Document Type: Research Paper
Department of Mining Engineering, Faculty of Engineering, University of Kurdistan, Sanandaj, Iran
School of Mining Engineering, College of Engineering, University of Tehran, Tehran, Iran
The paper describes an artificial neural network (ANN) model to predict the height of destressed zone (HDZ) which is taken as equivalent to the combined height of caved and fractured zones above the mined panel in longwall mining. For this, the suitable datasets have been collected from the literatures and prepared for modeling. The data were used to construct a multilayer perceptron (MLP) network to approximate the unknown nonlinear relationship between the input parameters and HDZ. The MLP proposed model predicted values in enough agreements with the measured ones in a satisfactory correlation, in which, a high conformity (R2=0.989) was observed. To approve the capability of proposed ANN model, the obtained results are compared to the results of the conventional regression analysis (CRA) method. The calculated performance evaluation indices show the higher level of accuracy of the proposed ANN model compared to CRA. For further evaluation, the ANN model results are compared with the results of available models and in-situ measurements reported in literatures. Comparative results present a logical agreement between ANN model and available methods. Obtained results remark that the proposed ANN model is a suitable tool in HDZ estimation. At the end of modeling, the parametric study shows that the most effective parameter is unit weight whereas elastic modulus is the least effective parameter on the HDZ in this study.