| Summary: | This report presents results of an experimental and numerical investigation into the response of partially-confined, thin-walled, stainless steel cylinders subjected to internal blast loading. "Partial-confinement" refers to an enclosure that may retain a significant, quasi-static pressure following an internal explosion, while "thin-walled" implies that the cylinder wall thickness is small relative to other geometric dimensions. The cylinder deformation is used to gauge the level of blast damage. The chosen cylinders are of length l = 300mm, inner radius a = 150mm, and wall thickness h = 2mm, and cut from seamless 304 stainless steel pipe. Partial-confinement is achieved by keeping one end of the cylinders closed in all tests. The experimental tests are conducted on the horizontal ballistic pendulum at the Blast Impact and Survivability Research Unit (BISRU), University of Cape Town. The blasts are generated by detonating radially-centred, spherical PE4 charges inside the cylinders. The charge mass is varied between 20g and 75g at two axial charge positions, specifically 150mm and 225mm, relative to the closed end. These axial positions are denoted 0.5 l and 0.75 l respectively. Polystyrene annuli are used to position the charges within the cylinders, and the influence of this polystyrene on the cylinder deformation is briefly investigated as an additional parameter. Details are presented of the development of an LS-DYNA Release 6.0.0 computational model that simulates the cylinder response to blast loading. Several 1D and 2D preliminary simulations and convergence studies are presented, the results of which inform the mesh sizes in the final model. The air and explosive are modelled using solid Arbitrary- Lagrange-Euler (ALE) elements, and the cylinders are modelled using Lagrange solids. Since the cylinders and explosive are all circular in section, the simulations are performed in 2D axisymmetry to reduce computational expense. The maximum cylinder deflections and selected final profiles, as well as the impulses imparted to the pendulum, are compared to the corresponding experimental results. With the exception of the 0.75 l tests at larger charge masses, the results exhibit generally good experimental-simulation correlation. For the 0.5 l tests, the cylinders exhibit a linear increase in deformation with increasing charge mass, while the relationship is an exponential increase for the 0.75 l axial charge position. For charges below 45g, the deformations from both axial charge positions are similar, however the responses diverge with increasing charge mass, indicating that the confinement effect of the cylinders is a function of the axial position and is influential only beyond a given mass of explosive. This confinement effect is greater when the charge is located nearer the open end of the cylinder. The computational models provide insight into the transient behaviour of the systems which cannot be achieved experimentally. The influence of the charge position is confirmed by comparing the simulated deformation-time histories for the different axial charge positions. Two pressure fronts are evident in the simulations: one moving radially and one axially. The significant structural damage is caused by the radial pressure incident on the cylinder wall, while the laterally moving pressure drives gas out from the open end. In the case of the 0.75 l simulations, the pressure incident on the cylinder wall has longer to act before it is expelled by the laterally moving pressure. For higher charge masses, the high pressure acting during this additional time is the cause of late-time deformation. Two tests are performed using a half-annulus of polystyrene. Relative to the other tests, these two exhibit greater radial disparity, with the deformation biased to the side with polystyrene. This preliminary result suggests that placing polystyrene between the charge and the cylinder increases the structural deformation, and necessitates further investigation.
|
|---|
