Study of the ballistic limit of carbon fiber laminates throughout MEF and neural networks methodologies combined

Abstract

Nowadays composite materials are widely used in aeronautics due to its high resistance-density relation. However, one of its main drawbacks is its elevated vulnerability against perpendicular impacts. The ballistic limit value of each laminate is a critical variable playing an integral role in the aforementioned impacts. The aim of this project is to analyze the influence of the laminates stacking sequence in its ballistic limit values. To this end, a numerical model through a commercial code of finite elements in LS-DYNA® will be developed and implemented to be compared and validated through an previous experimental model from [López-Puente et al., 2003][1]. The experimental model was carried out with impact trials of high and low velocity of spherical projectiles on carbon fiber test tubes with specific geometries and boundary conditions. The composite material utilized is AS4 carbon fiber and 8552 epoxy matrix, an amply used material in aeronautics. For a thorough interpretation and analysis of the influence between both parameters it is intended the usage of a new technology based in neural artificial networks through a code implemented in Python. These computational tools are increasing in use in the engineering industry due to its features of behavior prediction between variables without an apparent either particular or mathematizable physical relation.