Abstract: To enhance the performance of ultra-high performance concrete (UHPC), an improved Andreasen & Andersen particle packing model was employed to design the matrix mix. Using image analysis method, the study examined the influence of different matrix viscosities on the distribution of steel fibers in UHPC. By judiciously combining compressive strength considerations with fiber distribution coefficients, an optimal matrix viscosity was selected. The research further investigated the effects of steel fibers with varying shapes and volume fractions on mechanical properties, ultimately determining the optimal mix proportions for achieving the best mechanical performance. The test results show that achieving an appropriate viscosity in the UHPC matrix, through uniform steel fiber dispersion while minimizing internal air bubbles, further enhances UHPC compressive strength. Corrugated steel fibers exhibit superior dispersion compared to hooked and straight fibers at the same viscosity, resulting in a more pronounced improvement in the mechanical properties of the UHPC matrix. The optimal mix proportions for UHPC are as follows: a water-to-binder ratio of 0.16, a mass ratio of cement to silica fume to slag at 0.75 ∶0.2 ∶0.05, a 3% volume fraction of corrugated steel fibers, a superplasticizer mass fraction of 0.8%. At this point, the powder particles exhibit the closest packing, steel fibers are uniformly distributed, resulting in the optimal performance of UHPC.

Key words: ultra-high performance concrete, close particle packing theory, steel fiber, pull-out, compressive strength, image analysis method