Prediction of the creeping of AFC based on fuzzy reasoning and Bi-LSTM fusion iteration

Abstract

The creeping of Armoured Face Conveyor (AFC) is an engineering problem that needs to be avoided in coal mining production process. In this paper, a method for predicting the creeping accident of AFC based on fuzzy reasoning and Bi-directional Long Short-Term Memory (Bi-LSTM) fusion iteration is put forward. Firstly, through the force analysis of the AFC and the fuzzy correlation analysis method in the actual operation process, the reasons for the creeping of AFC are analyzed; Secondly, according to the propulsion characteristics of the AFC in the time sequence development, the method of the AFC running track based on Bi-LSTM neural network is proposed; Then, on the basis of the virtual transformation of the prediction results, a judgment mechanism for the extent of the creeping of the AFC based on fuzzy evidence reasoning based on fuzzy comprehensive evaluation method and Dempster-Shafer evidence theory (D-S evidence theory) is established; Finally, the analysis on the creeping of 9711 full-mechanized mining face in Kaiyuan Mine under virtual environment after 6 cycles of continuous advancement shows that the extent of creeping of AFC is relatively high and coal mining accidents are likely to occur.