ABSTRACT
This study examines the strength of siltstone from an underground mine in Victoria, Australia under four saturation levels and assesses the use of artificial neural networks (ANN) to predict the strength. Oven-dry siltstone shows an average uniaxial compressive strength (UCS) of 70 MPa, with saturation reducing UCS by 14%–32%. Brazilian tensile strength (BTS) drops from 17 MPa by 44%–49%, while axial and diametral point load strengths (Is50) fall from 11 MPa and 9 MPa by 37%–40% and 22%–31%, respectively. Weak planes controlled the strength in about 15% of samples. ANN models effectively estimated UCS, BTS, and Is50 from parameters like depth, size, density, and saturation, offering a strong complement to lab testing and enhancing understanding of siltstone behavior.
