Projectile impact of fluid backed metal beams and plates : experiments and numerical simulation
The growth of the nuclear power industry has provided a considerable stimulus for investigations into fluid-structure interaction problems. The safety case for nuclear reactors requires an understanding of the impact response of structures enclosing or surrounded by fluids. In many cases the structural response is in excess of that which can be predicted by elastic analyses and both material and geometrical non-linearities must be considered. The understanding of the interaction between the structure and the contained fluid poses additional problems which, in the extreme loading conditions envisaged, have received little attention. There is a lack of data relating to basic fluid-structure interaction problems involving dynamic plastic structural impact. Two sets of experiments are described which were carried out to provide some such data. The first set of experiments considered beams, both fully clamped (leading to large membrane forces) and partially clamped (preventing rotational and transverse motion while allowing the beam material to be fed in from the supports), struck centrally by a projectile. The second set of experiments considered a circular plate clamped around its periphery, sealing a volume of fluid, and struck centrally by a projectile. The shape of the plates and beams as they deformed were recorded, as were the pressure variations during the tests. In both sets of experiments the main contribution of the fluid to the beam or plate response was to localise the deformations. The early deformation of the beams was limited to the centre half span and the deformation only spread to the ends of the beams as the supporting effect of the fluid was lost due to the fluid escaping. In the plate experiments, where a good seal could be achieved, the deformation throughout was localised compared with a similar plate in air. The deformation in these cases was limited to a central disc of approximately half the plate diameter. The pressures recorded during the tests suggest that the fluid response was predominantly incompressible. A finite element program was written to model the response of beams and circular plates (axisymmetric problems). A brief history of the finite element method, the background theory and the development of the method to treat non-linear, large displacement, dynamic problems are given. The results are presented for a number of beam and plate problems, both those described above and other problems for which data was available. The finite element program was found to give good predictions of the deforming shapes of both the beams and the plates. No detailed analysis of the fluid was carried out, but two types of approximation to the effect of the fluid were investigated. Firstly a time varying pressure pulse (based on the measured pressure pulses) or a pressure loading derived from the beam velocity (acoustic and incompressible fluid approximations) were used to represent the loading on the beam due to the fluid. Secondly a mass was added to the plate mass to represent the inertia of the fluid. The applied pressure loading worked to a limited extent for the beams but no one pressure pulse shape gave good results for both end fixities. The best results for the plate problem were achieved with the added mass approach. Finally a number of areas of experimental and computational work are identified, which it is felt would benefit from further study.