A decision analysis model to optimise the feed strategy of a coal preparation plant
Thesis (MBA)--Stellenbosch University, 2011.
The objective of this study is to recommend the optimal coal preparation plant feed strategy for the Tweefontein Optimisation Project (TOP) at Xstrata Coal South Africa’s Tweefontein Complex operations. In this research report the specific challenges relating to the mining, stockpiling and quality of coal relating to this problem are presented together with the decision analysis modelling used to solve the optimal plant feed strategy. Interviews were held and relevant literature studies were consulted to gain a deeper understanding of the challenges relating to the opencast mining of coal pillars remaining from historically mined underground operations, together with the coal quality variability and complexity relating to the Tweefontein Complex. Further literature studies were conducted to gain an understanding of the most favourable decision modelling tools to be applied. Three different types of models were developed. The first was to test various washing philosophies influenced by the combination of coal from various sources and seams over individual week periods. The second was to simulate the coal process flow from arrival at the run-of-mine (ROM) pad through to the loading of trains, influenced by production profiles, system constraints and dynamic stochastic processes. The final model had to test the combined financial impact of washing and stockpile philosophies on the net present value of the Tweefontein Complex over the life of the mine. Due to the nature of the modelling techniques applied, the establishment of scenarios was necessitated. Six washing philosophy scenarios and three base stockpile philosophy scenarios (developed to 39 scenarios when sensitivity and limit tests are included) were tested. Thereafter the relevant combined permutations were used to establish 12 feed strategy scenarios that were compared for their financial impact. From the individual philosophies recommended from a washing and stockpiling perspective, a final three scenarios were recommended from a net present value perspective. These include: (a) the individual feeding of five different quality-groups of coal with a ROM stockpile capacity of 128 000 tonnes and excluding a plant buffer stockpile, applying a wash approach based on the “randomly” estimated cut point density of the respective quality-group; (b) the individual feeding of two quality-groups of coal (export and Eskom) with a ROM stockpile capacity of 50 000 tonnes and a plant buffer stockpile (using a bridge reclaimer system) with a capacity of 80 000 tonnes, applying a wash approach whereby the optimal cut point density of the previous week is utilised the following week; (c) a similar feed strategy as for b, but with the application of coal valves as the reclaimer system on the plant buffer stockpile. A final recommendation was made to implement the scenario given in c above. This was concluded following a combination of additional considerations. This research report contains the knowledge gained in the personal interviews and the literature reviewed. It further discusses the methodology followed to achieve the objectives of the study, it illustrates the models and scenarios that were developed and it discusses the analysis of the results and conclusions formulated. It finally concludes with a recommendation for the optimal feed strategy for the TOP, together with additional work proposed to be conducted subsequent to this study.