TY - CPAPER
TI - Bed Blending Design Incorporating Multiple Regression Modelling And Genetic Algorithms
AU - Kumral, M.
PB - The Southern African Institute of Mining and Metallurgy
AB - The efficiency of an ore-processing unit depends on the consistency of the characteristics of raw material entering the plant. When the mined ore is highly variable in quality, the only way to ensure consistency is to homogenize the ore prior to feeding to the processing plant. The homogenization can generally be achieved by the bed blending operation. Given that the stockpiling and reclamation processes are very expensive, it is necessary to design the process in such a way as to minimize variabilities of specified properties of raw material. In this paper, for alternative stacking types, optimal stockpile geometry is found in three stages: First, stockpile input is simulated by sequential Gaussian simulation, and then the variance reduction ratios (VRR) as a criterion of stockpile efficiency are calculated for various stockpile geometry scenarios by a stockpile simulator written in FORTRAN. Second, multiple regression analysis is performed to model the VRR by the use of stockpile length, the number of layers and stacker speed as the independent variables. Finally, the model is an optimization problem. Decision variables are the stockpile length, the number of layers, stacker speed and stacking type. The genetic algorithms (GA) are used to minimize the VRR. The approach was demonstrated on data from an iron orebody. The problem was to reduce fluctuations of iron, silica, alumina and lime contents in the stockpile output. The results showed that the approach could be used for the bed blending design efficiently. Keywords: bed blending design, multiple regression analysis, genetic algorithms, iron ore, content fluctuation
PY - 2006
ER -