A hierarchical local-model tree for predicting roof displacement in longwall tailgates

Abstract

An explicit challenge for mine productivity and personnel’s safety in mechanized underground mines is to envisage the uncontrolled roof displacements, usually occurred in underground roadways. For this purpose, support systems are to be installed to guarantee the proper functionality of underground structures during mining operations. Roof displacement is a reliable indicator to examine stability conditions of tunnels or roadways, which are driven in underground coal mines. In this research, a neuro-fuzzy-based method, namely hierarchical local model tree (HiLoMoT), was employed to predict the roof displacements and consequently indicate the unstable zones in a tailgate roadway. In this regard, the geomechanical and instrumentation information measured from a longwall panel at Tabas coal mine was employed to validate the developed HiLoMoT model. According to the results, the proposed HiLoMoT model could predict roof displacements in reasonable conformity when compared with measured ones. In order to examine the prediction capability of the HiLoMoT model, three indices of the coefficient of determination (\n $$R^{2}$$\n ), variance accounted for (VAF), and root mean square error (RMSE) were used. Introducing unseen test data, $$R^{2}$$\n , VAF, and RMSE were, respectively, obtained as 0.952, 95.13, and 0.0193, which showed high goodness of fit and low error for the proposed model. In comparison with ANFIS as a common fuzzy-based model, HiLoMoT could predict the roof displacement by a nonlinear partitioning based on the incremental tree-construction, which improves the quality of the model without further iteration loops or trial and error. Therefore, the HiLoMoT model may be implemented as a new applicable tool for predicting roof displacements ahead of time in mechanized longwall mining.