Developments in road vehicle crush analysis for forensic collision investigation.

The change of a vehicle’s velocity due to an impact, DeltaV (v) is often calculated and used in the scientific investigation of road traffic collisions. Two types of model are in common use to achieve this purpose, those based on the conservation of linear and angular momentum and the CRASH model which also considers the conservation of energy. It is shown that CRASH and major implementations of the momentum models are equivalent provided certain conditions are satisfied. Explicit conversions between the main variants of the models are presented. A method is also presented which describes a new formula for determining the total work performed in causing crush to a particular vehicle. This has the advantage of incorporating restitution effects and yields identical results to the momentum only models. Although the CRASH model has received adverse criticism due to perceived inaccuracies in the results, little work has been performed to determine the theoretical limitations on accuracy. This thesis rectifies that shortcoming. A Monte Carlo simulation and analytical model are developed here to provide two independent methods for determining the overall accuracy of the CRASH method. The principal direction of force was found to be the most likely to introduce error based on the CRASH assessment. It is shown how this and other sources of error in the CRASH model can be quantified for a particular collision suggesting priorities for minimising the overall uncertainty. The data from a series of well known crash tests are used with each of the models to provide comparison and validation data. It is recognised that without additional data velocity change is of limited use for forensic investigation. However DeltaV can be used as a proxy for acceleration and is particularly useful in studies involving injury causation. A method is also presented here which uses the change in velocity sustained by a vehicle in a planar collision to estimate the velocities of a vehicle before and after a collision. This method relies solely on conservation laws and is also applicable to situations where the coefficient of restitution is non-zero. An extension to the method is also described which allows an initial estimate to be modified to generate more realistic directions of force. This extension has the desirable effect of reducing uncertainty in the estimation of the direction of force which significantly improves the overall accuracy