Numerical Simulation of Mechanical Effects in Composite Structures by the Finite Element Method

Abstract

In electromagnetic railgun experiments at ISL, the composite projectiles are accelerated up to 2000 m/s with a maximal acceleration of 7×106 m/s2. The maximum forces typically will have been achieved after 200 μs. One consequence of this dynamic load is a strain rate in the order of ε̇=2000 s−1, a rate at which metals are also giving up their static properties. The aim of research at ISL is the construction of a fiber-reinforced sabot adapted for dynamic bearing. FE simulations support the construction process as well as allowing a detailed study of the mechanical effects occurring during the acceleration of the sabot. The most important problem of dynamical simulation of composites is an adequate material model. We show in this paper, that the classical homogenized laminate models are insufficient to be used in dynamics. To develop an appropriate material model, we investigated selected materials statically as well as dynamically and we compared the experimental results with those of their numerical simulation. For a first material model approach, we made quasi-static experiments with a special sabot geometry. The quasi-static investigations have revealed that different conditions, e.g., lateral support, involve a completely different system answer. We found that a specimen without radial strain has more than the double strength of a specimen that allows radial strain, whereby the system answer still remains linear elastic and only very little internal delamination occurs. A second step toward a material model consists of dynamic experiments with specimens in a simple cylinder geometry on a Split Hopkinson Pressure Bar test facility. The specimens are made of fiber-reinforced epoxy, as well as metal-matrix composites. The dynamic experiments proved the strain as principal failure criterion for dynamic bearing, whereby the failure strain depends upon the momentary strain rate. Failure strain is also a historical function, so that the deformation history is important. We can show that the viscous polymer matrix plays a very important role for the dynamic toughness of reinforced materials. When the material fails, the stress is far beyond that of static failure: up to two times the static strength has been measured. Experiments with real sabots in railguns have shown that the failure occurs a relatively long time after the stress has achieved its maximum.