A prediction model for surface deformation caused by underground mining based on spatio-temporal associations

Abstract

Accurate predictions of the surface deformation caused by underground mining are crucial for the safe development of underground resources. Although surface deformation has been predicted by artificial intelligence (AI) methods, most AI models are established based on the relationships between surface deformation and influential factors. The lack of consideration of the deformation state transition often leads to errors in the prediction results of catastrophic deformation by conventional AI methods. In this respect, this study introduces a surface deformation prediction model based on spatio-temporal association rule mining (STARM). Surface deformation is classified as excessive deformation zone (EDZ) and hysteretic deformation zone (HDZ), representing different surface deformation stage or state. The spatio-temporal association rules between the monitored EDZ and HDZ data are then mined. A surface deformation prediction model is established according to the spatio-temporal relationship between monitored EDZ and HDZ data. The proposed model is verified based on a practical case study of the Chengchao Iron Mine in China. The data collection of the influential factors is not requisite for the proposed model. It can achieve accurate prediction of the catastrophic deformation that was characterized by deformation state transition.