Abstract:This study aims to make a unique contribution to the existing body of knowledge about rock strength and deformation parameters and crack stress thresholds through intelligent and statistical approaches applied to a database comprising various rock types (i.e., sedimentary, igneous, and metamorphic rocks). The database contains physical–mechanical and ultrasonic parameters. Six distinct machine learning (ML) algorithms— artificial neural network (ANN), random forest (RF), decision tree (DT), K-nearest neighbor (KNN), support vector regression (SVR), and bagging regressor (BR)— along with the conventional linear regression techniques, were employed to develop predictive models. These models estimate uniaxial compressive strength ( ) and Tangent Young's modulus ( ) based on bulk density ( ) and P-wave ultrasonic velocity ( ). Furthermore, they predict crack stress thresholds (i.e., crack closure stress , crack initiation stress , and crack damage stress ) as a function of , , , , axial strain at failure ( ), and lateral strain at failure ( ). Various performance indices were utilized to evaluate and compare the performance of these models. The results indicated that the RF method outperformed other ML-based and linear regression-based approaches in predicting the output parameters. Additionally, the multiple parametric sensitivity analysis (MPSA) was carried out to determine the significance of input parameters in predicting the output variables. This analysis revealed that and have the highest and lowest impact on predicting and , respectively. On the other hand, was identified as the most influential parameter in predicting and , while parameters and showed the least impact on the foregoing outputs, respectively. This is while and were, respectively, found as the most important and least important factors in predicting . Finally, to facilitate easy access to the prediction results and enhance the practicality of the proposed RF model, a graphical user interface (GUI) was developed, which enables the practical application of the most performing developed prediction model.

Crack detection using tap-testing and machine learning techniques to prevent potential rockfall incidents

A sound-based machine learning method for crack-type recognition in hard rock

Vision-based Crater and Rock Detection Using a Cascade Decision Forest

End-to-End Intelligent Framework for Rockfall Detection

Rock Crack Recognition Technology Based on Deep Learning

A Method for Automatic Assessment of Rockfall Susceptibility Based on High-Resolution Point Clouds

PointCrack3D: Crack Detection in Unstructured Environments using a 3D-Point-Cloud-Based Deep Neural Network

Detection of Cracking Levels in Brittle Rocks by Parametric Analysis of the Acoustic Emission Signals

Intelligent Approaches for Predicting the Intact Rock Mechanical Parameters and Crack Stress Thresholds

Image-Based Monitoring of Cracks: Effectiveness Analysis of an Open-Source Machine Learning-Assisted Procedure

A Skeleton-based Approach For Rock Crack Detection Towards A Climbing Robot Application

Artificial intelligence enhanced automatic identification for concrete cracks using acoustic impact hammer testing

Probabilistic Data-Driven Framework for Performance Assessment of Retaining Walls Against Rockfalls

Transfer learning framework for multi-scale crack type classification with sparse microseismic networks

Automatic Crack Classification by Exploiting Statistical Event Descriptors for Deep Learning

Research on crack monitoring at the trailing edge of landslides based on image processing

Crack Detection and Classification for Reinforced Concrete Structures using Deep Learning

Impact-Detection Algorithm That Uses Point Clouds as Topographic Inputs for 3D Rockfall Simulations

Method for rock fracture prediction and early warning: Insight from fusion of multi-physics field information

A data driven real-time perception method of rock condition in TBM construction

Crack Detection and Evolution Law for Rock Mass under SHPB Impact Tests