Estimating the Mechanical Properties of Travertine Building Stones Due to Salt Crystallization Using Multivariate Regression Analysis

Abstract

Salt crystallization is one of the most powerful weathering agents that may cause a rapid change in the mechanical properties of stones, and thus limit their durability. Consequently, determining the mechanical properties of stones due to salt crystallization is important for natural building stones used in marine environmental and mild climatic conditions, which expose excessive salt crystallization cycles. In this study, multivariate regression analysis was performed for estimating the mechanical properties of travertine building stones after salt crystallization test. For this purpose, 12 travertine samples were selected and their physical and mechanical properties (density, porosity, uniaxial compressive strength, Brazilian tensile strength, and P-wave velocity) were determined. Then salt crystallization test was carried out at sodium sulfate solution (Na2SO4) up to 50 cycles and, after every 5 cycles, the uniaxial compressive strength, Brazilian tensile strength and P-wave velocity of the samples were measured. Using data analysis, regression equations were developed for estimating the mechanical properties of deteriorated samples at any cycle of the salt crystallization test. In these equations, the mechanical properties of the samples after salt crystallization were considered to be the dependent variable, which is dependent on the independent variables of the number of salt crystallization cycles, initial mechanical properties of the stones and their porosities. The validity of the equations was verified with the mechanical properties data of a researcher for salt crystallization test. The results showed that regression equations are in good accuracy for estimating the mechanical properties of stones, and thus making a rapid durability assessment.