A Markov Chain Monte Carlo Technique Based Optimal Mix Design of Porous Concrete

Abstract

Porous concrete is one of the innovative and promising concrete products, which is featured with a relatively high water permeability rate. Compared with conventional concrete products, due to the lack of fine aggregates in the mix design of porous concrete, the void spaces between the coarse aggregates remains unfilled and causes a large amount of porosity in the hardened concrete mass. On the other hand, the strength of porous concrete is usually lower than that of the conventional concrete products due to the lack of fine aggregates. For the purpose of achieving a relatively high strength of porous concrete while maintaining a good permeability of pavements, the mix design of porous concrete is modeled as a Markov Chain Monte Carlo (MCMC) system and a Gibbs Sampling method based approach is developed to approximate the optimal mix design. The simulation results show that, by using the proposed approach, the system converges to the optimal solution quickly and the derived optimal mix design achieves the tradeoff between the compressive strength and the permeability rate.