Abstract
The objective of this study was to assess the accuracy of using
high-speed frontal barrier crash tests to predict the impact speed, i.e.
equivalent barrier speed (EBS), of a lower-speed frontal barrier crash.
Force-displacement (F-D) curves were produced by synchronizing
the load cell barrier (LCB) data with the accelerometer data. Our
analysis revealed that the F-D curves, including the rebound phase,
for the same vehicle model at the same impact speed were generally
similar. The test vehicle crush at the time of barrier separation,
determined from the F-D curves, was on average 17±16% (N = 150)
greater than the reported maximum hand-measured residual crush to
the bumper cover. The EBS calculated from the F-D curves was on
average 4±4% (N=158) greater than the reported EBS, indicating that
using F-D curves derived from LCB data is a reliable method for
calculating vehicle approach energy in a crash test. Our method of
using F-D curves from high-speed tests to predict the EBS of a
lower-speed barrier crash overestimated the EBS of actual lower-
speed tests by an average of 21±9% (N = 129). Further work in
developing and reining our method is needed to improve the
accuracy of predicting a lower-speed EBS.
Introduction
One method of determining the severity of a collision is to calculate
the deformation energy based on the vehicle crush and an estimate of
the vehicle stiffness. Current methods estimate the vehicle stiffness
using linear or non-linear crush coeficients. Linear crush coeficients
assume constant vehicle stiffness and rely on measurements of
residual crush, i.e. static crush, of crash test vehicles [1, 2, 3]. Several
methods exist for determining non-linear crush coeficients. One
method is to create a polynomial it of the square root of 2E/w and
residual crush [4, 5, 6, 7]. In instances where there are no tests
conducted at lower-speeds, the deformation energy is determined by
integrating portions of the force-displacement (F-D) curve from a
load cell barrier (LCB) crash test [7]. Another method for
determining non-linear crush coeficients is by calculating a
polynomial it of a pressure-displacement (P-D) curve [8]. P-D curves
are generated from F-D curves of crash tests conducted at the highest
available impact speed for the vehicle model. Methods that rely on
linear crush coeficients or non-linear crush coeficients derived from
F-D curves assume that barrier impacts at lower-speeds would mimic
the F-D curve of a higher speed barrier impact.
Our study investigates this assumption using an alternative prediction
method that does not rely on crush coeficients or residual bumper
cover crush measurements to estimate vehicle stiffness. Instead, our
method shifts the rebound proile of a high-speed F-D curve to the
same crush at barrier separation as that of an actual lower-speed F-D
curve (see methods for more details). In other words, we assume that
the separation velocity of a high-speed test is the same as that of a
lower-speed test. Previous literature has shown similar rebound
proiles for F-D curves for repeated impacts of increasing severity of
the same vehicle [9]. Other literature has shown the same vehicle
model exhibits similar separation velocities for a wide range of
barrier impact speeds [10]. Since our prediction method directly
compares F-D curves from high- and lower-speed tests, any errors in
our lower-speed predictions would be directly correlated with
differences in high- vs lower-speed F-D data.
The National Highway Trafic Safety Administration (NHTSA)
conducts many crash tests, and in addition to reporting the impact
speed and residual crush data, they also record dynamic data from
load cells on a rigid barrier and linear accelerometers mounted at
various locations on the test vehicle. These accelerometer data can be
double integrated to estimate the vehicle displacement and then
synchronized with the load cell data to generate F-D curves, as
outlined in detail by Gilbert et al [8]. In theory, the area under the
F-D curve from LCB data is expected to equal the mechanical work
done to the vehicle, i.e. the deformation energy dissipated by the
vehicle, during the collision; however, previous literature has found
this method will over-represent the energy dissipated in the crash
[11]. We investigated the accuracy of the LCB data by comparing
NHTSA’s reported EBS to the EBS calculated from the F-D curve so
that readers could quickly identify which barrier impact speeds were
used in the comparisons.
Using Force-Displacement Data to Predict the EBS of Car into
Barrier Impacts
2016-01-1483
Published 04/05/2016
Ross Hunter, Ryan Fix, Felix Lee, and David King
MEA Forensic Engineers and Scientists
CITATION: Hunter, R., Fix, R., Lee, F., and King, D., "Using Force-Displacement Data to Predict the EBS of Car into Barrier
Impacts," SAE Technical Paper 2016-01-1483, 2016, doi:10.4271/2016-01-1483.
Copyright © 2016 SAE International
Downloaded from SAE International by Felix Lee, Friday, March 11, 2016