Behaviour, modelling and simulation of thin steel plates subjected to combined blast and impact loading

Abstract

This thesis studies the effects of combined fragment impact and blast loading on thin steel plates. It is here assumed that the fragments strike before the arrival of the blast wave. The fragments are represented through pre-cut defects of idealized geometries and by ballistic impact holes in the target plates before the exposure to blast loading. The effects of strength versus ductility on the dynamic response and fracture resistance of thin steel plates were investigated by studying plates with and without pre-formed defects exposed to blast-like loading conditions in a shock tube facility. Several geometries of the pre-formed defects were tested and compared using three different plate materials of target plates, i.e., two dual-phase steels and one martensitic steel. The three materials had different work hardening, an incremental increase in yield strength and a corresponding decrease in elongation to failure. The material with the highest strength and lowest elongation to fracture consistently resulted in the largest resistance to plastic deformation and lowest resistance to fracture. Fragment impact in thin steel plates typically results in perforation with surrounding plastic deformations and damage to the material. To investigate the effect of more realistic pre-formed defects, target plates exposed to high-velocity impact were compared to target plates containing pre-cut circular holes during blast loading. To establish a controllable and measurable test environment, the ballistic impact was conducted using small-arms projectiles fired from a fixed rifle. The blast load was applied separately in a shock tube facility. Even though the pre-cut holes and the ballistic impact holes were both circular with similar diameters, the ballistic impact holes introduced small petalling cracks and plastic deformations to the material around the perforation hole. As expected, crack propagation during blast loading initiated at these petalling cracks. The target plates exposed to ballistic impact showed reduced fracture resistance compared to the target plates with pre-cut circular holes. An interesting observation was that this became more pronounced with increased material strength. To increase the complexity of the blast environment, the target plates were exposed to partially confined detonations of C-4. This study was limited to only one plate material, where the target plates contained either pe-cut circular holes or holes from ballistic impact. Compared to the loading conditions from the shock tube, the partially confined detonations resulted in a much higher peak reflected pressure and a significant reduction of the positive duration phase. The corresponding pressure fields showed large time-dependent variations on the blast-exposed area of the target plates. Despite the changes in blast environment, no additional differences between the two plate configurations were observed. All blast tests were successfully measured both in terms of plate response and the corresponding pressure measurements. The global plastic deformations of the target plates showed a final deformed shape of a global dome with a superimposed local dome. The extensive dataset collected from the blast tests was used for the validation of various numerical techniques for modelling and simulation of blast loaded plates. The simulations with the best agreement to the experimental findings, were the shock tube experiments. That is, in the case where the numerical simulations were defined as purely Lagrangian analyses with the blast load applied as prescribed pressure histories. As the complexity of the blast environments increased for the partially confined detonations, both Arbitrary Lagrangian-Eulerian (ALE) and particle-based simulations had to be applied to describe the fluid-structure interaction (FSI) between the explosive, the confinement and the target plate. Considerable deviations between the two numerical approaches were observed. Compared to the experimental findings, the ALE simulations and the particle-based simulations underestimated and overestimated the loading conditions with approximately 10 %, respectively. This work provides insight into the behaviour, modelling and simulation of thin steel plates subjected to blast loading. Parameters influencing the dynamic response have been identified and investigated. The experimental dataset is therefore well suited to evaluate more numerical methods and develop new computational methods in future studies. This work also confirms the trade-off between CPU time and accuracy in the numerical modelling and simulation of combined fragment impact and blast loading. That is, simulations of fragment impact require a very fine discretization of the target plate to obtain reliable predictions, while simulations of the blast response have to use a coarser mesh to run within a reasonable CPU times. This is challenging when modelling the combined effect of fragment impact and blast loading of plated structures, especially if a realistic fracture mode from the ballistic impact and a feasible CPU cost during the blast loading phase of the simulation are important. The main finding of this work is that fragment impacts prior to blast loading can significantly reduce the blast resistance of thin steel plates. Particularly in the case of high-strength, martensitic steel plates.