Improved Model for Three-Dimensional Virtual Concrete: Anm Model
Edward J. Garboczi, Yang Lu, Stephen Thomas
Construction aggregate particles, fine or coarse, can be mathematically characterized using spherical harmonic series and used to simulate random parking of irregular aggregates to form a virtual mortar or concrete using the Anm model. Any other similar composite system, of irregular (star-shaped) particles in a matrix, can also be simulated. The task of modeling the random nature of the concrete microstructure is a computationally expensive task even using regular particle shapes such as spheres. The use of real-shaped aggregates makes it even more complex and time consuming, since detecting the contact of irregular shape particles is more complicated than spheres or ellipsoids. This paper integrates two new algorithms into the Anm model. The first new algorithm is the Extent Overlap Box (EOB) method that detects inter-particle contact, and the second is the capability of adding a uniform thickness shell to each particle. Parameter analysis has shown that the EOB method leads to a more accurate detection of inter-particle contact with a smaller computational cost than the previously used Newton-Raphson method. The uniform thickness shell provides a customizable tool to control the minimum inter-surface distance of particles during the parking process, as well as simulate processes and microstructure that are dependent on the Euclidean distance from a particle surface. For mortar and concrete, the uniform thickness shell can represent the observed interfacial transition zone (ITZ) structure. A parallel processing application programming interface (API) was integrated into the Anm model to accelerate the particle placement process by parallel optimization, which results in significant improvements in the packing efficiency on multicore processor systems. This significant speed-up, as well the improved contact function and new uniform thickness shell algorithm, greatly extends the range, size, and type of particle systems that can be studied.
Computer-Aided Civil and Infrastructure Engineering
, Lu, Y.
and Thomas, S.
Improved Model for Three-Dimensional Virtual Concrete: Anm Model, Computer-Aided Civil and Infrastructure Engineering, [online], https://doi.org/10.1061/(ASCE)CP.1943-5487.0000494
(Accessed June 1, 2023)