Evaluation of dynamic tensile strength of concrete using lattice-based simulations of spalling tests

The mechanical properties and failure modes of concrete are affected by strain rate. Therefore, various experimental methods have been used to quantify the strain-rate dependency of concrete properties, such as the spalling test by using a Hopkinson bar. In the test, pullback velocity at the free-end surface of the specimen is usually measured to evaluate the dynamic tensile strength of concrete, instead of directly measuring the critical stresses at the damaged location(s) due to experimental constraints. Herein, such indirect measurements of tensile strength are compared with direct determinations of tensile strength based on lattice models of the spalling tests. To represent rate-dependent material behavior, rheological units are introduced within the lattice elements. The parameters of the rheological units are calibrated through comparisons with experimental data. The calibrated values remain unchanged in subsequent simulations, which can be regarded as virtual spalling tests at various high strain rates of loading. The separation of viscous and inertial contributions to apparent tensile strength provides insights into the dependence of actual tensile strength on high strain rates. The simulation results indicate indirectly measured dynamic tensile strength, as commonly done in practice, is sufficiently close to directly measured strength.