مقایسه روش‌های زمین‌آماری و هوش مصنوعی جهت مدل‌سازی سه‌بعدی کانی‌زایی طلای اپی‌ترمال منطقه زایلیک شمال غرب ایران | |
| روش های تحلیلی و عددی در مهندسی معدن | |
| مقاله 1، دوره 14، شماره 38، فروردین 1403، صفحه 1-19 اصل مقاله (2.63 M) | |
| نوع مقاله: مقاله پژوهشی | |
| شناسه دیجیتال (DOI): 10.22034/anm.2023.19842.1591 | |
| نویسندگان | |
| محمد جعفر محمد زاده* ؛ محمد مهدی رجایی | |
| دانشکده مهندسی معدن، دانشگاه صنعتی سهند، تبریز، ایران | |
| چکیده | |
| هدف از این پژوهش، مقایسه و ارزیابی مدل‌سازی‌های مختلف، جهت تشخیص بهتر الگوهای ژئوشیمیایی توزیع Au و تفکیک دقیق‌تر زون‌های کانی‌سازی طلای رگه‌ای منطقه زایلیک در شمال غرب ایران است. در این منطقه، عیار Au در رگه 01S (یکی از 7 رگه محدوده اکتشافی) با استفاده از روش زمین‌آماری کریجینگ معمولی (OK) و همچنین روش‌های هوش مصنوعی مانند تلفیق شبکه عصبی مصنوعی (ANN) با الگوریتم‌های کرم شب‌تاب (FFA) و بهینه‌سازی ازدحام ذرات (PSO)، تخمین زده شد. داده‌های حاصل به بلوک‌ها و زیر بلوک‌های مربوطه در نرم‌افزار دیتاماین وارد گردیده و مدل‌سازی‌های سه‌بعدی به‌دست‌آمده با یکدیگر مقایسه شدند. مدل‌سازی در روش‌های هوش مصنوعی، با استفاده از کد نویسی در نرم‌افزار متلب و ارتباط دادن آن با نرم‌افزار دیتاماین در چهار گام مجزا انجام شد که در این روش‌ها، با کمک FFA و PSO، پارامترهای روش ANN مانند بایاس و وزن‌ها به‌روزرسانی و بهینه گردید تا نتایج بهتری نسبت به روش ANN به دست آید. جهت اطمینان از دقت مدل‌سازی‌ها، از پارامترهای آماری ضریب تعیین (R2) و تابع خطا جذر میانگین مربعات خطا (RMSE) استفاده شد. نتایج نشان می‌دهد، روش تلفیقی الگوریتم کرم شب‌تاب (ANN-FFA)، با توجه به حداقل بودن تابع خطا (134/0RMSE= ) و حداکثر بودن ضریب تعیین (66/0R2=)، دارای بیشترین دقت است. همچنین جهت اطمینان از صحت مدل‌سازی‌ها در روش‌های تلفیقی، مقایسه‌ای با روش مرسوم زمین‌آماری OK انجام شد و صحت آن نیز مورد تأیید قرار گرفت. در تمامی مدل‌سازی‌های انجام‌گرفته، محل مقادیر تخمین زده شده انطباق مناسبی با لیتولوژی و دگرسانی‌های مرتبط با کانی‌سازی Au در این منطقه داشت. | |
| کلیدواژه‌ها | |
| کریجینگ؛ شبکه عصبی مصنوعی؛ الگوریتم کرم شب‌تاب؛ بهینه‌سازی ازدحام ذرات؛ کانی‌زایی طلا زایلیک | |
| عنوان مقاله [English] | |
| Comparison of geostatistics and artificial intelligence methods for 3D modeling of epithermal gold mineralization in the Zailik region, northwest of Iran | |
| نویسندگان [English] | |
| Mohammadjafar Mohammadzadeh؛ Mohammadmahdi Rajaei | |
| Dept. of Mining Engineering, Sahand University of Technology, Tabriz, Iran | |
| چکیده [English] | |
| Summary For the three-dimensional modeling of the S01 vein from the Zailik exploratory area, the sampled data of the trenches and boreholes of this vein were used, and the gold grade was estimated using ANN, ANN-PSO, and ANN-FFA methods. To check the accuracy of the modeling, it was compared with the estimate of grade using the ordinary geostatistical kriging method, as well as the geological evidence of the area, such as lithology and alteration. Introduction Due to the lack of essential mathematical models to describe the magma migration and subtraction, the mineralizing in rocks, and the need to identify the anomalies associated with real mineralization from its false types, it is necessary to specific modeling for each mineral mass. Modeling is indispensable because after modeling, first of all, the ore reserve is calculated by the product of the block volume and the specific weight of the rock, and the tonnage-grade diagram is drawn (determining the economic limit grade), and according to the conventional classification systems, the type of ore deposit is specified. Another advantage of modeling is saving money and time instead of carrying out excess sampling in the exploration area [1]. Estimation operations are used to perform modeling and pattern recognition; when a data analyst model encounters a set of data, it must be able to estimate these data, which requires choosing the best estimator model to solve a specific problem. In new methods, achieving optimal modeling will only be possible with the simultaneous use of geological sciences, mathematics (statistics and probabilities), and computer engineering (artificial intelligence). This research aims to combine machine learning algorithms (support vector machine) and meta-heuristic optimizer algorithms (firefly algorithm and particle swarm optimizer) to perform three-dimensional block modeling and prepare Au-grade distribution maps in the Zailik region in the northwest of Iran. In this article, to model the S01 vein gold reserve in three dimensions, the combination of trenches and borehole sample data was used to obtain the Au grade distribution map using geostatistics and artificial intelligence methods. In the above modeling, comparing and evaluating the mathematical and numerical criteria, it is necessary to examine the qualitative validation of the predicted models to identify and extract the exploration pattern correctly. Qualitative validation means measuring the conformity of the obtained exploratory modeling with the known mineral deposit datasets that were not used in the preparation of the model [2]. The general process of conducting this research is as follows: Analyzing lithogeochemical data and pre-processing on Au values and corresponding paragenesis values. Prediction and estimation of Au values using the geostatistical method as well as artificial intelligence integrated methods (artificial neural network, firefly, and particle swarm algorithms) Comparison of quantitative evaluation criteria such as coefficient of determination and error function value [3] Qualitative assessments and compliance of the estimated points with the lithology and variation of the Zailik region Methodology and Approaches In the Zailik exploratory area, due to the spread and width of the mineral inside the excavated trenches, sampling was done unsystematically with variable distances and lengths (Fig. 1). These samples were analyzed in the laboratory for the gold using the Fire Assay method and for other elements using the ICP-OES method. In addition to determining the grade in each sample, that sample's lithology information and alteration type were recorded. The statistical parameters of the samples taken from this vein are represented in Table 1. Results and discussion Modeling in ANN, ANN-PSO, and ANN-FFA methods was done according to the following steps, and its results are shown in Fig. 2. The First step was Data compositing, drawing strings, creating a wire model, and finally creating a block model in the Datamine Studio software. The second step was calculating the grade of each of the blocks within the defined search radius using the constructed composite model. The third step was entering the grade values of the second-step blocks into artificial intelligence methods to estimate the blocks with specified coordinates and unknown grade values (blocks outside the search radius in the second step). The fourth step is assigning the grade values obtained in the previous steps to the corresponding blocks in the Datamine Studio software. Conclusions In this research, recognizing the geochemical patterns of Au in the Zailik region appropriately and more accurately distinguishing the anomaly from the background, modeling using geostatistics and artificial intelligence methods (ANN, ANN-FFA, and ANN-PSO) were performed and compared with each other. In artificial intelligence modeling for validation, by quantitatively comparing the accuracy evaluation criteria, it was shown that the ANN-FFA method has the highest coefficient of determination (R2=0.66) and the lowest error value (RMSE=0.134) compared to ANN and ANN-PSO methods. The location of the estimated values was compatible with the lithology and alteration related to Au mineralization. To approve the correctness of the modeling done in artificial intelligence methods, the location of the estimated values in these methods was in good agreement with each other and the area of the estimated values in geostatistical methods. Hence, the estimated anomalies were in the northern and eastern parts of the S01 vein, with a high amount of Au grade obtained in silicate minerals and silicified alteration zones. As a result, according to the obtained layers of information, the creation of a strong silicified alteration zone with high amounts of iron oxides and the formation of scattered alteration halos around this zone (advanced argillic with medium amounts of iron oxide and medium argillic with low amounts of iron oxide), It indicates the formation of a potential and promising anomaly for Au mineralization in the region. | |
| کلیدواژه‌ها [English] | |
| kriging, artificial neural network, Firefly algorithm, particle swarm optimization, Zailik | |
|
سایر فایل های مرتبط با مقاله
|
|
| مراجع | |
|
[1] Parsa, M., Carranza, E. J. M., & Ahmadi, B. (2022). Deep GMDH Neural Networks for Predictive Mapping of Mineral Prospectivity in Terrains Hosting Few but Large Mineral Deposits. Natural Resources Research, 31(1), 37-50.
[2] Fernández-Delgado, M., Cernadas, E., Barro, S., & Amorim, D. (2014). Do we need hundreds of classifiers to solve real world classification problems? The journal of machine learning research, 15(1), 3133-3181.
[3] Dumakor-Dupey, N. K., & Arya, S. (2021). Machine Learning—A Review of Applications in Mineral Resource Estimation. Energies, 14(14), 4079.
[4] Wu, X., & Zhou, Y. (1993). Reserve estimation using neural network techniques. Computers & Geosciences, 19(4), 567-575.
[5] Jafrasteh, B., Fathianpour, N., & Suárez, A. (2018). Comparison of machine learning methods for copper ore grade estimation. Computational Geosciences, 22, 1371-1388.
[6] Farhadi, S., Afzal, P., Boveiri Konari, M., Daneshvar Saein, L., & Sadeghi, B. (2022). Combination of Machine Learning Algorithms with Concentration-Area Fractal Method for Soil Geochemical Anomaly Detection in Sediment-Hosted Irankuh Pb-Zn Deposit, Central Iran. Minerals, 12(6), 689.
[7] Afzal, P., Farhadi, S., Boveiri Konari, M., Shamseddin Meigooni, M., & Daneshvar Saein, L. (2022). Geochemical Anomaly Detection in the Irankuh District Using Hybrid Machine Learning Technique and Fractal Modeling. Geopersia, 12(1), 191-199.
[8] Chudasama, B. (2022). Fuzzy inference systems for mineral prospectivity modeling-optimized using Monte Carlo simulations. MethodsX, 9, 101629.
[9] Abbaszadeh, M., Soltani-Mohammadi, S., & Ahmed, A. N. (2022). Optimization of support vector machine parameters in modeling of Iju deposit mineralization and alteration zones using particle swarm optimization algorithm and grid search method. Computers & Geosciences, 165, 105140.
[10] Mahmoudabadi, H., Izadi, M., & Menhaj, M. B. (2009). A hybrid method for grade estimation using genetic algorithm and neural networks. Computational Geosciences, 13, 91-101.
[11] Yang, X.-S. (2010). Nature-inspired metaheuristic algorithms. Luniver press.
[12] Fister, I., Fister, I., Yang, X.-S., & Brest, J. (2013). A comprehensive review of firefly algorithms. Swarm and Evolutionary Computation, 13, 34-46.
[13] Moini, H., & Mohammad Torab, F. (2019). Ant colony algorithm as a high-performance method in resource estimation using LVA field; A case study: Choghart Iron ore deposit. Journal of Mining and Environment, 10(3), 667-678.
[14] Soltani-Mohammadi, S., Hoseinian, F. S., Abbaszadeh, M., & Khodadadzadeh, M. (2022). Grade estimation using a hybrid method of back-propagation artificial neural network and particle swarm optimization with integrated samples coordinate and local variability. Computers & Geosciences, 159, 104981.
[15] Das, S., Maity, S., Qu, B.-Y., & Suganthan, P. N. (2011). Real-parameter evolutionary multimodal optimization—A survey of the state-of-the-art. Swarm and Evolutionary Computation, 1(2), 71-88.
[16] Al-Thanoon, N. A., Qasim, O. S., & Algamal, Z. Y. (2019). A new hybrid firefly algorithm and particle swarm optimization for tuning parameter estimation in penalized support vector machine with application in chemometrics. Chemometrics and Intelligent Laboratory Systems, 184, 142-152.
[17] Chung, C.-J. F., & Fabbri, A. G. (2003). Validation of spatial prediction models for landslide hazard mapping. Natural Hazards, 30(3), 451-472.
[18] Emamgholizadeh, S., & Mohammadi, B. (2021). New hybrid nature-based algorithm to integration support vector machine for prediction of soil cation exchange capacity. Soft Computing, 25(21), 13451-13464.
[19] Nabavi, M. (1984). An introduction to the geology of Iran, Geological Survey of Iran. Tehran University Publications.
[20] Calagari, A., Siahcheshm, K., & Sohrabi, G. (2019). Geochemical study of alteration zones around Au-bearing silicic veins at Zailic, East of Ahar, East-Azarbaidjan Province. Iranian journal of crystallography and mineralogy, 27(2), 347-360.
[21] Lin, N., Chen, Y., Liu, H., & Liu, H. (2021). A Comparative Study of Machine Learning Models with Hyperparameter Optimization Algorithm for Mapping Mineral Prospectivity. Minerals, 11(2), 159.
[22] Poli, R., Kennedy, J., & Blackwell, T. (2007). Particle swarm optimization. Swarm intelligence, 1(1), 33-57.
[23] Shamshirband, S., Esmaeilbeiki, F., Zarehaghi, D., Neyshabouri, M., Samadianfard, S., Ghorbani, M. A., Mosavi, A., Nabipour, N., & Chau, K.-W. (2020). Comparative analysis of hybrid models of firefly optimization algorithm with support vector machines and multilayer perceptron for predicting soil temperature at different depths. Engineering Applications of Computational Fluid Mechanics, 14(1), 939-953.
[24] TSai, P.-W., Pan, J.-S., Liao, B.-Y., & Chu, S.-C. (2009). Enhanced artificial bee colony optimization. International Journal of Innovative Computing, Information and Control, 5(12), 5081-5092.
[25] Kennedy, J., & Eberhart, R. (1995, 27 Nov.-1 Dec. 1995). Particle swarm optimization. Proceedings of ICNN'95 - International Conference on Neural Networks,
[26] Assareh, E., Behrang, M. A., Assari, M. R., & Ghanbarzadeh, A. (2010). Application of PSO (particle swarm optimization) and GA (genetic algorithm) techniques on demand estimation of oil in Iran. Energy, 35(12), 5223-5229.
[27] Yousefi, A., & Amirshahi, B. (2015). A hybrid meta-heuristic algorithm based on ABC and Firefly algorithms. Journal of Advances in Computer Engineering and Technology, 1(4), 53-58.
[28] Khaze, S. R., Hojjatkhah, S., & Bagherinia, A. (2013). Evaluation the efficiency of artificial bee colony and the firefly algorithm in solving the continuous optimization problem. arXiv preprint arXiv:1310.7961.
|
|
آمار
تعداد مشاهده مقاله: 495
تعداد دریافت فایل اصل مقاله: 374