Prediction of subsidence due to underground mining by artificial neural networks

Abstract

Alternatively to empirical prediction methods, methods based on influential functions and methods based on mechanical model, artificial neural networks can be used for the the surface subsidence prediction. In our case, the multi-layer feed-forward neural network was used. The training and testing of neural network is based on available data. Input variables represent extraction parameters and coordinates of the points of interest, while the output variable represents surface subsidence data. After the neural network has been successfully trained, its performance is tested on a separate testing set. Finally, the surface subsidence trough above the projected excavation is predicted by the trained neural network. The applicability of artificial neural network for the prediction of surface subsidence was verified in different subsidence models and proved on actual excavated levels and in levelled data on surface profile points in the Velenje Coal Mine.

Influence of model parameter uncerainties on forecasted subsidence

“…The second group includes the profile function (Asadi et al, 2004;Díez and Álvarez, 2000;Torano et al, 2000;Waddington and Kay, 1995), graphical method (NCB, 1975), application of artificial neural networks (Ambrožič and Turk, 2003;Gruszczyński, 2007;Weifeng and Xiaohong, 2013) and influence function (Álvarez-Fernandez et al, 2005;Cui et al, 2013Cui et al, , 2000Ghabraie et al, 2017;Huayan et al, 2002;Liao, 1993;Nicieza et al, 2005;Ren et al, 1987Ren et al, , 2014Sheorey et al, 2000;Singh and Singh, 1998). Based on large collections of observations, the relationships between measurement results and geometric conditions from a mining operation are determined.…”

Fossil karst in the Jurassic of the Kościuszko Mound in Kraków (southern Poland) – a discussion

“…Newer model stud ies con ducted on cen tri fuges (e.g., Craig, 1990;Iglesia et al, 1990;Abdulla and Goodings, 1996;Costa et al, 2009) en abled a better un derstand ing of the mech a nism of sink hole for ma tion and im plicated the ap pli ca tion of an other sci en tific area, i.e. an a lyt i cal and nu mer i cal mod el ling, for sink hole anal y sis (e.g., Tharp, 1999;Ambroziae and Turk, 2003;Augarde et al, 2003). Hav ing dis cussed the ap proaches to the in ves ti ga tions of sink hole ar eas, the at ten tion will be now drawn onto the sub ject of this re search.…”

A new multi-gene genetic programming approach to non-linear system modeling. Part II: geotechnical and earthquake engineering problems

“…Progress has recently been made in the ability to predict ground movement, but the state of the art is deficient in many ways. ANNs have recently been applied to the prediction of tunneling-induced ground movement [30,31]. Li et al [32] utilized a hybrid fuzzy and tree-based GP method to analyze the actual cases of excavation, mining, and ground surface movement.…”