12778

11693

كارشناسي ارشد

اكتشاف معدن

اصفهان: دانشگاه صنعتي اصفهان، دانشكده معدن

1396

سيزده، 83ص.: مصور، جدول، عكس(رنگي)، نمودار

ص.ع. به فارسي و انگليسي

نادر فتحيان پور، هوشنگ اسدي هاروني

رسول اميرفتاحي، احمدرضا مختاري

1396/06/21

كتابنامه

معدن

مهندسي معدن

ID11693

مطالب فهرس صفحه عنوان چكيده 9 فصل اول مقدمه و اهداف تحقيم 9 9 مقدمه 9 9 9 9 ضعف زمينآمار دونقطهاي 9 9 9 9 برتري روشهاي زمينآماري چند نقطهاي 3 9 9 ضرورت موضوع تحقيق 5 9 3 اهداف تحقيق 5 9 4 ساختار پايان نامه 5 فصل دوم سابقه موعو تحقيم 9 9 سابقه به كار گيري روشهاي مبتني بر واريوگرام 7 9 9 سابقه به كار گيري روشهاي مبتني بر شئ 8 9 3 سابقه به كار گيري روشهاي مبتني بر فرآيند 2 9 4 سابقه به كار گيري روشهاي مبتني بر آمار چند نقطهاي 19 9 5 سابقه به كار گيري روشهاي شبيهسازي زمين آماري چند نقطهاي مبتني بر تبديل موجك 99 فصل سوم روش شناسي تحقيم 3 9 مقدمه 49 3 9 مفاهيم اساسي در شبيهسازي زمين آماري 59 3 9 9 گره 59 3 9 9 شبكه 59 3 9 3 تصوير آموزشي 59 3 9 4 الگوي جستجو 89 3 3 الگوريتمهاي شبيهسازي زمينآماري 89 3 3 9 الگوريتم 98 FILTERSIM 3 3 9 شبيهسازي تصادفي الگوها 99 3 4 مقدمهاي درباره موجك 99 3 4 9 تبديل فوريه 39 3 4 9 تبديل فوريه زمان كوتاه 49 3 4 3 تبديل موجك 69 3 4 4 تبديل موجك گسسته 69 3 4 5 انرژي موجك 89 3 4 6 موجكهاي زوج متعامد 29 3 4 7 تبديالت موجك دوبعدي 13 3 5 كاربرد تبديل موجك در پردازش تصوير 93 3 5 9 تجزيه موجك تصوير آموزشي 63 فصل چهارم آشنايي با منطقه مورد مطالعه پردازش دادهها و شبيهسازي 4 9 مشخصات كانسار مس سرچشمه 73 هشت

Application of Wavelet Transform in Improving Copper Grade Estimations in Sarcheshmeh Copper Mine using Filtersim method Asal Zavarei a zavarei@mi iut ac ir Date of Submission 2017 06 19 Department of Mining Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language FarsiSupervisor Nader Fathianpour and Hooshang Asadi HarooniAbstractThe increasing need of human for natural and mineral resources has attracted significantattentions towards developing new and more efficient simulation algorithms in order toenable mapping complicated geological settings Among algorithms commonly used forsimulating complex and non linear variability the multiple point geostatistical methods aregradually gaining more credit and applications The aim of multiple point geostatistics isproducing more reliable and geologically consistent model of natural phenomena throughemploying training images In fact training images represent spatial patterns and structuresthat would be expected to exist throughout ore deposit The main objective of current studyis to produce simulated realizations preserving large variability structures of third orderhidden in the measured exploratory data Since the porphyry copper deposits are inherentlycontrolled by complex and non linear phenomena the central part of Sarcheshmeh copperdeposit encompassing a number of 135 exploratory bore holes was considered as a goodexample of applying the proposed new simulation methodology The well known Filtersimalgorithm was selected as the core engine for simulating such non linear structures to beapplied on four purposely selected plans In order to define training images indicative of large scale varions throughout 2D plans atrend surface analysis of up to 3rd degree was carried out for each of the plans separately Asecond stationary training image was then produced through repeating the main 3rd ordertrend in all directions In order to maintain the governing large scale variations and in themeantime reducing the local high wavenumber noises in the course of simulation a wavelettransform was applied on both of the training image and gridded data constructed throughnearest neighbor interpolation In this way the simulated realizations created in waveletdomain via Filtersim algorithm are enhanced In order to keep the image dimensions andinertia unchanged discrete stationary wavelet transform was adopted in this study Theentire data was divided into training and testing data sets Then a wavelet transform wasapplied on both of the training images and gridded hard data up to one level and theconditional and unconditional simulations were performed for 100 realizations Next an e type estimation were evaluated on simulated realizations after applying inverse discretewavelet transform was made Finally through using a number of error measures such asMSE RMSE and relative error the performance of the conditional and unconditionalsimulations were evaluated The average relative error of computed for all plans was foundto be in the order of 10 percent which is considered acceptable in the final stage of detailedexploration program It was also concluded that the application of stationary discrete wavelettransform on the stationary training image has considerably reduced the undesirable edgeeffect created by the performance of filters in the wavenumber domain Keywords Multiple Point Geostatistical Method Training Images Filtersim SimulationAlgorithm Wavelet Transform Sarcheshmeh Copper Deposit

نادر فتحيان پور، هوشنگ اسدي هاروني

رسول اميرفتاحي، احمدرضا مختاري