Probabilistic and Experimental Investigation of the Effect of Mineral Adsorbents on Porous Concrete Using Kriging, PRSM, and RBF Methods

Abstract

The use of porous concrete for paving urban pathways and reducing water pollution is particularly important. Pumice, scoria, zeolite, travertine, and metal sulfide nanoparticles can reduce water pollution. In this study, the effect of the volume percentage of several adsorbents, such as pumice, scoria, zeolite, travertine, and iron sulfide nanoparticles, as alternatives to concrete aggregates, was investigated with regard to the physical properties of concrete specimens, including compressive strength, porosity, and permeability coefficient. In order to study the effect of the adsorbents, a surrogate model trained by metamodel methods was used. Kriging, polynomial response surface method (PRSM), and radial basis function (RBF) methods were used as metamodel approaches. Kriging was found to have the least error in estimating compressive strength, porosity, and permeability coefficients. The failure probabilities of the performance functions defined by the kriging, PRSM, and RBF methods were estimated to be 0.0081, 0.0080, and 0.0019, respectively.