Prediction of Sulfate Ion Penetration in Concrete Containing Nano-Silica and Micro-Silica Using Machine Learning

Abstract

Global temperature changes have been increasing in recent years. This has caused damage caused by chemical compounds, especially sulfate ions, to increase in coastal and offshore concrete structures. Sulfate attack is a major cause of concrete durability deterioration. Mass loss, strength reductions, and expansive strain of concrete specimens are generally its effects. Supplementary cementitious materials (SCMs) affect the mechanical and chemical properties of concrete. In this research, specimens with Silica fume and Nano silica were built. Then sodium sulfate solution (10%) with three different temperatures of 298, 303, and 308 ◦K and drying-wetting cycles with a ratio of 3:4 for 360 days (51 cycles) have been evaluated. Sample measurements, mass changes, sulfate concentration profile in samples, and compressive strength were monitored over a 360-day period. Scanning Electron Microscopy (SEM-EDS), X-ray Diffraction (XRD), and Thermal gravimetric Analysis (TGA) are also used to analyze the microstructure of concretes after exposure to corrosive environments. Finally, with the help of machine learning and the Multi-Layer Perceptron (MLP) algorithm, the concentration of sulfate ion penetration has been predicted. The obtained results showed that with increasing temperature, the intensity of sulfate ion penetration into concrete increases. This intensity decreased with the increase in the percentage of cementitious materials. The values of statistical measurements such as R2, RMSE, and MSE determinated that for instance 0.9821 and 0.9741 for SF10 and nS3 respectively, which shows the high correlation and accuracy of the algorithm development.

Keywords: Sulfate concentrationMachin Learning (MLP)SCMsWet-Drying Cycles