS-DYNA)/ Property sets (PATRAN) 174

2 Number of material models 23 3 Number of nodes 67,788 4 Number of solid elements 9,612 5 Number of shell elements 63,271 6 Number of beam elements 712 7 Total number of elements 73,595

Figure 3: FE model of the transit bus.

5 Crashworthiness analyses

Several impact scenarios were studied to investigate crashworthiness and passenger safety of modified body-on-chassis buses. In this research, the most common type of accident events such as: frontal impact at 30 mph, side impact at 30 mph, and side impact at 50 mph velocity were studied in detail. As an example, the side impact of two identical Ford Eldorado buses, at a velocity of 30 mph, is presented below. The initial position of vehicles is shown in Figure 4. The initial velocity was applied to the right bus, which served as an impacting vehicle. The simulation started 0.33s before the vehicles came to a contact, to allow them to settle down under gravity, and to induce initial deformations and forces in springs and dampers of the suspension system. The initial distance between the front bumper of the right vehicle and the street wall of the left bus was assumed as 4.2 m (13’ 9”). Figure 5 shows an animation sequence for side impact simulation.



Left bus Right busFront view

48 km/h (30 MPH)

Figure 4: Initial position of buses.

B0

B2

B3

B4

Figure 5: Sequence of pictures showing both buses at characteristic time points.

Figure 6 shows change in the total kinetic energy with time. Points B0-B4 depict characteristic time points on time history curve. The initial total kinetic energy is 350.4 kJ (476.7 ft-lb). This value corresponds to the initial velocity of 30 mph (13.33 m/s) applied to the right bus. According to Figure 6 wheels of the right bus touched the road after about 0.07 s. During this initial period, presented in Figure 6 as a horizontal segment B0-B1, impacting bus moved (“levitated”) above the road. In the next B1-B2 phase the interaction between wheels of the moving vehicle and the road caused a loss of kinetic energy of about 5.2%. At the time 0.33 s (point B2) the right bus impacted the second vehicle. During the short period of about 0.06s, depicted as the segment B2-B3in Figure 6, 46.6% of the total kinetic energy (at the point B2) was transformed into strain energy due to deformation in the impacting zone. Figure 7 shows time history of total strain energy in the impacting zone



B0 B1 B2

B3

B4

0.070.33

0.39 [s]

[Nmm]

Figure 6: Time history of total kinetic energy (values in Nm).

B0

B1

B2

B3 B4

0.070.33

0.39 [s]

[Nmm]

Figure 7: Time history of total internal (strain) energy.

which matches with the loss of kinetic energy shown in figure 6 as the segment B2-B3. Most of kinetic energy was dissipated due to inelastic deformation in the impacted zone (street wall) in the left bus. Figure 8 shows deformation and damage in the street wall at the end of simulation at the time 1.2 s. The impacting bus was erased from the view. Figure 9 shows deformations in the steel cage and floor, at the same time. Windows, driver doors, and body were removed from the view. Figure 10 presents contours of effective stress (von Misses stress) for steel cage and the floor at the moment when they were the largest, at the time t=0.37s. Both pictures show that most of the impact was concentrated at the floor level, where the bus structure was relatively stiff. Figure 11 shows time history for relative lateral displacement for one of the nodes in the wall, in the impact area. The relative displacement was defined as a difference of the total displacement of selected node and the displacement due to rigid body movement. Maximum elastic – plastic penetration was found to be 300 mm at the time t=0.39 s. The inelastic maximum penetration was about 200 mm. Relatively limited damage in the bus body resulted from this crash at a speed of 30 mph, as shown by analysis. This conclusion is valid only for the material properties applied.



Figure 8: Inelastic deformations in the impacted bus, t=1.2 s.

. Figure 9: Deformations in cage and floor, t=1.2 s.

Figure 10: Contours of effective stress, t=0.37 s.



B0 B1 B2

B3

B4

0.070.33

0.39 [s]

[mm]

Figure 11: Relative lateral displacement in the street wall.

6 Summary and conclusions

Presented research was carried out to investigate the crashworthiness and passenger safety of modified body-on-chassis buses. The study was focused on a Ford Eldorado Aerotech 240 transit bus, which served as a sample representing a wider range of similar transit buses. The major goal of this research was to develop an effective tool for collecting data useful for future development of recommendations for bus body manufactures and operators. High cost of actual crash test experiments and difficulties with collecting extensive data from full-scale crash tests result in increasing interest in analytical and computational methods. Also, damage and failure mechanisms, especially those with sequential, complex and progressive failure could only be fully understood from computational mechanics and crash analysis. Material properties for the model were obtained from several sources available. Most of the material properties were adopted from other FE models of vehicles developed by the National Crash Analysis Center [11] and other university communities. Limited laboratory testing was conducted at FAMU-FSU College of Engineering to determine material properties. Other material properties were selected from common engineering handbooks, which often provide only a range of material properties. Results of this study highly depend on material properties used and hence are valid only for the selected material properties. Test coupons, especially for body wall structure (honeycomb and laminate layers), could provide exact material properties and the results obtained from crash simulations would represent the actual bus and they would increase the credibility and value of this analysis. In addition, simple laboratory tests of selected component connections could help evaluate their strength and failure criteria. A full-scale crash test of Ford Eldorado bus is necessary to validate the entire finite element model. Although the finite element model was not validated, it is



complete in the sense that it represents all the structural parts of the actual bus, their geometry and mechanical properties and connections between them. Numerical data presented shows that the FE model yields realistic results and, with possible modifications and improvements (depending on laboratory test data), can accurately represent the behavior of the actual bus. Results from computational mechanics analysis presented in this study provide detailed histories of stress, strain and other state variables. The study offers a valuable insight into the structural response of a bus impacted by another vehicle. The results show the potential for thorough evaluation of the bus crashworthiness and safety of its passengers. The existing model can be easily converted and modified to reflect potential changes in the body structure of new buses for detailed parametric studies. Comparison of results for models with different material properties or structural components can assist engineers in developing recommendations regarding optimal and safer design, and to establish new standards for bus body builders.

References

[1] Bathe, K-J. Crush simulation of cars with FEA. Mechanical Engineering, ASME, 1998, http://www.asme.org/pdf/.

[2] Li, G. Scalability of LS-DYNA on SGI systems. SGI Developer News, December, 2000.

[3] LS-DYNA Keyword user’s Manual (Nonlinear Analysis of Structures), Livermore Software Technology Corporation: Livermore, California, May, 1999.

[4] LS-DYNA Theoretical Manual, Livermore Software Technology Corporation: Livermore, California, May, 1998.

[5] Chenga, Z.Q., Thackera, J.G., Pilkeya, W.D., Hollowell, W.T., Reagana, S.W. & Sieveka, E.M. Experiences in reverse-engineering of a finite element automobile crash model. Finite Element Anal, 37, pp. 843–860, December, 2001.

[6] FARO ARM ANTHROCAM. Design Basic Training Workbook, Version 2.3/2.5, Faro Technologies Inc: Lake Mary, Florida, July, 1999.

[7] MSC/PATRAN Reference Manual, MSC. Software Corporation: Santa Ana, California, 2001.

[8] Bathe, K-J., What can go wrong with FEA? Mechanical Engineering, ASME, 1998, http://www.asme.org/pdf/.

[9] Omar, T.A., Kan, C.D., Bedewi, N.M. & Eskandarian, A. Major parameters affecting nonlinear finite element simulations of vehicle crashes. Crashworthiness, Occupant Protection and Biomechanics in Transportation Systems - (The ASME International Mechanical Engineering Congress and Exposition), Nashville, TN, USA, 1999.

[10] Kwasniewski L., Wekezer J.W., & Li H., Development Of Finite Element Model For Ford Eldorado Transit Bus, Theoretical Foundations of Civil Engineering, 10th Polish-Ukrainian Transactions, Warszawa, June, 2002.

[11] National Crash Analysis Center, web page- http://www.ncac.gwu.edu.



finite element analysis of a transit bus - wit press · preprocessor software. the finite element...

Documents