Model Prediction of Fracture Due to Explosions

Evidence indicates that small model systems using the same material, same explosive, and similar geometry give good quantitative predictions of strains in a number of rocks, but only qualitative predictions of fracture. It appears that an equation of state that is applicable to blasting in rocks does not involve strain rates, but that fracture is time dependent. Fracture caused by rapidly applied loads cannot, in general, be predicted by assuming that materials fail at a certain state of stress or strain. Some additional property of the material must be known. The consistency between nonsimilarity of fracture in model systems, and the existance of such a property is pointed out. Experiments demonstrating variation in time required to complete fracture of plaster bars with applied stress are described. A mechanism of fracture is proposed that accounts for variation in time required to complete fracture and for the difference in static and dynamic ultimate stresses that have been reported by various investigators. Some predictions that follow from this mechanism of fracture are presented.