Scaling of inelastic structures loaded dynamically.

This paper assesses the current situation on the scaling of geometrically similar structures subjected to large dynamic loads which cause an inelastic material response. It appears that the wholly ductile dynamic response (without fracture) of small scale prototypes might conform to the principles of elementary geometrical scaling, except for the well known size effect associated with materialstrain rate sensitivity, which might be exacerbated in structures with load deflection curves which fall after a peak value. This conjecture is supported by experimental tests while dynamic tests would not be likely to depart significantly from geometrical similitude. However, the existence of ductile brittle transitions which might occur in other test programmes cannot be predicted with any degree of certainty when varying specimen thicknesses. The dynamic tests on two quite different mild steel structures, and Vaughan's empirical predictions on the dynamic tearing of steel plates, show a significant departure from the predictions of geometrically similar scaling. The full scale prototypes suffer larger deformations and greater tearing than predicted from experimental results on small scale models using the principles of geometrically similar scaling. Thus, geometrical scaling is not obtained when tearing, cutting or ductile brittle transitions occur during a structural response. It is evident that further studies are required to predict the ductile brittle transitions and to quantify any departure from geometrical scaling.