Concrete cohesive curves for specimens with different sizes: A study of inverse analysis and size effect

Abstract

Inverse analysis is a possible and adequate manner to obtain a cohesive curve that mathematically characterizes a concrete specimen under tension stresses. However, there are concerns about the uniqueness of this result. Literature reports that the cohesive curve depends on the size of the specimen, the so-called size effect. Researches usually work with limited sets of specimens and, associated with the heterogeneity of concrete, such sets are not always representative. This work proposes to determine the possibility of obtaining a single cohesive curve that describes a set of 128 concrete specimens in three-point bending tests of different sizes and notch depths. A numerical procedure is implemented in an inverse analysis program for the determination of the cohesive curve in a finite element model. The analysis process is separated into 4 steps. Step 1 works with the individualized specimens to characterize the cohesive curves of each size. Steps 2 and 3 work with the specimens grouped according to the size and notch depth. Finally, Step 4 adjusts a single cohesive curve for all specimens. The final comparison of all experimental P-CMOD curves with the simulated P-CMOD curves shows a good agreement, demonstrating the feasibility of a single cohesive curve approach.