Prediction of the mechanical compressive behavior of granite using intelligent tools

Abstract

This paper aims to apply intelligent tools such as artificial neural networks, support vector machines and multi-ple regression to forecast the main parameters characterizing the compressive behavior of granites, namely the resistance under compression, fc, the crack initiation stress, fci, and the crack damage stress, fcd, based on physi-cal parameters like density, ρ, porosity, η, and ultrasonic pulse velocity (UPV). The granitic rocks selected are from the north region of Portugal existing in ancient masonry structures. Several experiments were performed to build a database of 55 records containing the mechanical and physical parameters mentioned above. The predictive capacity of the models was evaluated using the coefficient of correlation, R, and the root mean square error, RMSE. The results showed a good predictive capacity of the developed models.