Risk analysis and prediction of out-of-seam dilution in longwall mining

• Published in: International journal of rock mechanics and mining sciences (Oxford, England : 1997) England : 1997), 2014-09, Vol.70, p.115-122
• Main Authors: Bahri Najafi, Ali, Saeedi, Golam Reza, Ebrahimi Farsangi, Mohammad Ali
• Format: Article
• Language: English
• Subjects: Applied sciences; Artificial neural networks; Buildings. Public works; Computation methods. Tables. Charts; Dilution; Exact sciences and technology; Geotechnics; Kerman and Tabas coal mines; Learning theory; Longwall; Longwall mining; Mathematical models; Mining; Miscellaneous; Neural networks; Out-of-seam dilution; Risk; Risk analysis; Rock engineering systems; Soil mechanics. Rocks mechanics; Structural analysis. Stresses
• ISSN: 1365-1609; 1873-4545
• DOI: 10.1016/j.ijrmms.2014.04.015

A rock engineering systems (RES) based model is defined to predict the level of risk due to out-of-seam dilution (OSD) in longwall faces. Also, based on the level of risks obtained, a predictive model for the OSD is proposed. Furthermore, an artificial neural networks (ANN) model is developed to predict the OSD. The data collected from thirty-five longwall faces are used to carry out the risk analysis and to develop the predictive models. The results obtained show that the level of risk achieved for each face is in consistence with the corresponding OSD calculated. Also, coefficient of determination (R2) and root mean square error (RMSE) for the ANN model (R2=0.98, RMSE=1.24) and the RES-based model (R2=0.86, RMSE=3.74) have been obtained. These show the good performances of both models. However, the ANN model has a better performance than the RES-based model. •An RES-based model was developed to predict the level of risk due to the out-of-seam dilution (OSD) in longwall faces.•A predictive model for the OSD was proposed based on the level of risks.•An ANN model was developed to predict the OSD.•For both RES and ANN models, the level of risks estimated is in consistence with the corresponding OSDs calculated.•R2 and RMSE indicators for the ANN and RES-based models show the better performance of the ANN over the RES-based model.