Quantifying the forces in stabbing incidents.
PhD thesis, University of Glasgow.
Full text available as:
Stab wounds are an increasingly common cause of death or series injury and the high-risk groups in society are growing both in size and number. These facts make the study of mechanics of knife penetration more relevant than ever.
The aim was to quantify the penetration force needed to inflict a certain stab wound by modelling knife penetration via the Finite Element Method. Case studies of stabbing incidents were carried out to give some insight into the nature and type of problem to be modelled. It was decided to work with an idealised stab-penetration model including a section of target tissue simulants. This stab-penetration test could yield repeatable and comparable results both experimentally and computationally. Suitable target simulants were identified by the stab-penetration test and also by uniaxial tensile tests. Pig skin was found to roughly match the mechanical properties of human skin with gelatine as a realistic flesh simulant.
Computational modelling of knife penetration was attempted by use of Abaqus/Explicit, a nonlinear FEA package which features modelling of contact-impact problems. A true to scale finite element model of the stab-penetration test set-up was built including a material model of the target simulant. The computed penetration force was found highly mesh dependent for sharp blades and too high forces were predicted. Blunt penetrators were also tested both by experiments and computationally. By refining the constitutive model for skin computed values were obtained in reasonable agreement with the experiments for blunt penetrators. Mesh dependency was minimal in the computational model with blunt penetrators. It was concluded that modelling of knife penetration via finite element method is possible but analysis is time consuming due to the high mesh refinement required. Accuracy of the predicted penetration force is still too low for typical blade sharpnesses to be of practical use.
Actions (login required)