The severity of a planar crash is most commonly defined by the change in vehicle velocity, or delta-V (ΔV). In the National Automotive Sampling System – Crashworthiness Data System (NASS-CDS), ΔV is computed from post-crash vehicle damage using a CRASH3 – based computer program called WinSMASH. Prior studies have investigated the accuracy of NASS-CDS ΔV in real world frontal crashes. Those studies compared the WinSMASH ΔV estimates in NASS-CDS to the ΔV obtained from the crashed vehicles’ Event Data Recorders (EDRs). In those studies, the EDRs only measured/recorded ΔV in the longitudinal direction. Accordingly, the accuracy of WinSMASH ΔV in side crashes has been assessed only through comparison with controlled crash tests, not real world crashes. Many newer vehicles are now equipped with EDRs that record both longitudinal and lateral ΔV. The objective of this study is to use these newer EDRs to compare WinSMASH ΔV estimations to EDR ΔVs for real-world side crashes in the NASS-CDS.

This preliminary study examines 22 side impact crashes from the NASS-CDS. All struck vehicles were cars and all striking vehicles were either cars or light trucks and vans (LTVs). EDR measurements of side impact ΔV were used to evaluate the accuracy of WinSMASH ΔV estimates for these real world side crashes. WinSMASH systematically overestimated ΔV for the examined crashes. Overestimation for crashes involving cars struck by cars appeared greater than for those involving cars struck by LTVs. Observed systematic ΔV error varied by the area of the vehicle which was impacted.