Probabilistic Prediction of Ground Reaction Curves in Poor Rock Masses Using MARS

Document Type : Original Article

Authors

1 Assistant Professor, Department of Civil Engineering, Shahid Chamran University of Ahvaz, Iran

2 MSc student, Department of Civil Engineering, Shahid Chamran University of Ahvaz, Iran

Abstract

Tunnels are recognized as critical infrastructure components in modern so-cieties. The optimal design of tunnels requires a comprehensive understand-ing of their behavior under various geotechnical conditions. The ground re-action curve (GRC), which describes the relationship between tunnel con-vergence and support pressure, is a fundamental element of the conver-gence–confinement method in tunnel design. To date, numerous approaches have been proposed for constructing ground reaction curves, most of which are deterministic in nature. In the present study, a probabilistic framework is developed to derive ground reaction curves for shallow tunnels excavated in poor rock masses. For this purpose, appropriate probability distributions were assigned to the geomechanical properties of the considered rock class. Subsequently, 100 stochastic realizations were generated from these distribu-tions. All cases were analyzed using numerical modeling based on the finite element method, and the corresponding ground reaction curves were ob-tained. Finally, a predictive model for the ground reaction curve was estab-lished using multivariate adaptive regression splines (MARS). Statistical evaluation indicates that the proposed model achieves a coefficient of deter-mination (R²) of 0.982, demonstrating high predictive accuracy. Further-more, the results reveal that the at-rest lateral earth pressure coefficient (K₀) has a negligible influence on the resulting ground reaction curves.

Keywords

Main Subjects

Send comment about this article
CAPTCHA Image
Civil Infrastructure Researches

Articles in Press, Accepted Manuscript
Available Online from 07 June 2026

Files

Share

How to cite

Statistics