Method of Copper Deposit Mineralization Discrete Modelling With Use of Crystal Ball Application Based On Copper and Silver Grades Examples

摘要

The fundamental idea of simulation is replacing unknown variables with their statis-tical distributions. Selecting distributions fitting geological data collected for example describing variation of geological parameters such as copper and silver grades, reveals to be difficult. There are examples of stochastic, random modelling both with use of continuous or discrete distributions. The reason for inexistence of a simple solution is usually unpredictable; in this case it is a compli-cated character of the phenomenon that cannot be described with known mathematical function. The paper shows the method developed by the author allowing for distribution fitting to un-known (decision) variables which are basically discrete. Researches have been done on examples of statistical modelling of copper and silver grade in "X" copper ore deposit. Discrete methods are suggested to be used in case of modelling smaller collections of observations, which is very com-mon at an initial stage of ore development. Being conscious of input data limitation, the attempt to modelling with the best fitted discrete distribution was supported by Crystal Ball and Statistica software. The source of geological assays was exploratory boreholes (1,5x1,5 km blocks). As a re-sult of estimating, probability distributions of metals grade (Cu, Ag) were obtained as a weighted sum of uniform distributions. The evaluated method seems to be simple and reliable. The precision achieved depends on den-sity and quality of geological data. The method allows also for avoiding the problem of limited da-ta while maintaining an acceptable level of agreement with the actual. Along with growth of data availability the precision of the method remains high and resultant discrete distributions resemble continuous ones. The method can be also applied to other scientific areas. The only condition is to have defined sets of data and software allowing for relative probability estimating (for instance –Crystal Ball).