Evaluation of NASS-CDS side crash delta-V estimates using event data recorders

Area of Science:

• Crashworthiness and Biomechanics
• Vehicle Safety Engineering
• Accident Reconstruction

Background:

• Planar crash severity is typically quantified using delta-V (ΔV), the change in a vehicle's velocity vector.
• The National Automotive Sampling System - Crashworthiness Data System (NASS-CDS) utilizes the WinSMASH program for ΔV estimations.
• The accuracy of WinSMASH ΔV estimates in real-world side impacts has not been previously validated.

Purpose of the Study:

• To investigate and validate the accuracy of WinSMASH-generated ΔV estimates in real-world side-impact crashes.
• To compare WinSMASH ΔV predictions against biaxial data from event data recorders (EDRs).
• To assess the accuracy of the Principal Direction of Force (PDOF) estimates in NASS-CDS.

Main Methods:

• Utilized biaxial ΔV data from EDRs for single-event side crashes within the NASS-CDS database.
• Compared WinSMASH-estimated resultant ΔV with EDR-recorded resultant ΔV, adjusting EDR data for a 10% restitution value.
• Compared NASS-CDS PDOF estimates with PDOF calculated from EDR data.

Main Results:

• WinSMASH systematically overestimates ΔV by approximately 12.9% for car-vs-car side impacts and 2.4% for car-vs-light truck/van (LTV) impacts.
• ΔV error varied significantly based on the impacted area of the vehicle's side.
• The mean discrepancy in NASS-CDS PDOF estimates was -0.9°, with a standard deviation of 12.6°.

Conclusions:

• WinSMASH overestimates delta-V in side impacts, suggesting its stiffness parameters better reflect LTV impacts than car impacts.
• The accuracy of ΔV reconstruction is dependent on matching stiffness parameters to the specific crash mode and impacted area.
• NASS-CDS PDOF estimates show no systematic bias and exhibit random discrepancies consistent with prior research.