S Kianpouryan; M Farahmandian; M Karimi; A Bahroudi
Abstract
Considering the existence of many copper deposits in Iran and the importance of their exploration, mineral potential mapping with high accuracy is an important tool. The process of mineral potential mapping is a cumbersome process which can be performed using different methods. The Hybrid Neuro-Fuzzy ...
Read More
Considering the existence of many copper deposits in Iran and the importance of their exploration, mineral potential mapping with high accuracy is an important tool. The process of mineral potential mapping is a cumbersome process which can be performed using different methods. The Hybrid Neuro-Fuzzy approach is one of the powerful ones for mineral potential mapping in which both conceptual and empirical components of earth science information are employed, so using both components simultaneously increase the confidence interval. In this paper we were used Adaptive Neuro-Fuzzy Inference System (ANFIS) for mineral potential mapping in Chahar-Gonbad area 1:100000 sheet, Kerman province. The database consists of geology, geochemistry, airborne radiometric, regional faults, ETM+ data, and 22 deposit and occurrence locations. At first, the factor maps were provided in GIS environment in which each cell in the grid data represents a 100 m square on the ground, and then the outputs of this layer were used for training the network. As this technique requires some data for training the network, the occurrence locations were used for training and checking points. Since, the training points were not enough for this procedure, we assigned buffer from 100 to 1000 m for occurrence locations. The results showed that when the buffer is 500 m, the best classification which ANFIS identify about 80% of the known deposits and occurrence locations in high favorability zones.
S Alaei Moghadam; M Karimi; M Mesgari; N Saheb Zamani
Abstract
Due to the extensive areas of potential mineral reserves in the country, it seems necessary to have a systematic approach to identify and convert indices of mineral deposits into mines. Existing various conceptual models of mineral deposits, variety of both quantitative and qualitative data to explore ...
Read More
Due to the extensive areas of potential mineral reserves in the country, it seems necessary to have a systematic approach to identify and convert indices of mineral deposits into mines. Existing various conceptual models of mineral deposits, variety of both quantitative and qualitative data to explore mineral deposits and the expertise and different interests, cause the mineral potential mapping process to be very complicated. So far, various methods such as the overlap index, fuzzy logic, neural networks and weights of evidence are used for modeling this complexity. Consideration the fuzzy nature of mineral exploration in the process of modeling exploratory data, applying expert knowledge and flexibility for all types of mineral deposits in the form of an integrated system is essential. Compared with other methods fuzzy inference system has stated characteristics. To verify this, in this study, a fuzzy inference system for modeling mineral potential was proposed and for the Chah Firoozeh copper deposit was implemented. The main stages of this research include fuzzifying factor maps using the appropriate membership functions and linguistic variables, combining factor maps using fuzzy inference (by creating if_then fuzzy rules database and using an appropriate decision-making model) and generating mineral potential map with defuzzification output.
In the resulted mineral potential map, porphyry copper mineralization prone area is located in the central regions with north-south extension. For evaluation, 24 exploration boreholes in the area are complying with the mineral potential map. Based on the four classification types of mineral potential map, the compliance rate was calculated as 63.64%, 75%, 63.95% and 80.23%. Obtained mineral potential map is more accurate in the very low potential areas and 81.52% of the holes with very low state are located properly. In addition, resulted mineral potential map was compared with the mineral potential map generated using only fuzzy operators and without fuzzifying factor maps. The comparison shows that the mineral potential map that was generated using fuzzy inference system, in four classifications used in this study has 6% greater compliance with the exploration boreholes in average.
G. R. Elyasi; M. Karimi; A. Bahroudi; A. Adeli Sarcheshme
Abstract
Piles of maps from different sources with varying scales and formats and different styles and absence of a proper solution for integrating vast amount of information has resulted in a complexity for preparing mineral potential map. Using GIS not only organizes the information related to mineral exploration ...
Read More
Piles of maps from different sources with varying scales and formats and different styles and absence of a proper solution for integrating vast amount of information has resulted in a complexity for preparing mineral potential map. Using GIS not only organizes the information related to mineral exploration but also has the ability to produce and integrate information layers in different models with more precision and speed and supports spatial decision makings. In this article mineral potential map of Now Chun copper prospect has been produced for determination of drilling points. Used layers in this study include rock type, structure, alteration, mineralization indicators, anomaly zone of chargeability and apparent resistivity and metal factor, anomaly of copper and molybdenum and Cu-Mo additive indexes. After information preparation, Factor maps were weighted and integrated in the inference network. Integration use of Fuzzy logic and index overlay operators in inference network can eliminate defects in other models and provide more flexible integration of factor maps. Regarding to produce mineral potential map, mineral potential zones of porphyry copper were located in north-east parts of studied area. Eventually, the degree of correlation between mineral potential map and those operated exploration boreholes have been estimated for two different classes, 63.16 % and 64.52 %. Comparison between the high potential points indicated by our mineral potential maps with those previous drilled boreholes reveals about 26% discorrelation. It means that if such present study had been done before any drilling operation, it could have saved 200,000$ just for drilling expenditure.
M. Karimi; M.J. Valadan Zoej; H. Ebadi; N. Saheb Zamani
Abstract
Considering the vast area of Iran and extent of her potential mineral reserves (existence of volcanic belt of Urumieh-Dokhtar), a systematic view for mineral deposit exploration and mineral potential mapping is essential. Lack of a systematic view and appropriate models ...
Read More
Considering the vast area of Iran and extent of her potential mineral reserves (existence of volcanic belt of Urumieh-Dokhtar), a systematic view for mineral deposit exploration and mineral potential mapping is essential. Lack of a systematic view and appropriate models for collecting, managing and integrating various geo-spatial data from different sources based on various formats make it difficult to identify, evaluate and proioritize mineral potentials.
Since most of the data related to mineral deposit exploration activities are geo-spatial, Geographical Information System (GIS) can describe and analyze interactions, make predictions with models, and provide support for decision-makers. Mineral potential mappig composes of different steps including: identifying mineralization recognition criteria, data perparation and structuring, producing factor maps and integrating factor maps in the appropriate inference networks. In this research conventional models for integrating factor maps have been investigated. Index overlay and fuzzy logic models are selected to be appropriate models for mineral deposit exploration in semi-detailed (regional study) and detailed stages. An integrated model was also proposed based on Boolean, index overlay and fuzzy logic models . For experimental test, the mineral potential map of Rigan Bam copper deposit with appropriate methods in different inference networks have been produced and 3 appropriate inference networks (one network by Fuzzy Logic model and two networks by integrated models) were selected. Results of three selected networks are in a good accordance with drilling results (%75). Proposed model in Rigan Bam copper deposit capability with required variation can be used for other mineral potential areas and site selection of drilling wells.
