Abstract
The present study reports a computational methodology for predicting the barely visible impact damage (BVID) in carbon fibre-reinforced composite panels subjected to low velocity impact damage from a blunt nose impactor. Since the fibre-reinforced composite structures are susceptible to out-of-plane impact loading that occurs in many forms affecting the structural performance in differing degrees. In particular the BVID is invisible on the impact surface but lead to severe strength and stiffness degradation.
The extent of BVID damage is difficult to measure as it cannot be seen by visual inspection and damage is hidden among the layers that cause inter-laminar strength reduction which induces secondary invisible effects. Relationships between forces and deformations are much more complicated in composites than in conventional metallic materials due to composites? higher transverse shear and transverse normal stress deformability. Modelling of impacted composites becomes more difficult when a ply suffers damage. It affects performance of the whole laminate generating an iterative process each time a local ply or part of its material is degraded and continues until ultimate failure.
In order to model such material damages the approaches evolved in classical and modern layer-wise theories are based on impact damage initiation and propagation correlated with stress concentrations and deflected regions or damage between layers is introduced through configuration of damage to the idealized interfacial layer. The correlation and link between stiffness degradation and deflection as well as damage induced plies paves the way to develop the computational model in this work.
The model's ability to predict added response indicated a capability to predict progressive damage and design parameters. The results of static load-deflections for identifying induced damage areas are presented in the form of fully simulated images. Impacts due to flat nose impactors are considered appropriate when impact energies are at moderate levels and the back surface of the specimen is not protruded.
The computed and the available experimental results are compared and found to be in acceptable agreement for design work. The model could be applied to similar problems for further studies and design improvements. The methodology used here is efficient and also reduces the need for expensive and lengthy experimental testing.