Search Results

Given two or more photos of an accident vehicle (non-stereo pairs such as police photos) an estimate of the deformation (crush) of the vehicle may be obtained by application of camera reverse-projection, using two or more cameras and an exemplar vehicle. A single camera technique familiar to accident scene investigators is modified for this application. The methodology is described within the context of an experiment comparing results obtained by camera reverse projection to actual measured crush. A method of displaying crush results known as “displacement vectors” is presented and examples are illustrated. The technique has been found useful for measurement of 3-dimensional crush.

Rear impacts in the crash test data base compiled by the NHTSA are analyzed and compared to the CRASH3 rear stiffness coefficients. The CRASH3 values do not represent the test data adequately. This is because the values were derived from limited data, and because some of the rear moving barrier test data were miscoded as fixed barrier tests. A review of the larger NHTSA data base does not support the CRASH3 assumption that vehicles of similar size (wheelbase) have similar rear stiffness characteristics. Therefore, it is important when reconstructing individual accidents to use crash test data specific to the vehicles involved. Repeated rear fixed barrier test data on four vehicles are analyzed to study the data trend at speeds below and above the NHTSA test data. Constant stiffness and constant force models are compared and a combination of the two is shown to fit available test data.

The appropriateness of the set of eight frontal stiffness coefficients used by the CRASH3 program to estimate vehicle deformation energy (and to subsequently derive estimates of vehicle delta-V) is examined. This examination consists of constructing so-called CRASH energy plots based on 402 frontal fixed barrier impact tests contained in the NHTSA's Vehicle Test Center Data Base (VTCDB) digital tape file. It is concluded that the use of category coefficients within the CRASH3 program can result in large delta-V errors, reaffirming the inappropriateness of this program for use in individual accident reconstructions. The use of the CRASH3 category stiffness coefficients is seen to generally overestimate vehicle energy absorption for vehicles with small amounts of frontal crush and to underestimate vehicle energy absorption for vehicles sustaining large crush.

This paper summarizes work relating to the assessment of societal benefits of side impact protection. National Crash Severity Study (NCSS) and National Accident Sampling System (NASS) accident data technigues were reviewed with respect to the reliability of output information concerning the distribution of side impact accidents by impact severity and relationships between injury and impact severity. NCSS and NASS are confounded by errors and inadequacies, primarily as a result of improper accident reconstruction based upon the CRASH computer program. Based on review of several sample cases, it is believed that the NCSS/NASS files underestimate Lower severities and overestimate higher severities in side impact, with delta-V errors probably overestimated by 25-30 percent in the case of the more serious accidents. These errors cannot be properly quantified except on a case-by-case basis. They introduce unknown biases into NCSS/NASS.

An understanding of fundamental kinematic relationships among the several deforming surfaces of side-impacting bullet and target vehicle, occupant protection system and occupant is fundamental to rational design of crash injury counter-measures. Unfortunately, such understanding is not easy to achieve. Side impacts address the full range of bodily contacts and injuries in a way that challenges analysis. Each bodily area and organ requires individual consideration for adequate injury protection. This paper presents a simplified graphical analysis of occupant kinematics and injury exposure applied specifically to the NHTSA-proposed crabbed moving deformable barrier (MDB) compartment impact, as described in NHTSA's Notice of Proposed Rulemaking (NPRM) for Federal Motor Vehicle Safety Standard (FMVSS) 214, issued in January of 1988 [NHTSA 1988 (1)*]. Projections are offered regarding the potential of thoracic injury counter-measures.

A crash pulse representative of the accident event is often requested in addition to the reconstructed speed, deltaV, and PDOF. One approach to crash pulse generation is to scale available test data to the accident condition. Scaling formulas for time and acceleration are derived based upon commonly available accident reconstruction information from the crush profiles, closing speed, and vehicle deltaV. Scaling is based upon the compression phase of the crash pulse. A crash test similar to the accident may not be readily available unless a crash test is performed that is designed to represent a specific accident. Available test results may not reproduce the accident but may approximate it in several important aspects. In such situations it is necessary to scale a reconstructed crash pulse from the most representative test available based upon the test parameters and the reconstruction estimates.

A follow-up study on rollover testing was conducted along a section of a remote rural highway using six full-size sport utility vehicles (SUVs) of differing makes and models. The vehicles were instrumented and towed to highway speeds before being released, at which point an automated steering controller steered the vehicles through a series of maneuvers intended to result in rollover. A total of eight tests were conducted and documented, six rollovers and two non-rollover events. The six rollover events provide trip and tumbling conditions for each vehicle. The two non-rollover attempts produced cornering tire marks and allowed for the documentation of near roll conditions for the two out-of-control vehicles. All eight tests presented are instrumented real-world type tests that were later correlated based upon the data obtained.

Frontal, side, rear, pole and offset car to car data sets are examined using familiar damage analysis models: constant stiffness, bilinear stiffness, and force saturation. In addition to these, a non-linear power-law formulation is introduced and compared to the others. The power-law provides a nonlinear stiffness coefficient that transitions between a constant force model and constant stiffness model as the power goes from 0 to 1. It also provides a continuous, single valued function that is easily integrated and used in the analysis. Power-law nonlinearity can be used to smoothly fit low through high crush data. Geometric integral parameters are developed which represent irregular crush profiles. These permit graphical comparison of tests with non-uniform crush data (such as offset, side, and narrow object) with uniform crush test data. They also provide a means for comparison of accident damage with the test data set.

The data from a set of 45 reference cases, involving the collision of two generic automobile models, is presented as a basis for comparing collision algorithms used in motor vehicle accident investigation. The set was generated by the SMAC algorithm.

The relatively rare occurrence of injury or fatality in fuel-fed fires has received considerable attention in automotive safety rulemaking and products liability litigation. The literature related to fatalities associated with fire is confirmed by recent FARS data, and there are no reliable field data which confirm a need for further injury-reducing effect related to FMVSS 301. NHTSA has acknowledged this by removing crash fire rulemaking from its priorities plan. The police-reported crash fire data now available must be supplemented with in-depth investigation by trained teams before informed judgements can be made regarding further safety improvements with respect to crash fire injury.

A vehicle trajectory simulation called VTS has been developed as an aid for reconstruction of automobile accidents. The two dimensional vehicle has longitudinal, lateral and yaw degrees of freedom, a point mass at the center of gravity) yaw inertia about the center of gravity and four contact points (“tires”) which can be arbitrarily positioned. No collision or aerodynamic forces are modeled. The traction surface is represented as a flat plane with a specified nominal friction coefficient. Several quadrilateral “patches” may be applied to the surface to change the friction coefficient in specific regions. User vehicle control consists of timewise tables for steering angle and traction coefficient for each of the four wheels. When used individually or in conjunction with other computer modules, VTS provides a convenient, accurate modular tool for trajectory simulation.

The CRASH3 computer program, a well known and useful tool in accident reconstruction, is shown to be innaccurate by comparison with car-to-car crash test data. Claims for accuracy of about 10 percent cannot be validated. Both the impact model and the damage only model yield results which are in error. Cases involving error well in excess of 20 percent are demonstrated. These inaccuracies are due primarily to the omission of terms in the formulation of the energy equation and to the sensitivity of the solution to the input estimate of principle-direction-of-force.

The computer program “IMPAC” (impact Momentum of a Planar Angled Collision) was developed for use in accident reconstruction to study the impact phase of a collision. It may be used for vehicle to vehicle or vehicle to fixed object impacts. Collisions are modeled as a vector impulse in two-dimensions taking place at a specified point in each two-dimensional vehicle. A common velocity is reached at this point for the inelastic collision with complete lockup. An optional condition is available to study sideswipe type collisions in which the colliding points are permitted to slide relative to each other at a prescribed speed along a defined plane of slip. The program has been validated by comparison with the data from 16 staged crash tests conducted by the NHTSA.

The “IMPAC” collision algorithm is a comparatively simple application of momentum conservation in a collision. This 2-D model may be used in a number of applications: to reconstruct car to car collisions, to study car to barrier collisions, to evaluate proposed crash test conditions, to refine and check reconstruction calculations made using the “damage” option of “Crash3”, or as a predictor for the “SMAC” program to reduce the number of runs required to obtain a reconstruction. The program also provides a means of rapidly evaluating questions of sensitivity of results to changes in input. The essential features of the model are reviewed herein and two collision configurations are examined. The most recent version of the program provides output for purposes of comparison with the method employed by the “Damage” option of the “Crash3” program.

Damage analysis methods in accident reconstruction use an estimate of vehicle stiffness together with measured crush to calculate crush energy, closing speed, and vehicle delta-V. Stiffness is generally derived from barrier crash test data. The accident being reconstructed often involves one or more conditions for which vehicle stiffness is not well defined by existing crash tests. Massive moving barrier (MMB) testing is introduced as a tool to obtain additional and accident specific stiffness coefficients applicable for reconstruction. The MMB impacts a stationary vehicle of similar structure as the accident vehicle under accident-specific conditions like impact location, angle, over-ride / under-ride, offset and damage energy. A rigid or deformable structure is mounted to the front of the MMB, representative of the impacting structure in the accident. Four illustrative tests are presented.

Wheels and tires on vehicles, are often directly (or indirectly) involved in collisions with other vehicles or fixed objects. In this study, the effects of the pneumatic tire and rim, as it contributes to a dynamic collision, was isolated and studied. A total of 15 mounted tires of various common sizes were selected to conduct 35 dynamic impact tests into the flat face of an instrumented concrete barrier. The tires and rims used in the tests ranged from heavy truck, light truck, down to common passenger vehicle tires. Each of the 15 tires and rims were impact tested individually to failure in order to explore the dynamic response and performance of pneumatic tires in collisions. Of the 35 tests, 28 were conducted with a single tire and rim configuration and 7 tests were conducted simulating a dual truck tire configuration. It was determined that the coefficient of restitution for 22 of the tire impacts into the rigid flat faced barrier were remarkably similar, around 0.9 ± 0.1.

Three full-size sedans were towed to highway speeds along a section of a remote rural highway. Upon release, an automated steering controller steered the vehicles through a series of maneuvers intended to result in rollover. Repeated attempts to roll each vehicle were made until rollover resulted. Non-rollover attempts produced cornering tire marks by the out-of-control vehicle. Out of numerous runs, 3 rollover and 2 non-rollover tests were selected for documentation and analysis. One additional steer-induced rollover test is presented that was conducted along a simulated road section at a closed test-track facility. All six tests presented are instrumented real-world type tests that were later reconstructed based upon the data obtained from on-board instrumentation, videotape, survey measurements, and still photographs obtained of each respective test.

Crash behavior in narrow object impacts was examined for the perimeter of a 4-door full size sedan. Additional test data was obtained for this vehicle by impacting four sedans with a rigid pole mounted to a massive moving barrier (MMB) in the front, right front oblique, right side, and rear. The vehicles were stationary when impacted by the MMB. Two of the four cars were repeatedly impacted with increasing closing speeds in the front and side, respectively. Each test was documented and the resulting deformation accurately measured. The stiffness characteristics were calculated for the perimeter of car and were presented using the power law damage analysis model. The vehicle's crash performance in these pole tests was compared to that of NHTSA's flat fixed barrier tests (deformable and non-deformable) for the front, side, and rear of this vehicle.

Vehicle accident reconstruction methods based on deformation energy are argued to be an increasingly valuable tool to the accident reconstructionist, provided reliable data, reasonable analysis techniques, and sound engineering judgement accompany their use. The evolution of the CRASH model of vehicle structural response and its corresponding stiffness coefficients are reviewed. It is concluded that the deformation energy for an accident vehicle can be estimated using the CRASH model provided that test data specific to the accident vehicle is utilized. Published stiffness coefficients for vehicle size categories are generally not appropriate. For the purpose of estimating vehicle deformation energy, a straight-forward methodology is presented which consists of applying the results of staged crash tests. The process of translating crush profiles to estimates of vehicle deformation energies and velocities is also discussed.

Nine crash tests were conducted at various speeds on three vehicles in three locations under conditions that resulted in similar damage. The objective was to study the differences in crash pulse, deltaV, crush depth, and impact location with change in closing velocity from 20 to 55 mph. Three equal-weight Nissan Sentra vehicles were impacted in the front, rear, and side by an associated narrow object impact device. The three impactors were identically shaped, flat-faced, one-foot wide, and rigid; but each was designed to have a different weight (light, moderate, and heavy weight). The heavy, moderate, and light weight impactors collided with their associated test vehicle at low, medium, and high impacting speeds, respectively, in order to produce damage corresponding to a 20 mph BEV (Barrier Equivalent Velocity) in all nine tests. Impacts at the same location on the three vehicles produced nearly identical damage yet substantially differed in deltaV.