Abstract

Double punch test is used to indirectly assess the tensile strength of plain concrete, ${\displaystyle f_{t}}$. For this normalized test, the tensile strength is obtained as a function of the failure load, ${\displaystyle P}$, which is expressed as ${\displaystyle f_{t}=F(P)}$. Different authors have proposed different expressions for the relation ${\displaystyle F(\cdot )}$, yielding scattered values of ${\displaystyle f_{t}}$. None of these alternatives is universally recognized as being more suitable than the others. In fact, these expressions are mainly based on elastic models considering the maximum tensile stress under the load ${\displaystyle P}$ and ${\displaystyle f_{t}}$ is obtained as an output of the linear model. A numerical simulation allows using models in which ${\displaystyle f_{t}}$ is an input of the material model and the corresponding failure load ${\displaystyle P}$ is obtained associated with each value of ${\displaystyle f_{t}}$. In the present work, double punch test is simulated numerically considering two alternatives for modeling plain concrete accounting for damage and cracking: (a) the nonlocal Mazars damage model and (b) an heuristic crack model including joint elements in an a priori defined crack pattern. Numerical results are validated with experimental data and compared with the analytical expressions available in the literature.

Document information

Published on 30/10/19

DOI: 10.1016/j.ijsolstr.2011.01.006