3932

3710

كارشناسي ارشد

اكتشاف معدن

اصفهان: دانشگاه صنعتي اصفهان، دانشكده معدن

1386

سيزده، 65، [II] ص.: مصور، جدول، نمودار

ص. ع. به فارسي و انگليسي

نادر فتحيان پور، حسن طباطبائي

هوشنگ اسدي هاروني

07/03/87

سعيد صدري، لهراسب فرامرزي

1396/03/03

كتابنامه

معدن

مهندسي معدن

ID3710

به فارسي و انگليسي: قابل رؤيت در نسخه ديجيتال

Abstract Due to presence of inherent uncertainties in geoscience data caused by various unknown andeven known geological phenomenon applying simple boolean logics to infer from such datawould eventually lead to significant estimation errors One way out of this difficulty is toemploy knowledge based methods such as Fuzzy logic inference models which handles suchuncertainties through considering gradual nature of properties of qualitative parameters underinvestigation Fuzzy logic models are considered as knowledge based techniques and whenthey exploit the advantages of data driven techniques such as neural network form a verypowerful class of inference model called neurofuzzy Such integration of Fuzzy inferencepower of knowledge and robustness of neural network models is particularly useful in dealingwith estimating and forecasting in geoscience problems where we are faced with huge amountof data and genetic models The studied area in current research lies between Kashan to Naienin Uromieh Dkhtar volcanic belt structural zone This area has been selected for its wellknown metallic ore potential such as iron copper lead and zinc The mineral occurrencepotential of these metals have been evaluated using geological attributes such as host rocklithology tectonic settings and alterations derived from geological maps and remotely sensedimagery plus airborne magnetic data and other spectrally distinguishable features such asferrous iron and hydroxide bearing alterations in conjunction with spectral supervisedclassification maps using maximum likelihood algorithm The geophysical raw data wereprocessed through reducing them to the pole and applying analytic signal filter on them inorder to detect and enhance the spatial location of mineralization phenomenon and geologicalfeatures and contacts Through processing various geological geophysical and remotelysensed data seven thematic layers were created and allocated to three main alteration geological and structural classes The alteration class includes ionic dominance maps derivedfrom satellite imagery The geological class was formed through combining classified rocktypes of geological maps and remotely sensed classification map In the third class thestructural features derived from geological maps and satellite imagery were combined withanalytic signal map Finally the above thematic data were integrated using both neural network and neurofuzzyalgorithms to predict the favorable metallic occurrences Four training sites were adopted forboth techniques and based on the obtained predicting models the entire were processed forfavorability The final results for neurofuzzy prediction show that 95 2 percent of the knowncopper deposits were granted as favorable locations compared to the 76 1 percent for that ofneural network prediction This figure for known iron deposits was reduced to 72 2 forneurofuzzy method compared to that of 54 5 for neural network while for lead and zincdeposits both methods gave the same results covering 62 5 percent of known deposits

نادر فتحيان پور، حسن طباطبائي

هوشنگ اسدي هاروني

سعيد صدري، لهراسب فرامرزي