Title: Estimation of thermal conductivity of rocks through artificial neural network modelling based on physico-mechanical properties

Authors: A.K. Tripathi; Gurram Dileep; S.K. Pal; Aman Raj

Addresses: Department of Mining Engineering, Faculty of Mining Engineering, National Institute of Technology Karnataka, Surathkal, 575025, Karnataka, India ' Department of Mining Engineering, Adhoc Faculty of Mining Engineering, Visvesvaraya National Institute of Technology, Nagpur, 440010, India ' Department of Mining Engineering, Faculty of Mining Engineering (Retd.), Indian Institute of Technology Kharagpur, Kharagpur, 721302, West Bengal, India ' Indian Institute of Technology Kharagpur, Kharagpur, 721302, West Bengal, India

Abstract: Accurate estimation of rock thermal conductivity is vital in geotechnical, geothermal, and mining engineering for underground design, thermal modelling, and energy storage applications. Conventional laboratory methods, though accurate, are time-consuming and impractical for large datasets or field conditions. This study explores artificial neural networks (ANN) to predict rock thermal conductivity using measurable physico-mechanical properties such as density, porosity, P-wave velocity, and uniaxial compressive strength (UCS). A feedforward backpropagation ANN model was developed and trained in MATLAB using experimental data. Model performance was evaluated using correlation coefficient (R), mean squared error (MSE), and regression analysis. The ANN architecture was optimised by varying hidden neurons from 5 to 15, with optimal performance at 12 neurons. The model achieved high R-values for training (0.98361), validation (0.94635), testing (0.95408), and overall R of 0.97679, with consistently low MSE values. Results confirm ANN as an accurate, efficient, and reliable alternative to conventional methods.

Keywords: thermal conductivity; artificial neural network; ANN; rock properties; P-wave velocity; porosity; density; uniaxial compressive strength; UCS; predictive modelling.

DOI: 10.1504/IJMME.2026.154344

International Journal of Mining and Mineral Engineering, 2026 Vol.17 No.2, pp.125 - 140

Received: 05 Aug 2025
Accepted: 22 Oct 2025
Published online: 23 Jun 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article