Selection of the optimal plan for maintaining the roadway affected by mining operations, when mining coal seams in weak rock formations with complex conditions of occurrence, has been studied a lot over the past decades, which has brought about many successes.
The optimal design of roadway maintenance plan requires a precise understanding of the mechanical properties of the surrounding rock mass, the interaction between the rock mass and the installed support system, and the stress distribution induced around the mining face and around the tunnel. Hence, researchers have been trying to solve the roadway maintenance problem by empirical methods such as mechanical, statistical and visual methods and numerical simulation analysis.
Empirical methods, where the complex mechanical interaction between rock mass and roof support is represented by simple equations, are supported by a large database, but they generally neglect the stress distribution generated by mining, the geological relationships of roof support management systems, and the interaction between the support system and the rock mass. On the other hand, numerical simulation analysis, which is dependent on the subjective experience of humans, cannot reflect well the nonlinear relationship between the factors affecting the stability of workings, and thus, it is not reliable for stability determination.
To solve this problem, researchers have used GA and ANN or their combination to establish nonlinear relationships, predict the displacement of the roadway influenced by the roadway depth, principal stress, drilling method, and rock load, and determine the displacement of the rock mass around roadway and the geological parameters for mining. This intelligent analysis method is very effective for its objectivity when evaluating the stability of roadways in a region of relatively regular coal seams. However, when the coal seams are in an irregular state and the roadway is strongly affected by mining operation, the nonlinear relationships between the factors should be taken into account for stability determination.
Pak Tae Song, a section head at the Faculty of Mining Engineering, has found the optimal maintenance scheme of roadways strongly affected by mining operation for nonlinear problems by combining ANN with high adaptive capacity and GA with strong adaptive optimization ability.
He applied a two-stage ANN-GA. First, he obtained the installation space and failure rate of the support from the combination of ANN and GA in the second stage. Then, he found the optimal roadway maintenance scheme from the combination of ANN and GA in the first stage.
After that, he compared the selected scheme with field observations. The results show that the scheme is effective in optimizing the maintenance system of workings, which is affected by the nonlinear relationship among mining geological parameters.
The ANN-GA method can be effectively applied to the study of the geo-mechanical behavior of rock mass affected by the uncertainty relations of qualitative and quantitative factors caused by mining operations.
© 2021 Kim Chaek University of Technology