Page 1 of 1

What is the SMAC Inter-vehicle friction coefficient?

Posted: Sat Jun 20, 2009 2:01 pm
by MSI
Q: What range of values should be used in SMAC for inter-vehicle friction?
A: From our SAE paper 970947 SMAC-97:
  • "The Inter-vehicle friction coefficient of the SMAC collision model (AMU, SMAC input card 13, field 7) is based on coulomb friction. Coulomb friction is independent of the sliding velocity. Coulomb friction should have a coefficient value of less than or equal to 1.0. During the collision, as relative sliding of the surfaces occurs, the magnitude of the inter vehicle friction force depends on the existence and magnitude of the collision normal force. Supplemental tensile forces and/or moment constraints sometimes occur in vehicles as the vehicles begin the separation phase when the normal force may be very small or zero. After the primary collision, as a result of the impact configuration, the vehicle contact area and/or the vehicle component contacted, intermeshed components of the two vehicles can offer additional resistance to separation in the absence of the normal forces required for coulomb friction. Using a value for the inter-vehicle friction coefficient greater than the recommended range (normal value 0.55, recommended range 0.3-1.0) may unrealistically and/or adversely affect the primary impact phase, since the additional forces and moments produced by the elevated friction can act to redirect the forces and moments during the primary impact."
In a paper by Marine, et al, SAE 2007-01-0744 "On the Concept of Inter-Vehicle Friction and Its Application in Automobile Accident Reconstruction"
The appendix includes an attempt to justify using intervehicle friction >> 1.0.
Why? In our paper "SMAC2003: The Automatic Iteration of SMAC", SAE paper 2003-01-0486,
  • ALL values for inter-vehicle friction coefficient were < 1.0
  • Paper "2003-01-0486" validated the SMAC program by producing results for impact speeds and deltaV all less then 10% of the measured values while starting with CRASH/linear momentum results and then automatically iterating to a 'match' of the evidence (damage and scene). (NO human intervention in the iteration process and the initial starting values from CRASH which were sometimes >>20% in error). Marine, et al never contacted us about what values we used in our 2003 paper (and as we mentioned all were less than 1.0).
From our 1997 SAE paper SMAC-97:
  • "A secondary problem which may occur with original SMAC (EDSMAC) use of a high value of inter vehicle friction can occur in the post-processing program which may not be able to "match" the accelerations to the damage region. As a result invalid Delta-V values and clock direction may be reported. An EDSMAC application summary page (used by an expert witness as an exhibit to deposition testimony) which contains invalid Delta-V values for the collision partners was produced by the use of a value for the inter vehicle friction coefficient of 2.0 and is shown in in the attached Figure. (Note that in the EDSMAC application m1*DeltaV1 does not = m2*DeltaV2). SMAC creates a supplementary diagnostic page which reports all of the impact speed changes calculated for a given run to permit a check of the values reported by the post-processing routines. This auxiliary page appears to have been omitted in the EDSMAC program"
As per our recommendations in our 1997 paper and the values used in our 2003 paper intervehicle friction should always be less that 1.0.
  • Any need for the use of a value of intervehicle friction > 1.0 indicates that the force of collision is not a function of the normal force and most likely a snag or inter-locking of the vehicle occurred and therefore that the SNAG option of SMAC should be used.
  • Also in the SMAC Input Manual (Help->SMAC Input Manual) OR also online in the SMAC input Manual DISUCSSION: SNAG Option

Re: What is the SMAC Inter-vehicle friction coefficient?

Posted: Wed Oct 28, 2009 4:52 pm
by MSI
Q: What values for inter-vehicle friction coefficient did you use in your 2003 paper ‘SMAC2003 – The Automotic Iteration of SMAC’.
A: For the paper we first examined the values used in previous SMAC validations studies. The first being by Jones & Baum, “RICSAC – Vol IV – Staged Collision Reconstructions”. We found several of the values of inter-vehicle friction used in that study to be excessively high.
In 1990, Day & Hargen repeated the RICSAC SMAC runs performed by Jones & Baum as validation for the EDSMAC computer program in the paper Further Validation of EDSMAC using the RICSAC Staged Collisions. They did not make any changes in the inputs for inter-vehicle friction coefficient.
In 1996, Day published "Validation of Several Reconstruction and Simulation Models in the HVE Scientific Visualization Environment," SAE paper 960891. EDC did not vary any of the the values for inter-vehicle friction coefficient used for the RICSAC validation runs.
In our 1997 paper SMAC97 – Refinement of the Collision Algorithm’ we discussed inter-vehicle friction (see the thread above).
The following table contains a list of the values used by Jones & EDC and then the values we used in our automatic iteration validation simulations. We reduced the excessively high values of inter-vehicle friction used in previous SMAC validation runs to be less than 1.0 (coulomb friction should always be less than or equal to 1.0).
RICSAC_AMU.jpg (15.33 KiB) Viewed 2208 times
We are currently working on a sensitivity study of SMAC using SMACITER aka ‘SMAC2003 – The Automotic Iteration of SMAC’, to help objectively determine the effects of changes in the range of values for inter-vehicle friction. For example, we will re-run each RICSAC SMAC reconstructions with SMACITER with variations (+/-20%?) in the input for the inter-vehicle friction coefficient and report how much does that vary the resulting speed and impact speed change, if at all?