Ítem
Acceso Abierto
Identificación automática de facies litológicas de una secuencia sedimentaria basado en registros de pozo
| dc.contributor.advisor | Villarejo Mayor, John Jairo | |
| dc.creator | Montealegre Pallares, Tomás Andrés | |
| dc.creator.degree | Magíster en Matemáticas Aplicadas y Ciencias de la Computación | |
| dc.creator.degreetype | Full time | |
| dc.date.accessioned | 2023-09-15T21:09:02Z | |
| dc.date.available | 2023-09-15T21:09:02Z | |
| dc.date.created | 2023-08-11 | |
| dc.description | La identificación precisa de la litología es esencial en la caracterización de yacimientos, ya que impacta significativamente la calidad de los yacimientos de petróleo y gas. La convencional interpretación manual de los datos de registro de pozo requiere un volumen masivo de datos y es subjetiva al depender de la experiencia del geofísico. En los últimos años se han desarrollado métodos automáticos basados en inteligencia artificial para identificar la litología mediante el análisis de los registros de pozos. No obstante, muchos de estos enfoques utilizan valores de una sola medición y tienen dificultades para distinguir las características de respuesta de las litologías, lo que lleva a predicciones inexactas. Este estudio tiene como objetivo desarrollar un modelo de aprendizaje automático efectivo para la clasificación de facies litológicas en pozos. Se propusieron modelos de redes neuronales como CNN1D y LSTM para aprovechar la naturaleza secuencial de los registros. Además, se exploraron modelos ramificados que combinan diferentes tipos de redes neuronales, incluyendo un mecanismo de autoatención. Comparando estos modelos con los enfoques tradicionales KNN y FC basada en una única medición se encontró que el CNN1D fue más efectivo en términos de métricas de evaluación, superando las limitaciones de los enfoques basados en datos puntuales. Además, un análisis de importancia de características reveló que todos los registros de pozo son relevantes en la clasificación, destacando GR, RDEP, RMED y DTC como los más influyentes. La importancia asignada a estos registros en el modelo propuesto coincidió con la atención dada por un petrofísico experto durante su identificación manual. Los resultados obtenidos con los modelos propuestos presentan alternativas eficientes y satisfactorias para su aplicación en el campo de la industria de gas y petróleo. | |
| dc.description.abstract | The accurate identification of lithology is crucial in the characterization of reservoirs as it significantly impacts the quality of oil and gas fields. The conventional manual interpretation of well log data requires a massive volume of data and it is subjective, relying on the expertise of geophysicists. In recent years, automated methods have been developed to identify lithology by analyzing well log data based on artificial intelligence. Nevertheless, many of these approaches rely on single-measurement values and struggle to distinguish the response characteristics of different lithologies, leading to inaccurate predictions. This study aims on developing an effective machine learning model for the classification of lithological facies in wells. Neural network models were proposed, including CNN1D and LSTM, which leverage the sequential nature of the well log data. Furthermore, branched models combining different types of neural networks, including a self-attention mechanism, were explored. Comparing these models with traditional approaches such as KNN and a single-measurement-based FC, it was found that the CNN1D outperformed others in terms of evaluation metrics, surpassing the limitations of point-based approaches. Additionally, an analysis of feature importance revealed that all well log measurements were relevant in the classification process, with GR, RDEP, RMED, and DTC standing out as the most influential. The importance assigned to these measurements in the proposed model aligns with the attention given by expert petrophysicists during manual lithology identification. This convergence between automated approaches and human expertise reinforces confidence in the model's effectiveness. The achieved results with the proposed models present efficient and satisfactory alternatives for their application in the field of the oil and gas industry. | |
| dc.format.extent | 84 pp | |
| dc.format.mimetype | application/pdf | |
| dc.identifier.doi | https://doi.org/10.48713/10336_40985 | |
| dc.identifier.uri | https://repository.urosario.edu.co/handle/10336/40985 | |
| dc.language.iso | spa | |
| dc.publisher | Universidad del Rosario | spa |
| dc.publisher.department | Escuela de Ingeniería, Ciencia y Tecnología | spa |
| dc.publisher.program | Maestría en Matemáticas Aplicadas y Ciencias de la Computación | spa |
| dc.rights | Attribution-NonCommercial-ShareAlike 4.0 International | * |
| dc.rights.accesRights | info:eu-repo/semantics/openAccess | |
| dc.rights.acceso | Abierto (Texto Completo) | |
| dc.rights.uri | http://creativecommons.org/licenses/by-nc-sa/4.0/ | * |
| dc.source.bibliographicCitation | Zhu, Liping; Li, Hongqu; Zhongguo, Yang; Li, Chengyang; Ao, Yile (2018) Intelligent Logging Lithological Interpretation With Convolution Neural. En: Petrophysics. Vol. 59; pp. 799-810 Disponible en: http://dx.doi.org/10.30632/PJV59N6-2018a5. Disponible en: 10.30632/PJV59N6-2018a5. | |
| dc.source.bibliographicCitation | Busch, J M; Fortney, W G; Berry, L N (1987) Determination of Lithology From Well Logs by Statistical Analysis. En: SPE Formation Evaluation. Vol. 2; No. 4; pp. 412-418 0885-923X; Consultado en: 2022/11/6. Disponible en: https://doi.org/10.2118/14301-PA; http://dx.doi.org/10.2118/14301-PA. Disponible en: 10.2118/14301-PA. | |
| dc.source.bibliographicCitation | Delfiner, Pierre; Peyret, Olivier; Serra, Oberto (1987) Automatic Determination of Lithology From Well Logs. En: SPE Formation Evaluation. Vol. 2; No. 3; pp. 303-310 0885-923X; Consultado en: 2022/11/6. Disponible en: https://doi.org/10.2118/13290-PA; http://dx.doi.org/10.2118/13290-PA. Disponible en: 10.2118/13290-PA. | |
| dc.source.bibliographicCitation | IEA (2021) Oil 2021. Disponible en: https://iea.blob.core.windows.net/assets/1fa45234-bac5-4d89-a532-768960f99d07/Oil_2021-PDF.pdf. | |
| dc.source.bibliographicCitation | Jaikla, Chayawan; Devarakota, Pandu; Auchter, Neal; Sidahmed, Mohamed; Espejo, Irene (2019) FaciesNet: Machine Learning Applications for Facies Classification in Well. | |
| dc.source.bibliographicCitation | Honório, Bruno; Drummond, Rodrigo; Vidal, Alexandre; Sanchetta, Alexandre; Leite, Emilson (2012) Well Log Denoising and Geological Enhancement Based on Discrete Wavelet. En: Energy Exploration & Exploitation. Vol. 30; pp. 417-434 Disponible en: http://dx.doi.org/10.1260/0144-5987.30.3.417. Disponible en: 10.1260/0144-5987.30.3.417. | |
| dc.source.bibliographicCitation | Serra, Oberto; Serra, Lorenzo (2004) Well Logging: Data Acquisition and Applications. pp. 674 : Serralog; 9782951561250; | |
| dc.source.bibliographicCitation | Misra, Siddharth; Li, Hao; He, Jiabo (2019) Machine Learning for Subsurface Characterization. pp. 442 : Gulf Professional Publishing; 9780128177372; | |
| dc.source.bibliographicCitation | Glover, Paul (2001) Petrophysics MSc Course Notes. : University of Aberdeen; | |
| dc.source.bibliographicCitation | Rider, M H (1996) The Geological Interpretation of Well Logs. pp. 280 : Whittles Publishing; 9781870325363; | |
| dc.source.bibliographicCitation | Bateman, Richard M (2012) Openhole Log Analysis and Formation Evaluation. pp. 653 : Society of Petroleum Engineers; 9781613991565; | |
| dc.source.bibliographicCitation | Dubois, Martin K; Bohling, Geoffrey C; Chakrabarti, Swapan (2007) Comparison of four approaches to a rock facies classification problem. En: Computers & Geosciences. Vol. 33; No. 5; pp. 599-617 0098-3004; Consultado en: 2022/11/11. Disponible en: https://www.sciencedirect.com/science/article/pii/S0098300406001956; http://dx.doi.org/10.1016/j.cageo.2006.08.011. Disponible en: 10.1016/j.cageo.2006.08.011. | |
| dc.source.bibliographicCitation | Hingle, (1959) The use of logs in exploration problems. En: SEG 29th Annual Meeting. | |
| dc.source.bibliographicCitation | Ellis, Darwin V; Singer, Julian M (2010) Well Logging for Earth Scientists. pp. 728 Dordrecht 9789048169474; | |
| dc.source.bibliographicCitation | Hochreiter, Sepp; Schmidhuber, Jürgen (1997) Long Short-term Memory. En: Neural computation. Vol. 9; pp. 1735-1780 Disponible en: http://dx.doi.org/10.1162/neco.1997.9.8.1735. Disponible en: 10.1162/neco.1997.9.8.1735. | |
| dc.source.bibliographicCitation | Powers, David (2008) Evaluation: From Precision, Recall and F-Factor to ROC, Informedness,. En: Mach. Learn. Technol. Vol. 2; | |
| dc.source.bibliographicCitation | Directorate, The Norwegian Petroleum; Well definitions. 2022/11/15; Consultado en: 2022/11/15. Disponible en: https://www.npd.no/en/facts/wells/well-definitions/. | |
| dc.source.bibliographicCitation | Reading, Harold G (1996) Sedimentary Environments: Processes, Facies and Stratigraphy. pp. 2 : Wiley; 9780632036271; | |
| dc.source.bibliographicCitation | Puthusserypady, Sadasivan (2021) Applied Signal Processing. : Now Publishers; 9781680839791; Consultado en: 2022/11/15. Disponible en: https://library.oapen.org/handle/20.500.12657/47870; http://dx.doi.org/10.1561/9781680839791. Disponible en: 10.1561/9781680839791. | |
| dc.source.bibliographicCitation | Liu, Fei Tony; Ting, Kai Ming; Zhou, Zhi-Hua (2008) Isolation Forest. En: 2008 Eighth IEEE International Conference on Data Mining. pp. 413-422 Disponible en: http://dx.doi.org/10.1109/ICDM.2008.17. Disponible en: 10.1109/ICDM.2008.17. | |
| dc.source.bibliographicCitation | Bonaccorso, Giuseppe (2018) Machine Learning Algorithms. pp. 522 : Packt Publishing; 9781789347999; | |
| dc.source.bibliographicCitation | Fukushima, Kunihiko (1980) Neocognitron: A self-organizing neural network model for a mechanism of. En: Biological Cybernetics. Vol. 36; No. 4; pp. 193-202 1432-0770; Consultado en: 2022/11/15. Disponible en: https://doi.org/10.1007/BF00344251; http://dx.doi.org/10.1007/BF00344251. Disponible en: 10.1007/BF00344251. | |
| dc.source.bibliographicCitation | Aggarwal, Charu C (2018) Neural Networks and Deep Learning: A Textbook. pp. 512 : Springer; 9783319944630; | |
| dc.source.bibliographicCitation | Matthews, B W (1975) Comparison of the predicted and observed secondary structure of T4 phage. En: Biochimica et Biophysica Acta (BBA). Vol. 405; No. 2; pp. 442-451 0005-2795; Consultado en: 2022/11/15. Disponible en: https://www.sciencedirect.com/science/article/pii/0005279575901099; http://dx.doi.org/10.1016/0005-2795(75)90109-9. Disponible en: 10.1016/0005-2795(75)90109-9. | |
| dc.source.bibliographicCitation | Chollet, Francois; Kalinowski, Tomasz; Allaire, J J (2022) Deep Learning with R, Second Edition. pp. 566 : Simon and Schuster; 9781633439849; | |
| dc.source.bibliographicCitation | Distributed collaborative prediction: Results of the machine learning. 2022/11/15; Consultado en: 2022/11/15. Disponible en: https://library.seg.org/doi/10.1190/tle36030267.1. | |
| dc.source.bibliographicCitation | Santos, Daniel Theisges dos; Roisenberg, Mauro; Nascimento, Marivaldo dos Santos (2022) Deep Recurrent Neural Networks Approach to Sedimentary Facies. En: IEEE Geoscience and Remote Sensing Letters. Vol. 19; pp. 1-5 1558-0571; Disponible en: http://dx.doi.org/10.1109/LGRS.2021.3053383. Disponible en: 10.1109/LGRS.2021.3053383. | |
| dc.source.bibliographicCitation | Bormann, Peter; Aursand, Peder; Dilib, Fahad; Manral, Surrender; Dischington, Peter (2020) FORCE 2020 Well well log and lithofacies dataset for machine learning. Disponible en: http://dx.doi.org/10.5281/zenodo.4351156. Disponible en: 10.5281/zenodo.4351156. | |
| dc.source.bibliographicCitation | Shier, D E (2004) Well log normalization: Methods and guidelines. En: Petrophysics. Vol. 45; pp. 268-280 | |
| dc.source.bibliographicCitation | Boughorbel, Sabri; Jarray, Fethi; El-Anbari, Mohammed (2017) Optimal classifier for imbalanced data using Matthews Correlation. En: PLoS ONE. Vol. 12; No. 6; pp. e0177678 1932-6203; Consultado en: 2022/11/17. Disponible en: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5456046/; http://dx.doi.org/10.1371/journal.pone.0177678; https://www.ncbi.nlm.nih.gov/pubmed/28574989; https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5456046. Disponible en: 10.1371/journal.pone.0177678. | |
| dc.source.bibliographicCitation | Cover, T; Hart, P (1967) Nearest neighbor pattern classification. En: IEEE Transactions on Information Theory. Vol. 13; No. 1; pp. 21-27 1557-9654; Disponible en: http://dx.doi.org/10.1109/TIT.1967.1053964. Disponible en: 10.1109/TIT.1967.1053964. | |
| dc.source.bibliographicCitation | Frazier, Peter I (2018) A Tutorial on Bayesian Optimization. No. arXiv:1807.02811; arXiv; Consultado en: 2022/11/18. Disponible en: http://arxiv.org/abs/1807.02811; http://dx.doi.org/10.48550/arXiv.1807.02811. Disponible en: 10.48550/arXiv.1807.02811. | |
| dc.source.bibliographicCitation | Vaswani, Ashish; Shazeer, Noam; Parmar, Niki; Uszkoreit, Jakob; Jones, Llion; Gomez, Aidan N; Kaiser, Lukasz; Polosukhin, Illia (2017) Attention Is All You Need. No. arXiv:1706.03762; arXiv; Consultado en: 2023/6/8. Disponible en: http://arxiv.org/abs/1706.03762; http://dx.doi.org/10.48550/arXiv.1706.03762. Disponible en: 10.48550/arXiv.1706.03762. | |
| dc.source.bibliographicCitation | Lundberg, Scott M; Lee, Su-In (2017) A unified approach to interpreting model predictions. En: Proceedings of the 31st International Conference on Neural Information. pp. 4768-4777 : Curran Associates Inc.; Consultado en: 2023/6/11. | |
| dc.source.instname | instname:Universidad del Rosario | |
| dc.source.reponame | reponame:Repositorio Institucional EdocUR | spa |
| dc.subject | Registros de Pozo | |
| dc.subject | Litología | |
| dc.subject | Caracterización de yacimientos | |
| dc.subject | Eliminación de ruido | |
| dc.subject | Importancia de las características | |
| dc.subject | Aprendizaje automático | |
| dc.subject | Petrofísica | |
| dc.subject.keyword | Well Logs | |
| dc.subject.keyword | Lithology | |
| dc.subject.keyword | Reservoir Characterization | |
| dc.subject.keyword | Noise Removal | |
| dc.subject.keyword | Importance of Features | |
| dc.subject.keyword | Machine Learning | |
| dc.subject.keyword | Petrophysics | |
| dc.title | Identificación automática de facies litológicas de una secuencia sedimentaria basado en registros de pozo | |
| dc.title.TranslatedTitle | Automatic identification of lithological facies in a sedimentary sequence based on well logs | |
| dc.type | bachelorThesis | |
| dc.type.document | Trabajo de grado | |
| dc.type.spa | Trabajo de grado | |
| local.department.report | Escuela de Ciencias e Ingeniería |
