مرکز منطقه ای اطلاع رساني علوم و فناوري - تحليل حساسيت عوامل مؤثر بر مدل سازي كانسار مس سونگون

چكيده لاتين :

The successfulness of a mining project requires a precise definition of the subsurface deposit situation. Deposit characterization is potentially a critical part of any mine design process including reserve estimation and production planning. Over the past two decades, several sophisticated modeling techniques have been developed in the mining and mineral industries. In the available commercial systems, various methodologies have been implemented. The process of deposit characterization and modeling consist of determination of the spatial variation of certain geological and mining parameters such as deposit extensions, ore grade and density. This process includes several steps; construction of database, generation of topography, modeling the geometry of the deposit, grade analysis, blocks modeling and reserve estimation. The modeling and visualization process and software application are technically and practically a challenging issue. In this regard, two main groups of challenges can be realized. The first is related to the algorithms used in the modeling and the second is concerned to the user engineering judgments in evaluating of geometry and limits of the deposit. These challenges are related to complexity of ore deposit geometry, unknown spatial limit of ore body, non sufficient number of samples and using of interpolation techniques. In this research, variations of the different parameters of second group that affect the results of modeling are examined. Certain parameters such as those related to block model and wireframe construction are specific to the modeling method and software implementation. Some of the other parameters for example search distance and minimum numbers of samples for estimation are common in the majority of modeling techniques. In the first step, a base model of the Songun copper deposit is constructed using Datamine software and then variations of the effective modeling parameters are studied, i.e 22 parameters which can be grouped in 5 categories: search samples, variogram, grade estimation, geometry and general data. Thereafter applying sensitivity analysis, the parameters were classified according to the degree of influence in the modeling results (ore tonnage, metal content and mean grade). On the basis of obtained results it was found that the variables including search distance and minimum number of samples for estimation are the most critical variables. On the other hand, the number of maximum sample and the random variance parameters were determined as the least effective in this regard. This research provides possibility of differentiating sensitive parameters and perception of behavior of various parameters in modeling process.