Quevedo, R, Sari, YA & McKinnon, S 2024, 'Machine learning framework application for modelling geomechanical instabilities: a caving case study', in P Andrieux & D Cumming-Potvin (eds), Deep Mining 2024: Proceedings of the 10th International Conference on Deep and High Stress Mining, Australian Centre for Geomechanics, Perth, pp. 853-866, https://doi.org/10.36487/ACG_repo/2465_53
(https://papers.acg.uwa.edu.au/p/2465_53_Quevedo/)
Abstract: The field of machine learning (ML) has had a significant impact on, and adoption in, many fields of science and engineering, yet in mining is still not very well developed, with many publications exploring scopes of application and potential areas of integration.
As mining reaches deeper environments where most traditional methods of stability analysis have yet to be calibrated, there are good opportunities to apply ML methods to diverse types of, for example, failure phenomena. Still there is a necessity to properly account for adequate data inclusion and problem definition to apply these kinds of analysis, which is why data representation and availability with regards to a particular problem are crucial
In this paper a case study of the application of data extraction and the ML modelling process applied to rock mass failure phenomena taking place in an underground cave mine is presented as an illustrative example of the practical application of ML methods of analysis in mining.
The results show that ML methods have high potential in mining applications when coupled with careful consideration of input variables and the correct choice of ML approaches. The fundamentals and practical aspects are outlined such that the methodology of the case study is generalisable to different kinds of geomechanical problems.

Keywords: machine learning, mine design, caving mechanics, caving projects