Use Maximum Crush For CRASH A&B Crush Coefficients?

Questions/Topics Related to the CRASH computer program
#CRASH #EDCRASH #PC-CRASH
Post Reply
MSI
Site Admin
Posts: 1792
Joined: Thu Jun 18, 2009 12:37 pm

Use Maximum Crush For CRASH A&B Crush Coefficients?

Post by MSI »

Q: Can you use Maximum Crush for the creation of CRASH Crush Coefficients?

Got a note requesting additional clarification on HOWTO Calculate CRASH3 A & B Coefficients and here is some additional information:

Quick response:
  • Q: Can you use Maximum Crush for the creation of CRASH Crush Coefficients?
  • A: NO! It violates the CRASH simplifying assumptions, it has NO scientific basis, and it creates meaningless A & B Coefficients.

Now for the long answer:

Apparently a vendor who sells crush coefficients sometimes randomly uses maximum crush to calculate the A & B coefficients if and when there is incomplete or insufficient static data of the residual crush. The vendor claims this creates 'conservative' coefficients.
but
does it pass the smell test? they claim it create "conservative" results...does it? what is it?

From Review of CRASH Damage analysis and the NHTSA "reformulation" (and all other publications on CRASH)

CRASH is a
  • simple linear relationships between the impact speed-change and the extent of residual crush

If the use of some arbitrary value for maximum crush to create A & B Crush Coefficients for Damage Analysis is RANDOM and WRONG! By arbitrarily fitting impact speed (Or do you use Total deltaV?) to the "maximum crush" is meaningless
  • It has NO relationship to the scientific basis of damage analysis and CRASH
The simplifying beauty of using CRASH damage analysis,
is the ability to measure a fully restituted vehicle at any time after the crash
and then determine an approximate impact speed in a uniform manner.

To randomly use some characterization of the "maximum crush", which depending on structural restitution properties, may be a little or a lot over the residual crush, to create A & B crush coefficients is erroneous and meaningless.
  • And since they use it when there is no or questionable residual crush measurements, how do you know how it might compare to the residual crush?
First, maximum ‘measurements’ aren’t ‘measurements’
  • since depending on the test facility what they might be reporting is:
    • Double integrated accelerometer results to produce velocity and then displacement are subject to ringing, position in vehicle issues, etc etc
    • Or are they based on GPS measurement devices which have their own set of issues.
    • Or are they calculated by film analysis and monitoring targets or a single target on the vehicle?
    • Or are they simply reporting the Maximum Static Crush?
      You need to check each and every report to determine!
  • Unless you have a comparison of valid residual crush measurements there is NO way to know how any reported maximum crush relates to the residual crush.
  • And if you have valid residual crush measurements, WHY would you need to use the maximum crush EXCEPT to try to characterize the restitution properties of the vehicle?
DO NOT USE MAXIMUM CRUSH TO CREATE A & B COEFFICIENTS!!!
  • It violates the CRASH simplifying assumptions, it has NO scientific basis, and it creates meaningless A & B Coefficients.
    • The argument that since the Maximum may or may note be larger than residual so then it is 'conservative'
      • if that is true do a research study of the reported maximum v residual crush areas and see whether that it true
    • The determination of what is reported as "the maximum crush" may be through three processes listed above
      OR
    • it may simply be they indicate the 'maximum residual crush'??
    Only through a properly prepared research project can you quantify and clarify exactly what any reported 'maximum crush" might represent in the NHTSA reports and based on what testing organization did the crash test and reported the measurements
Taking a step back, lets rehash some things that have been published and posted about CRASH:
  • From Is CRASH damage analysis a virtual relationship?:
    • The original CRASH program, which utilized both piecewise-linear trajectory solution procedures and a damage analysis procedure, was created to provide an initial estimate for the SMAC computer program for use in the NHTSA NASS investigations.
    • NHTSA adopted the CRASH program for their NASS statistical studies since they considered the average error in severity determination as more important than any individual errors.
    • NHTSA understood the importance in statistical research for the uniform interpretation of crash evidence
    • Now fast forward 40+ years and many software vendors and crash reconstructionists use CRASH damage analysis (or clones of it) in individual crash reconstructions.
    • Crash Reconstructionists use CRASH without ANY modifications (see notes on CRASH accuracy)
    • NHTSA distanced themselves from CRASH by creating WinSmash (with minor cosmetic changes)
      • This was to get away from litigation related subpoenas related to CRASH! (since folks using it in individual case reconstructions)
    In 1997 we published research Effects of Restitution in the Application of Crush Coefficients which showed a pathway to updating/improving the original CRASH formulation and accuracy.
    • We got a lot of misguided and erroneous pushback from the litigation fueled cottage industry built around CRASH. They had a financial interest in being the self proclaimed ‘crash coefficient experts’ and so we were strapped with incompetent and inaccurate ''peer review' garbage attached the paper as part of the flawed SAE paper ‘peer review’ process.
Another thing folks are trying to sell to try to imply 'improved accuracy' is adding more than 6 points of measurements:
  • This is silly since the accuracy of damage analysis does not merit any more than a few measurement…just need the general area of crush
    Some companies sell using a cloud scanner to measure in detail the 3 dimensional damage!! silly! NO improvement in accuracy!
    • the raw data, the actual scientific basis for damage analysis, is based on the crash test database and the 2,4 or 6 CRUDE MEASUREMENTS OF THE RESIDUAL CRUSH made at a single elevation. The vehicle is crashed into a barrier as part of the NCAP certification testing
    Tests have been done using electron microscope type scanner measurements vs simple tape measure
    • the results are similar/same/no significant difference!
    • CRASH damage analysis is limited in accuracy

      See our October 26, 1987 SAE Committee Correspondence
      • We’ve been trying to reel in the claims of accuracy related to damage analysis which mostly is based on a single crash test of questionable veracity and an assumption of a ‘no damage’ intercept.
      But folks still love it and use it and the best we can do is try to temper the claims of accuracy.
    SUMMARY: Randomly basing A&B crush coefficients on questionable "maximum crush" reporting because there is no residual crush available, or simply to compare as a ‘what if’, is of questionable scientific value.
    • No damaged car from a crash or barrier test is at ‘maximum crush’ (unless the vehicle has plastic crush properties!) and so coefficients based on it are meaningless and have no scientific basis.
LINKS: LATEST TOPICS, Info on LEASE pricing, Forum Index, McHenrySoftware.com,
Question? Comment? Please email us (all communications considered confidential)
(c)McHenry Software, Inc ALL Rights Reserved.
Post Reply