Crashworthiness Performance of Mass-Efficient Extruded Structures

Theoretical, numerical and experimental studies were conducted on the axial crushing behavior of traditional single-cell and innovative four-cell extrusions. Two commercial aluminum alloys, 6061 and 6063, both with two tempers (T4 and T6), were considered in the study. Testing coupons taken from the extrusions assessed the nonlinear material properties. A theoretical solution was available for the one-cell design, and was developed for the mean crushing force of the four-cell section. Numerical simulations were carried out using the explicit finite element code LS-DYNA. The aluminum alloy 6063T4 was found to absorb less energy than 6061T4, for both the one-cell and four-cell configurations. Both 6061 and 6063 in the T6 temper were found to have significant fracture in the experimental testing. Theoretical analysis and numerical simulations predicted a greater number of folds for the four-cell design, as compared to the one-cell design, and this was confirmed in the experiments. The theoretical improvement in energy absorption of 57% for the four-cell in comparison with the one-cell design was confirmed by experiment. The good agreement between the theoretical, numerical and experimental results allows confidence in the application of the theoretical and numerical tools for both single-cell and innovative four-cell extrusions. It was also demonstrated that these materials have very little dynamic strain rate effect.

Rails Deformation Pattern in Frontal Impact

During a vehicle crash stress waves can be generated at the impact point and propagate through the vehicle structure. The generation of these waves is dependent, in general, on the crash type and, in particular, on the impact contact characteristics. This has consequences with respect to different crash barrier interfaces for vehicle evaluation. The two barriers most commonly used to evaluate the response of a vehicle in a frontal impact are the rigid barrier and the offset deformable barrier. They constitute different crash modes, full frontal and offset. Consequently it would be expected that there are different deformation patterns between the two. However, an additional possible contributor to the difference is that an impact into a rigid barrier generates waves of significantly greater stress than impacts with the deformable one. If stress waves are a significant component of real world final deformation patterns then, the choice of barrier interface and its effective stiffness is critical. To evaluate this conjecture, models of two types of rails each undergoing two different types of impacts, are analyzed using an explicit dynamic finite element code. Results show that the energy perturbation along the rail depends on the barrier type and that the early phase of wave propagation has very little effect on the final deformation pattern. This implies that in the real world conditions, the stress wave propagation along the rail has very little effect on the final deformed shape of the rail.

Use of Stochastic Analysis for FMVSS210 Simulation Readiness for Correlation to Hardware Testing

The FMVSS210 regulation establishes requirements for seat belt assembly anchorages to be strong enough for effective occupant restraint. The belt separation from the vehicle structure in crash tests needs to be avoided. Federal government mandate requires use of Pelvic and Torso Body Blocks for testing belt anchor strengths for lap and shoulder belts respectively. The belt anchorages are expected to withstand loads of 13.34 kN if both lap and shoulder belts are used and 22.24 kN if only lap belts are used. The analytical simulation of the hardware test is done using explicit dynamic code LS-DYNA. Hardware testing is of quasi-static nature while the simulation uses the dynamic code. However the analysis could be made to approach the quasi-static test by adjusting some input parameters in the simulation. In addition some input parameters need adjustment for making the model robust and to make it correlate to the hardware test. This study involves the use of Optimal Symmetnc Latin Hypercube Design to explore the design space, and to develop a fast surface response model. This response model can be viewed as a surrogate model to the actual LS-DYNA simulation and is used in this work to rank the input parameters by the percent contdbution they make towards the variation of the desired output responses. After determining the fit of the response model, it is used to perform the stochastic simulation. The confidence interval for test correlation prediction can then be estimated. This technique can further be used to do design sensitivity studies and for optimizing the vehicle structure with respect to FMVSS210 regulation.

Locomotive Crashworthiness Testing at the Transportation Technology Center

This paper describes a series of full-scale impact tests to be conducted at the Federal Railroad Administration’s Transportation Technology Center (TTC), Pueblo, Colorado. The tests will be performed to investigate locomotive crashworthiness.

Dynamic Axial Crush of Automotive Rail-Sized Composite Tubes: Part 2 — Tubes With Braided Reinforcements (Carbon, Kevlar®, and Glass) and Non-Plug Crush Initiators

This paper documents the braided reinforcement portion of a successful fundamental study of the dynamic axial crush of automotive rail-sized composite tubes. Braided reinforcements were comprised principally of carbon fiber but also of Kevlar® and E-glass and combinations of the three. Fourteen different braids were used, six of which were tri-axial and the remainder bi-axial. Tubes were manufactured using Resin Transfer Molding (RTM) with processing and molding techniques that are suitable for the low cost high volume needs of the automotive industry. Braids were obtained as continuous rolls of tubular sock-like material and pulled over metal mandrels one ply at a time. Carbon fiber tow sizes ranged from 6k to 48k. Dow Derakane 470 vinylester resin was used for all tubes. Tube geometry, a 88.9×88.9 mm square cross section with 2.54 mm thick walls, approximated that of the first 500 mm of the lower rail of a typical mid-size vehicle. Note in particular that tube wall thickness was fixed at a single value in this study. A 45° bevel on the outside edge of the lead end of each tube served as the crush initiator. In total 71 dynamic axial crush tests were conducted. In terms of important findings, consistent with the woven fabric portion of this program [1], desirable dynamic axial crush response was demonstrated for RTM’d automotive rail-sized carbon fiber reinforced tubes. For almost all parameter configurations, the tubes exhibited stable and progressive crush with a reasonably flat plateau force level and an acceptable crush initiation force, i.e. one that can be withstood by the backup structure. Additionally, crush debris from such tubes was found to neither contain objectionable sharp brittle splinters nor pose a health risk. Displacement average values of dynamic axial crush force ranged from 11.88 to 26.51 kN and values of SEA (specific energy absorption) ranged between 10.42 and 22.44 kJ/kg. In terms of parameter effects, the fiber type and reinforcement architecture were found to be capable of more than doubling/halving the dynamic axial crush force and SEA.

Target Setting for Vehicle Side Impact Using a Statistical Approach

Designing a vehicle for superior side impact performance is an important consideration in automotive product development. The challenges involved in this design are many as side impact is a relatively short duration event and interaction between the crash test dummy and vehicle side structure including side airbag (if present) is involved, and engagement of the moving deformable barrier with the stiffer underbody of a car may not always take place (unlike in the case of frontal NCAP test in which the front rails will necessarily play a major role in energy-absorption). The safety engineer nevertheless has to set targets for relevant side crash-related variables in the initial phase of design when very limited information on vehicle structure is available. As shown in the present study, linear regression-based relationships can be utilized for setting quantitative targets for structural response and packaging-related variables for side impact safety design of a new vehicle.

Emergency-Locking Retractor Performance in Rollover Accidents

Generally accepted accident statistical analyses indicate that seat belted occupants involved in automobile accidents fare far better than those that are not belted. This is especially true for rollover accidents, with the primary reason being that seat belts help prevent ejection of the occupant from the vehicle. Ejected occupants are far more likely to incur serious or fatal injuries than those that remain inside the vehicle occupant compartment. Nonetheless, even belted occupants can be seriously or fatally injured in rollovers. The excursion of belted occupants during rollover accidents has been a topic of research over the past several years. Much work has been reported on the effects of belt anchor geometry. More recently published analyses have looked at the performance of the seat belt retractor in rollover accidents as well as other accident scenarios. One theory, put forth by various analysts, is that the seat belt webbing can “spool-out” from vehicle-sensitive emergency-locking retractors (ELR’s). According to this theory, the “spool-out” mechanism occurs because the retractor may cycle between a locked condition to an unlocked condition as the vehicle is overturning. Seat belt webbing can then be spooled-out from the retractor if the occupant engages the seat belt at a time that the retractor is in an unlocked condition. The added webbing introduced into the seat belt system mitigates the effectiveness of the seat belt during the subsequent roll motion. In this paper, we specifically address the performance of ELR’s in rollover accidents. A detailed analysis of the various phases of a multiple-roll rollover sequence, with an emphasis on vehicle dynamics and occupant kinematics as they relate to the physics of the sensing mass and operation of the retractor spool and locking mechanism(s), is presented. Additionally, the results of full-scale rollover testing are analyzed. The conditions to effect a retractor “spool-out” require that the sensing mass of the ELR must move to a neutral position, and the occupant must move in such a way to release tension in the seat belt thereby allowing webbing to retract back onto the spool. This retraction motion is necessary to release the ELR lockup components from a locked position. After conditions have been achieved, the sensing mass must then remain in a neutral position while occupant moves sufficiently, relative to the vehicle, to withdraw seat belt webbing from the The analysis presented in this paper and the results of testing indicate that the circumstances necessary for retractor spool-out to occur are not present in rollover accidents. A condition where sensing mass of the ELR will remain in a neutral position long enough and coincident with the occupant moving relative to the vehicle in such a manner to withdraw appreciable webbing from the does not occur. The external inputs to the vehicle that induce occupant motion also induce mass motion. The sensing mass need only move fractions of an inch to activate the retractor mechanisms. As a result, the retractor will be locked before webbing can be extracted from webbing spool.

Experimental Study on the Crash Performance of Aluminum and Steel Rails

Increasing government mandated CAFE´ standards are forcing the OEMs to aggressively reduce vehicle weight. Aluminum, with a density of about a third of that of steel, has been established as a viable alternative to steel for the construction of the automotive body structure. However, for aluminum sheet metals, there are still lingering concerns about the reliability and robustness of the available joining techniques such as spot-welding, riveting etc. The investigation reported in this paper was aimed at evaluating the relative performance of self-pierced riveted aluminum rails as compared to spot-welded mild steel and high strength steel rails. A series of straight and curved (S-shaped) rails made of aluminum, mild steel, and high strength steel have been tested. Other design parameters considered in this study include sheet metal thickness, rivet/weld location, rivet/weld spacing, adhesives, temperature, and impact speed. As were observed from the tests, axial crush mode dominated the deformation of all straight rails while bending dominated the deformation of the curved rails. Statistical analysis was performed to find the relative importance and effects of each variable on the average crush load, maximum load and energy absorption. For aluminum rails, the thickness of the sheet metal was found to be the primary controlling factor for both straight and S-rails. Other factors i.e. rivet spacing/location, adhesives, temperature and impact speed, had no significant affect on the performance of the rails. For the steel rails, the sheet metal thickness, impact speed, temperature and material properties, were all found to be significant for the crash behavior. It was also found that the aluminum rails have higher specific energy absorption than the steel rails confirming that aluminum as a material is more efficient in absorbing crush energy than steel.

Eliminating the Unwanted Dynamics From the Controllable Suspensions for Vehicle Applications

This paper presents two alternative implementations of skyhook control, named “skyhook function” and “no-jerk skyhook,” for reducing the dynamic jerk that is often experienced with conventional skyhook control in semiactive suspension systems. An analysis of the relationship between the absolute velocity of the sprung mass and the relative velocity across the suspension are used to show the damping force discontinuities that result from the conventional implementation of skyhook control. This analysis shows that at zero crossings of the relative velocity, conventional skyhook introduces a sharp increase (jump) in damping force, which, in turn, causes a jump in sprung mass acceleration. This acceleration jump, or jerk, causes a significant reduction in isolation benefits that can be offered by skyhook suspensions. The alternative implementations of skyhook control included in this study offer modifications to the formulation of conventional skyhook control such that the damping force jumps are eliminated. The alternative policies are compared with the conventional skyhook control, using a laboratory implementation on a heavy truck seat suspension that represents a base-excited system with a semiactive suspension. An evaluation of the damping force, seat acceleration, and the electrical currents supplied to a magneto-rheological (MR) damper that is used for this study, shows that the alternative implementations of skyhook control can entirely eliminate the damping force discontinuities and the resulting dynamic jerks caused by conventional skyhook control.

Automated Finite Element Modeling of Tube Frame Structure Parametric Design

This research was intended to address the last step in the development of a tube-frame (termed B2B) parametric crashworthiness model - automated finite element modeling of the parametric design. We have added the generation of finite element models to the previously built Unigraphics Version 16 (UG V16) parametric model, so that finite element models could be quickly built. UG/WAVE was used to design the vehicle parametrically and UG/SCENARIO, a pre- and post-processor integrated in UG, was used to automatically construct the finite element mesh. We established the quality of the finite element meshes, generated for two new designs, which were created by changing overall dimensions of the vehicle. This was done using objective criteria for the finite element mesh. The component data was added to the automatically generated mesh, and the results from the crashworthiness analysis of this model compared favorably with the ‘hand-built’ model using traditional model building techniques. The results from this work will be useful in the development of the parametric design process. The use of automatically generated finite element meshes will also be useful for the automated evaluation of these parametric designs.