Prediction of flow-induced fiber orientation and its effects on hardened behavior of ultra-high performance concrete: An experimental and numerical study

Ultra-high-performance concrete (UHPC) is a fiber-reinforced cementitious composite known for its superior strength and durability, but it is also highly brittle. Short fiber reinforcement is crucial for imparting tensile capacity, with fiber orientation significantly enhancing mechanical behavior. In this paper, fiber orientation due to flow during casting is predicted using a computational fluid dynamics (CFD)-based stochastic fiber orientation model, which is commonly used to predict fiber orientation during the processing of fiber-reinforced polymer composites. The resulting stochastic fiber orientation serves as an input to a mesoscale discrete numerical model, specifically the lattice discrete particle model for fiber-reinforced concrete (LDPM-F), to further predict the solid mechanical response of the UHPC composite elements. To validate the numerical model, a novel nozzle-based casting method was developed to align fibers in the direction of UHPC flow. The nozzle design features gravity-induced converging flow with a nozzle exit width slightly smaller than the fiber length. To evaluate the fiber orientation efficiency, slabs were cast using layered deposition of UHPC on a moving mold. The slabs were tested under a three-point bending scheme and compared to randomly cast slabs. The results showed a significant enhancement in the flexural tension capacity of the aligned specimens. Additionally, direct tension tests were conducted on notched prismatic specimens cut from both the edges and center of the slabs, in parallel and perpendicular directions relative to the casting flow. Higher tensile capacities were observed for all samples cut in the parallel direction of casting compared to those cut perpendicularly or from randomly cast slabs. Subsequently, the experimental data was used to predict the stochastic fiber orientation state during casting via nozzle flow. This orientation state was used as a calibration parameter to inform the LDPM-F model, enabling simulations of the axial tension tests on notched prismatic samples and the three-point bending tests on slabs. The numerical simulations showed very good agreement with the experimental results. Collectively, this numerical framework serves as a virtual simulation platform to predict the impact of flow-induced fiber anisotropy on the mechanical properties of UHPC.