Computer modelling limitations and new directions-Part 2

"Part 1 (CIM Bulletin, August 1991) of this article identified the limitations of the traditional modelling methods and demonstrated the approximations and errors introduced by their application to complex mining situations. The objective of Part 2 is to describe the evolution of new, proven technology which eliminates these deficiencies and provides precise and practical alternatives. These promote the use of realistic geological and mining shapes and eliminate altogether the rectangular approximations of traditional methods. Limitations of the Traditional MethodsThe limitations of the traditional modelling method s (viz. simple block modelling, gridded modelling and their more recent variations) were discussed in detail in Part I . These methods are all based on the use of rectangular discretization to represent the boundaries and volumes of ore deposits and mining excavations. As a result they are unsuited to the representation of nonrectangular geological complexities such as folding, faulting and irregular surfaces . They are equally unsuited to the representation of realistic excavation limits, particularly in the case of underground mining.Recently these methods, with limited variations in technique, have been applied to increasingly more complex geological and mining situations. The resulting potentials for error and conflict within the modelling pro cess were demon strated in Part I . The effects can be summarized as follows:• volume and boundary approximations introduced by therepresentation of irregular three-dimensional boundaries andvolumes by rectangular modelling elements ; • geostatistical errors and conflicts caused by neglect of the relationships between grade estimates and the geometric/ spatial aspects of irregular volumes and associated observations (samples), including an inability to produce precise estimates with meaningful confidence limits;• a compounding of these effects by using rectangular element sizes other than the geostatistical optimum, and by using differentelement sizes within a model; and• general inefficiencies introduced throughout the modelling pro cess by the requirement to represent geology, grade distribution and excavation limits with a singular modelling method and model."