Planar Impact Model assumptions

[brian]

New: Use this link to go to the latest post on this thread

As a part of responding to some questions on planar impact models as compared to CRASH and/or SMAC (note SMAC is an impact simulation model), we are posting the following:

• Assumptions which are inherent in some planar impact mechanics models make their practical application to real world collisions fraught with 'intuitive' and subjective variations.
  The following quotes are from SAE Paper 2007-01-0737 "Residual Crush Energy Partitioning, Normal and Tangential Energy Losses" by Brach, Welsh, Brach

• (items in quotes and "bold" are taken directly from the paper, most from page 14)

"Summary of assumptions for planar impact mechanics:"

• 1. "A single dynamic contact, taking place over a short duration."
  • ------- 'short' actually means instantaneous. real world collision take 100 to 200 milliseconds
  2. "Forces other than the contact force and impulses of forces other than the contact force are negligible"
  • ------- Tire forces and other external forces during the collision are ignored which can lead to significant errors.
    -------Please see **External Forces Discussion below.
  "3. Rotational motion of the masses can be significant"
  
  "4. Initial velocities are known and final velocities are unknown".
  • ------- Collision reconstruction is used to determine the impact speeds.
    -------Apparently planar impact mechanics must start with the answer?
    That is great for 'validation' but in real world collisions, the answer isn't known
  "5. Deformation is localized and small compared to the size of the bodies"
  
  "6. During the contact duration, position and orientation changes are negligibly small, velocity changes are instantaneous and accelerations are large"
  • -------This would lead one to believe the 'short duration' mentioned in 1 means 'instantaneous'.
    -------Real world collisions occur over time, 50 to 150 milliseconds or more, and for some impact configurations
    -------See 97-0949 and others which demonstrate the movement between point of impact (POI) and point of separation(POS) in collisions.
    • Here is a simple illustration from the 1997 paper which is based on real world tests
    • Also see RICSAC97 SAE 97-0961 and other 'real world' tests which demonstrate vehicle move between POI and POS.

      

      sample POI to POS movement.png (33.58 KiB) Viewed 2041 times

    -------This also means DO NOT use planar impact models for side-slap collisions
    Main point: Real world collisions take time and have changes in positions and orientations
  "7. The effects of the normal (crush) and tangential (sliding, shearing, entanglement, crush, etc.) contact processes are known (through coefficients)"
  • ------- From the paper "The energy partitioning process is intuitive, and does not follow directly from any principle of mechanics".
    -------So how are they 'known'?
    'known' must mean subjectively assumed
  "8. A point (impact center), C, common to both vehicles and on the line of action of the contact impulse is known"
  • ------- The collision interface between two colliding bodies in an automobile or truck collision it not a simple point.
    -------The vehicles move relative to each othe during the actual collision and the forces act along the damage interface.
    NOT at a simple point!
  "9. A common crush plane defined by an angle is known"
  • ------- The assumed 'known' angle is only ‘known’ through the "energy partitioning process" mentioned in 7 which is “intuitive and not based on and any principle of mechanics”
    Definition of intuitive is "using or based on what one feels to be true even without conscious reasoning; instinctive"...NOT scientific!

The many arbitrary and subjective assumptions of planar impact mechanics reveal that it may not a practical procedure for use in real world collision reconstructions.


**External Forces Discussion:

• Since the 1970’s it has been recognized that tire forces need to be considered in collision analyses.
  • From the 1973 McHenry SAE paper 73-0980 A Computer Program for Reconstruction of Highway Accidents
    • “The general practice in automobile collision analyses is to consider the impact and the trajectory phases of the event separately. This division of the analytical task is based upon an assumption that the effects of tire forces are negligible during the existence of collision forces. While the assumption appears to be a reasonable one, its application has been found to produce significant errors in the case of moderate-speed intersection collisions in which multiple contacts frequently occur-for example, front-side followed by side-side and/or rear-side contact.
      If secondary contacts are neglected, major errors can be produced in predictions of spin-out trajectories. On the other hand, if the tire forces are neglected throughout the time during which the collision contacts occur, significant errors can be introduced in the lateral motions of the vehicles between impacts. Thus, it is essential in a general procedure for reconstruction calculations that both the collision and tire forces be considered simultaneously.”

[MSI]

Additional information on the Planar Impact model assumptions, specifically related to the Brach book "Vehicle Accident Analysis and Reconstruction Methods". is contained in detail in this thread:

• What is the CRASH Energy Correction Factor (ECF)?

[MSI]

We moved the following Question and Full Response from the ECF thread since more related to Planar Impact Models:
Q: In planar impact mechanics spreadsheet in the Brach book "Vehicle Accident Analysis and Reconstruction Methods", there are PDOF angles, not correction factors. For example, in that example PDOF 1= -1.5 degrees and PDOF 2 = 65.5 degrees. Thats why i asked if PDOF is limited to 45 degrees in CRASH 3. Because when we take that alpha angle as PDOF and use 65.5 degrees for ECF, it is higher than 2.

A: There is no limit on the Alpha angle in CRASH. However, the maximum ECF is limited to 2.0. So if the angle produces an ECF greater than 2.0, simply use 2.0.
In the example you cited in the Brach book (p 160, 161) the comparison of CRASH3 damage (uncorrected for tangential impules) with his Planar Impact mechanics calculations is rigged!

• Brach chose to use
  • ‘the uncorrected DeltaV values given directly from the crush energy equations in Fig 8.5’ (p160). In other words, he did not include the ECF portion of CRASH.
  • For this type of impact CRASH3 would severly underestimate the results.

• If the Brach book were to properly present the CRASH3 results including the use of the ECF then the results in Figure 8.5 for the CRASH crush energy and therefore deltaV’s would be basically the same as those in Figure 8.6 for Planar Impacts Mechanics.
  • (NOTE: When comparing analytical techniques one should consider comparing the techniques with reality, such as a full scale test and measurements, since otherwise the truth is not known and if comparing two techniques with each other who know which is correct? Both might be wrong!

Brach takes issue with the simplified characterizations and concepts used to CRASH3 (see Brach p143, 158. etc).
CRASH3 was created as a simple and direct approximation technique, mainly as a pre-processor for SMAC, and then later embraced by NHTSA for NASS, their statistical studies of accidents throughout the country.

• The concept of ECF was a simplification created to compensate for the fact that work against friction is done in angled impacts.
  The concept of PDOF was used so that visually an initial estimate could be obtained:
  • • What direction do you think the Force or Impulse came from on each vehicle?
  • The concept of PDOF also more correctly instructs the analyst to consider both vehicles in the overall exchange between the vehicle during a collision. It includes the concept of a line of action (PDOF) between the two vehicles. The analyst considers each vehicle at their orientation at impact and maximum engagement while thinking about the PDOF;
    • do the vehicles and PDOFs line up?
    • Does the line of action of each vehicle coincide for the impact location and angles?
  • NOTE: Collision interactions are actually produced through Impulses, as in Forces over time, so PDOF perhaps should be called PDOI. However since it is utilized with the assumption of an instantaneous exchange of momentum and with planar impact model applications, there is no 'time variation' so it can also be referred to as a Force!)
    • (For a discussion of PDOF see How does SMAC determine the DeltaV and PDOF.

The Brach book complicates the process and comes up with the additional burden of 3 extra arbitrary and subjective inputs for his Planar impact analysis:

• 1) Gamma, defined as their ‘crush surface angle’,
• 2) Mu, defined as their impulse ratio, and
• 3) Eps, defined as their value for restitution

Some notes on these inputs:

• 1) Gamma can lead to serious problems when variations in the local stiffness of vehicles produce variations in the ‘angle of the damage surface’. How do you compensate? A whole new slew of instructions will be required to approximate Gamma to handle local variations in the structural stiffnesses.
  • Using a simple PDOF concept makes it much more easy to explain and train: Main concept is from what clock direction do you think the main force came from?
    (and of course in a collision it is not a single force but an impulse (changing force over time of the collision)
    • BUT
      since CRASH and Planar Impact models and Momentum solution procedures ALL assume an INSTANTANEOUS EXCHANGE OF MOMENTUM you are approximating the MAIN FORCE DIRECTION!)
    • Look at the vehicles, or drawings of the vehicles with damage displayed, fit them together like at MAXIMUM ENGAGEMENT and voila! Very easy to approximate the PDOF clock direction for each vehicle!
    • To model the actual Forces time history you need to use a simulation model like mSMAC3D
• 2) Mu, The impulse ratio appears to be a subjective input. You start with the 'theoretical value' for MU and apparently change it until you get the results you want or need. But what if want to do an objective scientific analysis? What value do you use then?
  Looking at the Brach book Table 6.2, the listing for the RICSAC staged collision analysis, the value for MU for identical impacts varies arbitrarily:
  • • Tests 1,6,7 are 60 degree front to side impacts, MU=.966, .824, & .772),
    • Tests 8,9,10 are 90 degree front to side impacts, MU=.413,.486,.590
    • Tests 11,12 are 10 degree front to rear impact, MU is .038, .031
    • Tests 3,4,5 are 10 degree front to rear and Mu is -.065, -0.50, -0.09
      Then look at:
    • Tests 9 above, 90 degree front to side impact, MU used was .486
      YET...
      • Figure 7.14, p 145, RICSAC 9 example, Mu is 0.512
      • Figure 7.13, p 142, RICSAC 9 example, MU is 0.806
    • WHY??? Is their ANY objective scientific reason for this variation on the RICSAC 9 test? And any justification/explanation for the variations between identical impacts?
    SMAC uses the concept of inter-vehicle friction coefficient, Mu. With a typical value of 0.55, the range is 0.35 to 0.90 depending on whether the collision interaction included more or less friction then "typical".
    It could be added as an input to CRASH and used in the ECF calculation, but that's one more input requirement.
• 3) EPS, the restitution factor we have written suggestions for including restitution in CRASH.
  • NOTE: CRASH does not include restitution! (See CRASH and Restitution for additional discussion of why it does not include restitution

Whereas for CRASH3 these are all handled internally (less arbitrary adjustments). Only the PDOF is required as input.
KISS the concept (Keep It Simple Stupid!). PDOF and ECF are KISSed concepts!
It is important to note that the same results will be obtained between the simplified CRASH damage analysis concept and the Brach book "Vehicle Accident Analysis and Reconstruction Methods" for the examples on page 160-161 if CRASH3 is applied correctly with the ECF!

[MSI]

This Q&A moved from the ECF Thread to here:

Q: In the ECF thread you mention "Looking at the Brach book Table 6.2, the listing for the RICSAC staged collision analysis, the value for MU for identical impacts varies arbitrarily" and also present values for RICSAC 9 from the book in Figure 7.14, p 145 and Figure 7.13, p 142.
I currently do not have the book, can you provide those tables?

A: Yes here they are.

• NOTE: the handwritten notes and yellow highlights we added while reading the book. We'd also suggest you visit a library or obtain the book so you can see if you find any presentation of the rationale for the arbitrary variations of the values in these tables/figures.

Brach fig 6.2.jpg (120.48 KiB) Viewed 2170 times

Brach fig 7.13.jpg (140.05 KiB) Viewed 2170 times

brach fig 7.14.jpg (144.98 KiB) Viewed 2170 times

[MSI]

This Sept 2019 post moved from What is ECF to here
The Brach book was updated to 2nd edition (edition 1 2005, edition 2 2011) Vehicle Accident Analysis and Reconstruction Methods, Second Edition R-397
We recently viewed a copy of 2nd edition and initially it appears no changes were made to items discussed in this forum topic.
The following is a guide for the tables/pages

brach 1st v 2nd edition.jpg (32.4 KiB) Viewed 2164 times