ABSTRACT
For crash tests conducted by organizations other than the NHTSA, it is generally the case that digital accelerometer data files are not available. Rather, the testing laboratory will have applied their own filtering (according to SAE J211, one hopes) and presented the results on a hard-copy plot. Most often, these are graphs of acceleration versus time, called crash pulses. There may also be plots of velocity and displacement versus time, but often that is not the case. In such situations the analyst does not have the burden of filtering the data, but must take on the task of digitizing the hard-copy plot and integrating the results. Chapter 13 describes the process, the quality of the results, and some factors that affect the quality. It also presents an example of how to interpret results, whether obtained from hard-copy plots or digital data. Specific topics discussed include integrating the accelerations, the repeatability of digitizing hard-copy plots, the effects of the quality of the original plot, the accuracy of the integration process, the accuracy of the filtering process, the effects of filtering on acceleration and velocity data, and the effect of accelerometer location on the crash pulse.
Of all of the methods investigated for analyzing crash pulse data, the numerical processing of raw data comes the closest to reproducing the peak accelerations and velocity change reported by the test lab (within 0.1% on both parameters). Hand-digitizing a high-quality Class 60 plot is not as accurate in reproducing the ΔV but is still within 1.0%. If the crash pulse plot being hand-digitized is of poor quality, the variability of the results of integration can be doubled. However, the variation of ΔV is still within 2%.
One should be careful in choosing which accelerometer traces to analyze for the purpose of characterizing the crash. The location(s) selected for that purpose should be in the compartment, as far away from the crush zone as possible, and on sturdy structure.
