Application of Fuzzy Logic for Prediction of Roof Fall Area in the Goaf in Coal Mines

Abstract

Roof falls are common problems in underground coal mines which may have serious effects on miners in the form of injuries, disabilities or fatality as well as other damages like production delay, interruptions in mining operations, equipment breakdown events. Roof fall and side fall accidents in coal mines alone accounted for about 21%, 13%, 22%, and 13 % of all fatal accidents during the year 2014, 2015, 2016, and 2017 respectively. The complexity of geological deposit and variability of mining parameters leads to the occurrence of roof fall. According to the importance of the problem, proposed various relationship between the roof fall and the effective parameters on occurrences of the roof fall in underground mines. Fuzzy logic is a useful mathematical tool for modeling the existing uncertainty and complexity. This method is successfully applied in different aspects of science and engineering to real-world problems.

In this present work, Fuzzy Logic techniques like Mamdani Fuzzy, and Adaptive Neuro-Fuzzy Inference System were used for the prediction of roof fall area in goaf in underground coal mines of bord and pillar mine method. The Fuzzy techniques like Mamdani Fuzzy and ANFIS was applied to the two mines namely Mine A and Mine B. The Fuzzification, Defuzzification, If – THEN rules were framed for the Mamdani Fuzzy model. Another Hybrid fuzzy model ANFIS model was also used for the prediction of roof fall area. In ANFIS model the field data was trained, tested and check validated for the prediction of roof fall area. The results of both models indicated that ANFIS model predicted results gave better estimation than Mamdani Fuzzy model.