مقایسه طبقه بندی کانسنگ و باطله براساس نتایج حاصل از روش‌های لاگ کریجینگ و شرطی سازی یکنواخت محلی (Luc)

درشرایطی که فاصله داری داده ها در مقایسه با ابعاد بلوک مورد تخمین بزرگ باشد استفاده از تکنیک های تخمین مبتنی بر رگرسیون خطی برای مدلسازی عیار بلوک های کوچک نامناسب است. برای غلبه بر این مشکل، از روشی جدید تحت عنوان شرطی سازی یکنواخت محلی (luc) که نتیجه توسعه روش شرطی سازی یکنواخت (uc) است می توان استفاده نمود. در این پژوهش کاربرد این روش برای تفکیک واحدهای استخراج انتخابی به باطله و ماده معدنی براساس عیار حد 2/0 درصد در معدن مس میدوک مورد بررسی و با نتایج حاصل از روش متداول تخمین مقایسه شده است. براساس چالهای انفجاری در معدن برای دو پنل استخراجی 2450 و 2465 نتایج نشان داد که از مجموع کل بلوک ها 4 درصد باطله و 96 درصد کانسنگ می باشد.  بنابراین درصد تغییرات کانسنگ جهت نتیجه گیری مد نظر قرار گرفت. برای روش لاگ کریجینگ،  در این دو پهنه 49 درصد موارد تفکیک ماده معدنی با خروجی کنترل عیار تطابق دارد در حالیکه در 51 درصد موارد ماده معدنی به صورت باطله تفکیک شده است. همچنین براساس خروجی های روش شرطی سازی یکنواخت محلی در این دو پهنه 98 درصد رده بندی ماده معدنی با خروجی کنترل عیار تطابق داشته و 2 درصد تفکیک ماده معدنی به عنوان باطله تفکیک شده است. بنابراین حتی با فرض یکسان دانستن هزینه های ناشی از خطای کم رده بندی و بیش رده بندی، روش شرطی سازی یکنواخت محلی از مطلوبیت بیشتری برخوردار است.

Comparison of ore and tailings classification based on the results of ordinary erosion methods and local uniform conditioning

Summary In this research, the application of localized uniform conditioning (LUC) method for SMU classification into waste and ore based on 0.2% cutoff grade has been compared with the conventional estimation method in the Miduk copper mine. Based on the blast holes, for the two extraction panels. Finally, even assuming the same level of cost due to underestimation and overestimation, the LUC method is more desirable than Logkriging.   Introduction In the cases of the borehole data have large grid compared with the dimensions of the blocks, using the estimation techniques based on linear regression for modeling small blocks are. To solve this problem, Abzalov in 2006 introduced a new method of LUC which was result of a series of corrections on uniform conditioning (UC). In this method, after calculating the grade distribution functions for large panels, first, based on the uniform conditioning method, the large panel is divided into small blocks according to the increase of grade; small blocks located on the panels are ranked. This method is able to present the grade distribution functions for large panels based on UC method, and also can localized results of UC models.   Methodology and Approaches In this research, we used Logkriging and LUC techniques in order to classify waste and ore materials. The dimensions of the SMU units were chosen as 5×5×15 meters. Then, using Logkriging and LUC methods the grade of each SMU blocks in the block model was estimated. Then, according to the 0.2% cutoff grade, the SMU was divided into waste and ore. The best way to compare these two methods is to compare them with the output of the ore control unit. According to the short distance between blast holes, the separation of the ore control unit was considered as real data. Based on this data obtained from the ore control unit, a block model was prepared and separation of waste and ore blocks was done. Then, the results of the separation based on the LUC, Logkriging and blast hole model output were compared.   Results and Conclusions Based on the blast holes, for the two extraction panels, 2450 and 2465, their results showed that the total blocks were separated to 4% waste and 96% of the ore; therefore, the ore percentage changes were considered for the conclusion. For the method of Logkriging, in these two panels, 49% of the separation of the ore is matched with the ore control output, while in the 51% of the cases the ore was classified as waste. Also, according to the outputs of the LUC method, in these two benches, 98% of the classification of the ore was matched to the ore control output, and 2% of the separation of the ore was classified as waste. Therefore, even assuming the same level of cost due to under estimation and over estimation, the LUC method is more desirable.