Transitioning a unidirectional composite computer model from mesoscale to continuum

Ballistic impact on composites has been a challenging problem as seen in the abundant literature about the subject. Continuum models usually cannot properly predict deflection history on the back of the target while at the same time giving reasonable ballistic limits. According to the authors the main reason is that, while continuum models are very good at reproducing the elastic characteristics of the laminate, the models do not capture the behaviour of the "failed" material. A "failed" composite can still be very effective in stopping a projectile, because it can behave very similar to a dry woven fabric. The failure aspect is much easier to capture realistically with a mesoscale model. These models explicitly contain yarns and matrix allowing the matrix to fail while the yarns stay intact and continue to offer resistance to the projectile. This paper summarizes the work performed by the authors on the computationally expensive mesoscale models and, using them as benchmark computations, describes the first steps towards obtaining more computationally effective models that still keep the right physics of the impact.