Development of Physical and Numerical Model to Study Cranial Backspatter

Abstract

The study of gunshot-related blood-spatter is a common and often critical task for investigators. Simulating the formation of this spatter in a reliable manner to answer case-related questions is difficult. Furthermore, the mechanism of spatter projection is not well understood. This project addressed both these concerns. Two physical models were designed and constructed, and corresponding two numerical models were developed to study cranial gunshot wounding and spatter. A physical model of human head was developed, with correct anatomical details and biological structure of a head represented as skin, bone and brain layers. Extensive range of simulant material candidates for the skin and the bone layer were screened using another physical model of a simpler geometry. Based on the result, the optimal simulant combination to model human cranial ballistic response was recommended. Numerical models of the physical models were developed to allow more detailed analysis of the backspatter mechanisms. Ballistic impacts were simulated using a SPH method. The quality of the simulation was improved by reducing the model particle size and modulating material property inputs. The concurrent development of both the physical and numerical model allowed cross-validation of the results. A dynamic material characterisation using impact test was carried out to collaborate the findings from the physical and numerical models. A new systematic quantification criteria to allow better comparison of the results were established. The results and findings from the research was used to explain the detailed workings of the mechanisms of backspatter, and identified the important factors. Overall, the results of computational modelling and physical experiments provided valuable resources to relate actual events in crime scenes with the back-spatter observed, thus adding more credibility of this form of forensic evidence. Keywords: SPH, backspatter, ballistic modelling, modelling validation