Development of Multi-Criteria Decision-making Model using Fuzzy-AHP Technique for Selection of Underground Metal Mining Method

Abstract

Underground mining method selection is one of the most crucial decision-making tasks for the mining engineers or mine planners at the planning stage. The selection of mining method depends on multiple factors, including the geometry of ore body, geo-mechanical conditions of the ore body and adjacent strata, and available technology. Underground mining method selection is multi-criteria decision-making (MCDM) problem, and thus, the mine planners face the challenges in the selection of the appropriate mining method for a typical ore deposit. The selection of appropriate mining method for ore body extractions of a typical deposit is very important in order to maintain the profitability, safety, and productivity.
The proposed study aims to achieve the four objectives viz. development of a Fuzzy-Analytical Hierarchy Process (Fuzzy-AHP) based MCDM model for selection of optimum underground metal mining method, sensitivity analysis of the Fuzzy-AHP model for examining the robustness, comparative study of the results of the Fuzzy-AHP model with the other MCDM models (TOPSIS, VIKOR, improved ELECTRE, PROMETHEE II, and WPM), and development of a Graphical User Interface (GUI)-based software based on the developed algorithm.
The first step of the Fuzzy-AHP model development is to define the influencing factors or criteria of underground metal mining method. An individual factor may be either intrinsic or extrinsic in nature. The present study considered 11-criteria as intrinsic factors (dip, shape, thickness, depth, grade distribution, RMR of ore, RMR of hanging wall, RMR of foot wall, RSS of ore, RSS of hanging wall, and RSS of foot wall) and 5-criteria as extrinsic factors (productivity, recovery, dilution, flexibility, and safety). These criteria were further classified into 56 sub-criteria for evaluating the seven alternatives or mining method (block-caving (BC), sublevel stoping (SS), sublevel caving (SC), room and pillar mining (RP), shrinkage stoping (SH), cut and fill stoping (CF), and square set stoping (SQ)). The proposed Fuzzy-AHP model was formulated ina 4-layer hierarchical structure. The first layer of the hierarchy defined the nature of variables viz. intrinsic and extrinsic factors. The second layer listed different criteria under intrinsic and extrinsic factors. The third layer shows the sub-criteria for each criterion (listed in the second layer). The last layer defines alternatives or underground metal mining methods. In the next step, the local weights of each factor/criteria/sub-criteria/alternatives (listed in the different layer of the hierarchy) were determined using Fuzzy-AHP technique. The local weights were further used in determining the global weights of each mining method. The alternative/mining method received the highest weight or score for a particular ore deposit characteristic has been given first priority or first rank. In the same way, the mining method, which received the second highest weight has been given second priority or second rank and so on. The developed model was validated using two Uranium ore deposit (UCIL Tummalapalle deposit and UCIL Turamdih deposit) located in India. It was observed that the most suitable mining method for the UCIL Tummalapalle deposit was room and pillar and that of UCIL Turamdih deposit was cut and fill stoping as per the proposed MCDM model. To mine the respective ore deposits, similar mining methods (as that of the respective model output) were adopted in both the cases.
The next objective of the study is to conduct the sensitivity analysis of the Fuzzy-AHP model. The sensitivity analysis of the proposed Fuzzy-AHP model was conducted by varying the fuzzification factor (α) and decision-making attitude (λ) of the mine planners. The model output was analysed for six fuzzification factors (α= 0, 0.2, 0.4, 0.6, 0.8 and 1) in the range of 0 to 1 and three decision-making attitudes (λ= 0, 0.5, 1) in the range of 0 to 1. The fuzzification factor, α equal to 0, indicates no uncertainty in the observed variable, whereas α equal to 1 indicates maximum uncertainty. The three values of λ indicates pessimistic (λ = 0), unbiased (λ = 0.5), and optimistic (λ =1) decision-making attitude. The decision-making model output was analysed for each combination of α and λ. The sensitivity of the ranking of seven mining methods was analysed by considering the fuzzification factor in 2-factor, 16-criteria, 56-sub-criteria, and 7-mining method. The results indicated that the ranking or priorities of seven mining methods were not altered by either changing in the fuzzification factor or changing in the decision-making attitude. Therefore, for any combination of λ and α, the rank of a particular mining method remains the same. This indicates the robustness of the model under uncertainty in the variables.
The third objective of the proposed study is to make a comparative study of the results of Fuzzy-AHP model with the other MCDM models like TOPSIS, VIKOR, improved ELECTRE, PROMETHEE II, and WPM. The comparative results of six MCDM models (Fuzzy-AHP, TOPSIS, ELECTRE, PROMETHEE II, VIKOR, and WPM) were determined for two Uranium ore deposits (Tummalapalle and Turamdih) for analysis. The results revealed that five MCDM models (Fuzzy-AHP, TOPSIS, ELECTRE, PROMETHEE II, and WPM) give room and pillar mining method as a first priority mining method for Tummalapalle ore deposit, whereas, VIKOR model gives equal preference to two mining methods (room and pillar and sublevel stoping). Similarly, the most suitable mining method obtained for Turamdih ore deposit is cut and fill by four MCDM models (Fuzzy-AHP, TOPSIS, ELECTRE, and WPM), whereas, VIKOR model gives equal priority to cut and fill and sublevel stoping methods. PROMETHEE II model results assigned sublevel stoping as rank one and cut and fill method as second best mining method for the specified deposit. The comparative study results of six MCDM models indicate that all the models provided the same mining method as the first priority for both the ore deposit except in one case (PROMETHEE II for Turamdih deposit). Furthermore, the Fuzzy-AHP model outputs were matched with the adopted mining method for both the deposits. The additional advantage of Fuzzy-AHP model is its robustness and consideration of the uncertainty and decision-making attitude in the model.
The final objective is to develop a GUI-based software using the developed algorithms for the selection of underground metal mining method. The software also has the option of making the sensitivity analysis of Fuzzy-AHP model. The software tool can be easily implemented for the selection of underground metal mining method for a typical ore deposit without an in-depth analysis of the model by the user.