A simple approach to model SFRC

Experimental research that shows the improvement in structural behavior of concrete with the addition of fibers has been developed in the last years. Fibers control cracking and thus increase concrete toughness and ductility. Much effort has been devoted in the last decade to model this material.A simple homogenization approach based on a modified mixture theory is proposed in this paper to model Steel Fiber Reinforced Concrete (SFRC). The proposed and calibrated model takes information from the micro-scale to model the macro-scale. SFRC is considered as a composite material composed by concrete matrix and fibers. Concrete is modeled with an elastoplastic model and steel fibers are considered as orthotropic elastoplastic inclusions that can debond and slip from the matrix. In order to include this inelastic phenomenon without explicitly modeling interface, constitutive equations of fibers are modified including information from the debonding–slipping phenomena. The model requires concrete properties, fibers material, geometry, distribution and orientation as input data. The fibers bond–slip behavior is automatically derived from concrete properties and fibers geometry or it can be alternatively obtained from pull out tests.As illustration, the tension response of SFRC with different fiber contents is numerically simulated. The model is verified with the results of bending tests of beams extracted from a SFRC slab that present different fibers distribution due to the slab casting process. Comparisons with other numerical approaches modeling SFRC as an equivalent homogeneous material are also included in the paper.

► A homogenization approach for Steel Fiber Reinforced Concrete (SFRC) is proposed.
► It presents advantages in comparison with equivalent isotropic homogeneous models.
► The model takes information from the micro-scale to model the macro-scale.
► Fiber content, distribution and pull out response are explicitly taken into account.
► Flexure response of SFRC beams with variable fiber content is properly modeled.