| Ceramic materials have become increasingly prevalent as components of armour systems to provide ballistic protection to a variety of potential targets, from personnel to armoured vehicles. In order to characterize the ballistic performance of, i.e. the protection offered by, armour materials in general, and ceramic materials in particular, depth of penetration (DOP) testing has become a common methodology. The interaction between projectile and ceramic armour components, however, is one that is still not completely understood and no one approach has fully described all the complex processes involved. Artificial neural networks are analytical tools that can identify non-linear patterns between input and output values, by learning complex, non-linear relationships.; Data collected from DOP testing, primarily at Defence Research and Development Canada-Valcartier, has been used to develop an artificial neural network model that was able to predict the performance of ceramic armour materials (specifically, Al2O3, SiC, B4C, and TiB2). The network provided good results (prediction of DOP) in the form of a general model that interpolated well over the input space of the specific experimental conditions and ranges, where there were sufficient data. In areas of limited data, not unexpectedly, the network performed less well. The network, employing the back propagation paradigm with a single hidden layer architecture and considering up to 39 input variables and parameters, was able to establish a strong correlation between predicted DOP and impact velocity, ceramic density, ceramic thickness, projectile length, projectile diameter, and projectile mass. |
