The test study area is the Miocene reservoir of Nam Con Son Basin, offshore Vietnam. In the study we used unsupervised learning to automatically cluster hydraulic flow units (HU) based on flow zone indicators (FZI) in a core plug dataset. Then we applied supervised learning to predict HU by combining core and well log data. We tested several machine learning algorithms. In the first phase, we derived hydraulic flow unit clustering of porosity and permeability of core data using unsupervised machine learning methods such as Ward’s, K mean, Self-Organize Map (SOM) and Fuzzy C mean (FCM). Then we applied supervised machine learning methods including Artificial Neural Networks (ANN), Support Vector Machines (SVM), Boosted Tree (BT) and Random Forest (RF). We combined both core and log data to predict HU logs for the full well section of the wells without core data. We used four wells with six logs (GR, DT, NPHI, LLD, LSS and RHOB) and 578 cores from the Miocene reservoir to train, validate and test the data. Our goal was to show that the correct combination of cores and well logs data would provide reservoir engineers with a tool for HU classification and estimation of permeability in a continuous geological profile. Our research showed that machine learning effectively boosts the prediction of permeability, reduces uncertainty in reservoir modeling, and improves project economics.
Geo log i cal mod els play a cru cial role in the de scrip tion and sim u la tion of fluid flow of both hy dro car bon-and wa ter-bear ing strata. Meth od ol ogy, based on the hy drau lic flow unit build on the ba sis of core plug data com bined with rock types de termined from logs and 3D seis mic cubes gen er ated on the ba sis of 2D seis mic sec tions is pre sented. It works as a pos si ble explo ra tion tool for the Mio cene gas ac cu mu la tions in the Carpathian Foredeep of Po land. De ter min is tic and sto chas tic, geostatistical meth ods were used to con struct a static res er voir model from 2D seis mic sec tions, lithological data and hydrau lic flow unit data. A pseudo-3D seis mic vol ume was gen er ated from all of the 2D seis mic data avail able, in or der to aid the mod el ling of hy drau lic flow units. This ap proach is ap pli ca ble to other res er voirs, where the avail abil ity of seis mic data is lim ited. This study dem on strates that even with out 3D seis mic data and with lim ited well log data, the pro posed hy drau lic flow unit ap proach can be suc cess fully ap plied to res er voir mod el ling through the in te gra tion of di verse data sets for a wide range of scales.Key words: hy drau lic flow units, res er voir static mod el ling, po ros ity, per me abil ity, well log ging, 2D seismics.
Bài báo giới thiệu phương pháp minh giải địa chấn toàn phần (global seismic interpretation method) được phát triển bởi Pauget và nnk. [1]. Mô hình 3D thời gian địa chất tương đối (3D relative geologic time, RGT) được xây dựng trực tiếp từ tài liệu địa chấn là kết quả của phương pháp này. Trong mô hình RGT, tuổi địa chất có sự tiếp diễn liên tục, được nội suy và xác định trên mọi điểm của tài liệu địa chấn 3D. Tài liệu sử dụng trong nghiên cứu này là khối địa chấn Maui 3D, bể trầm tích Taranaki, ngoài khơi New Zealand. Mô hình RGT với số lượng 400 mặt phản xạ được đưa ra nhanh chóng trong quá trình minh giải. Kết quả cho thấy rõ ràng và chi tiết các đặc điểm địa chất ngay cả với khu vực địa chất phức tạp mà phương pháp minh giải địa chấn truyền thống khó minh giải. Ngoài ra, việc tích hợp các thuộc tính địa chấn (như Root Mean Square - RMS, Spectral Decomposition…) cho phép minh giải chi tiết hơn về địa tầng, chính xác hóa các yếu tố về cấu trúc địa chất, đặc trưng vỉa chứa và môi trường cổ trầm tích, từ đó có thể phát hiện các bẫy chứa địa tầng.
A sed i ment bud get for the cen tral Viet nam shelf off Nha Trang over the last deglacial Ho lo cene highstand pe riod has been inves ti gated on the ba sis of shal low seis mic and sed i ment core data and em pir i cal equa tions. The an nual sus pended sed i ment dis charge to the Nha Trang shelf ranges from: 4.3 to 5.4 Mt/year. Es ti mates based on pub lished em pir i cal equa tions sug gest that the sed i ment dis charge by three main lo cal moun tain ous rivers (the Cai, Dinh and Van Phong rivers) that en ter the Nha Trang shelf ranges be tween 1.7 and 4 Mt/year, which im plies that the lo cal rivers dis charge ap prox i mately 75% of the to tal an nual sed i ment in put to the shelf. The an nual sed i ment sup ply of the Cai River is ap prox i mately 2 and 6 times higher than that of the Dinh and Van Phong rivers, re spec tively. The highstand sed i ment depocentre of the Nha Trang shelf is mostly attached to the lo cal river out flows, in di cat ing their im por tance as the prin ci pal sed i ment sup ply sources to the shelf. Ad di tional sources of sed i ment sup ply to the Nha Trang shelf can prob a bly be re lated to along-shore trans port from the nearby shelves. Cal cu la tions based on seis mic and sed i ment core data in di cate that the net sed i ment vol ume stor age on the Nha Trang shelf is ap prox i mately 2.15 Mt/year. Ap prox i mately 50% of the to tal sed i ment yield sup plied to the shelf is prob a bly trans ported along-shore to the south. The sed i ment bud get model for highstand de pos its on the Nha Trang shelf is typ i cal for a small moun tain ous river ba sin, which is sig nif i cantly dif fer ent from that of the large river delta sys tems in Viet nam such as the Mekong and Red rivers where 90% of the river sed i ments are cap tured on the delta plain/sub aque ous part and only 10% of the river sed i ments are trans ported to the nearby shelf. In con trast, most of the sed i ments sup plied by small moun tain ous rivers off Nha Trang are trans ported to the mid-shelf, form ing a shore-par al lel mud depocentre. Key words: sed i ment bud get, Nha Trang shelf, cen tral Viet nam, Ho lo cene highstand, se quence stra tig ra phy.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.