A Negative Coefficient of Restitution? An Irresponsible vCrash Video!

General Questions related to the Momentum Based Analysis programs
#pc-crash #virtualcrash #crash
Site Admin
Posts: 2290
Joined: Thu Jun 18, 2009 12:37 pm

A Negative Coefficient of Restitution? An Irresponsible vCrash Video!

Post by MSI »

In a recent post on the NAPARS facebook page which was presenting what type of information is contained in NHTSA Crash test database reports they presented results from test #10156 which was a 2018 Chevrolet Traverse 4D SUV 35 MPH barrier crash. As part of the section on restitution (they have daily postings about various parts of the NCAP reports) the coefficient of restitution for the test was calculated (see SIDEBAR below)
  • The following comments were posted by Facebook users in response to the NAPARS facebook page restitution post:
    • "I have to disagree with the absolute value and that restitution is always positive. For example, see this Virtual Crash video. Negative restitution values are physically possible"
      • "Example, Mercedes over 90 mph into the side of a Honda Accord. The Honda split in two pieces and the Mercedes passed right through. The restitution is negative. How about the vehicle that slide sideways into a tree at high speed and broke into two pieces and continued pretty much in the same direction as it came? That too is a negative restitution"
    • There was also posted a reference to a planar impact model related discussion from a book mentioning that a value of negative restitution worked in their planar impact equations but it also included the term 'rare' and provided no examples (see footnote below for quote from book)
    • There was also a link to a research paper on the possibility of molecular level negative restitution which was research unrelated to vehicle crashes.
OUR RESPONSE:The following is our response to the possibility of 'negative restitution' in crash reconstruction:
  • NAPARS facebook page and this forum deal with crash reconstruction and
    There is NO NEGATIVE RESTITUTION in car crashes!!
  • If a vehicle breaks apart, it is a structural failure and therefore an incomplete collision, NOT negative restitution.
  • If the vehicles structurally stick together after/during an impact such that one vehicle appears to be pulled forward that is an indication of forces beyond compressive crush forces and coulomb friction and NOT negative restitution.
  • Also see What is the Coefficient of Restitution?

    If an analysis of a vehicle crash seems to indicate a negative restitution then it is either an indication of vehicle structural failure or structural interlock or a limitation of the crash analysis technique.

    As cited above they also cited a irresponsible and misleading video with the title:
    • vCRASH Academy | Veteran Level Course | Using Crush and Negative Restitution
    In the video, which i wouldn't recommend wasting time on, they include a presentation of a 1-Dimensional crash evaluation which indicated negative restitution! However they mention that it was a 1-D analysis which included no consideration of rotation or lateral movement so the vehicles seem to run thru each other (they didn't!)
    I believe it is a silly irresponsible and confusing presentation particularly since they included in the title of the video negative restitution!!
    • Negative restitution is ONLY due to the use of a simple 1-Dimensional analysis:
      • 1-D ignores rotation and ignores lateral movement. It's one dimensional! It is silly like 'gee wiz if i ignore movement aspects of a crash (lateral movement and rotation) i can get crazy 'negative restitution' results! WHAT IS THE POINT??!!
    • The example they used in the video was from Daily, et. al Fundamentals of Traffic Crash Reconstruction Volume 2 Example 19.1, pages 678-689 in the chapter on Conservation of Linear Momentum (COLM).
      • In the book it was a simple demonstration of using COLM analysis,energy balance and simultaneous equations. It was a good example for the purposes it was used for in the book.
        • For VCrash to use the incomplete data presented in the book as a demonstration of a simulation program, particularly with starting the discussion on the video with talk of negative restitution is irresponsible.
        • The example in the book presented no measurements of positions of rest or tire marks or debris to allow a determination of the approximate accurate Impact and rest positions.
        • There was only one photograph from the roadway of the vehicles at rest
        • There were some crush measurements and one side photograph of each vehicle's damage.
          example 19.1 photos.jpg
          example 19.1 photos.jpg (36.46 KiB) Viewed 5710 times
        • It was simply an example in a book with incomplete scene and vehicle data presentation/documentation.
      There is no negative restitution in vehicle crash reconstruction!
      The only thing demonstrated in the video was that after a lot of tweaking of the arbitrary and subjective point and maximum tangent line friction required for the VCrash instantaneous momentum exchange they can get results which generally agree with the COLM and energy basis hand calculation solution contained in Example 19.1 by Daily, et al.
      There is no check on either of the results since the actual physical data, impact speeds and impact speed changes, for the example in the book are not known. It was not a measured full scale crash test, it was simply a highway crash the authors of the book chose to use as a example
    MAIN POINTS: In CRASH Reconstruction there is NO negative restitution. If ANY program requires that a user input a value of negative restitution to model any collision it is due to a LIMITATION of the program and NOT because there is negative restitution.

  • As part of the NAPARS facebook page presentation on restitution they calculated the coefficient of restitution for the test which was 18%!!
    • That is very high for a 35 MPH barrier crash and from my examination of the test report (which unfortunately did not include any of the 8 vehicle accelerometer data, only occupant) however careful examination of the video demonstrated that the tires participated in the bounceback from the barrier.
      See our writeup on this 18% Restitution at 35 MPH??

    The following plot shows generally at 35 MPH (~56 kph)the coefficient of restitution is lower than 0.18
    Obviously from this data some values are above that value
    Question is: is it related to a smaller front overhang and additional rebound produced by the tires?
    restitution over a range.png
    restitution over a range.png (85.63 KiB) Viewed 5728 times
    Also see our post What is the Coefficient of Restitution?
  • In a book by Brach, etc., now in it's 3rd edition in Section 6.3.1 (in 1st & 2nd edition) they include the following,
    • Edition 1
      Brach 1.jpg
      Brach 1.jpg (22.71 KiB) Viewed 3358 times
    • Edition 2
      Brach 2.jpg
      Brach 2.jpg (20.16 KiB) Viewed 3358 times
    Although Edition 2 says "discussed elsewhere [6.1]" there is NO DISCUSSION on negative restitution in [6.1]
    • Note in [6.1] they do say
      'So long as the contact duration is small and the contact force large, momentum in a crash is approximately conserved"
      • Note 'duration is small' they mean instantaneous (an assumption of planar impact models)
      • Note 'approximately conserved' (another assumption because they ignore external forces during the crash
      For more information see our discussion on assumptions of planar impact models
    But in [6.1] there is no discussion of negative restitution
    In the book, in several places, they correctly emphasize that for vehicular collisions the coefficient of restitution is ALWAYS between 0 and 1!