deformation ._.....,.., .............. ..,. "''"-'·..,· lS is not the case C'tiH"I'""'"' have shown that (1992), Markeset (1994), Vonk (1993), van Mier (1997). Due to the localization phenomenon the complete stress-strain curve becomes specimen size dependent Most load-bearing elements, such as…
deformation ._.....,.., .............. ..,. "''"-'·..,· lS is not the case C'tiH"I'""'"' have shown that (1992), Markeset (1994), Vonk (1993), van Mier (1997). Due to the localization phenomenon the complete stress-strain curve becomes specimen size dependent Most load-bearing elements, such as beams, columns and slabs, experi- ence strain gradients. As localization of deformations occurs when the failure stage is approached, it is clear that one single stress-strain curve obtained from centric loading test will not realistically represent the behaviour of concrete subjected to strain gradients. Improperly characteriz- ing the post peak behaviour of concrete can lead to poor and unsafe prediction of the ductility and ultimate axial deformation of columns and the rotational capacity, balanced steel ratio, and ultimate moment capacity of beams. The observed size effect on the softening branch and its influence on the analysis of concrete members subjected to compressive strain gradients is discussed in this paper. 2 Strain softening and size effect From a Round Robin test, (van Mier et al. (1997)) it was found that the pre-peak stress strain curve was independent of the slenderness when low- friction loading platens were used. In the softening regime, however, an increase of ductility in terms of stress and strain with decreasing specimen slenderness was found in all experiments.