Auto-generation methodology of complex-shaped coarse aggregate set of 3D concrete numerical test specimen

摘要

In this paper, a novel numerical method for random generation of aggregates in concrete is presented. Compared with some other models, the aggregates model generated by this method offers a better approximation to the geometric shape of real aggregates and the target gradation. The research incudes that (1) a surface reconstruction method is developed first on the basis of the implicit T-spline surface algorithm through defining a modified knot vector, determining the off-set point locations and their sign distances, which aims at avoiding occurrence of unclosed curved surface or spurious sheets and reconstructing accurately the complex surface according to a set of scattered surface points of a gravel aggregate; (2) a sinusoidal generatrix function, of which the amplitude represents the aggregate size range and the period reflects the aggregate flatness, is proposed to auto-generate a set of scattered points on the surface of a gravel aggregate. Further modifications are made to improve the quality of the scattered points that agree well with the real crushed aggregate surface; (3) An efficient aggregate packing method is proposed by combining and modifying the "occupation and removal method" and the "layering disposition method" to improve the efficiency of aggregate packing; and (4) a MATLAB computing program for auto-generating aggregates in concrete is developed and validated by examples. The simulation results have demonstrated that the method presented in this paper can generate aggregates for any given mix proportions and gradations and the method can serve as an effective tool for numerical evaluation of mechanical properties of concrete materials. (C) 2019 Elsevier Ltd. All rights reserved.