GOCAD(外文资料) (31)
GOCAD(外文资料) (29)
TU Bergakademie FreibergThe TU Bergakademie Freiberg’s central concept is to endeavour in research areas that respond to the demands of industry. With its four core themes — Geosciences, Materials, Energy and Environment — it has a very distinct profile addressing the specific issues of our modern industrial world. http://tu-freiberg.de/ProMineProMine is an EU FP7 research project working to integrate the mapping of metal and mineral resources across the Union, to produce the first pan-European GIS-based database. On from this, new sustainable extraction and processing technologies are being developed along with a range of new nano products manufactured from European sourced raw materials. The ProMine project seeks to employ an inclusive and consultative approach, with an aim to improve the image of mining whilst reinforcing industrial links. Particular attention is being paid to long term sustainability and environmental concerns, reflecting current directives from the Raw Materials Initiative. The project commenced on May 1st 2009, and is coordinated by the Geological Survey of Finland (GTK), and made up of 27 partners in 11 countries working closely together over a four year period.Nancy-Université - INPLInstitut National Polytechnique de Lorraine (INPL) is a public University of Technology and a member of the ‘Nancy Université’ federation. INPL has 28 research laboratories dedicated to teaching and to research, integrating the latest technological innovations. http://english.inpl-nancy.fr/gOcad Training Course: Numerical modelingand database buildingfor 3D geological modeling28th Sept - 2nd Oct, 2009 Freiberg, GermanygOcadgOcad is a world leading enterprise of geomodeling featuring the common Earth model aiming at a unique approach to geometry, topology and properties. The gOcad software was developed by the gOcad Consortium and is now commercialized by Paradigm. The gOcad consortium consists nowadays of a multidisciplinary team of researchers striving to define advanced new approaches to build and update 3D subsurface models. /www/http://promine.gtk.fi3d modeling of a bedding and a fault planeCOURSE PROGRAMME SCOPETo provide practitioners form the exploration and mining industry theoretical concepts and training requested for building realistic 3D geological models, and resources assessment tools using gOcad. This five day course will alternate theoretical lectures and practical training exercises ranging from basic concepts (such as, importing data from external database, manipulating borehole properties, building geological contacts and faults, regular and irregular grids, solid models, visualization), to more advanced geostatistical tools (variography, kriging, simulation, resource assessment). A real case study will be used to conduct trainingexercises. Lecturers: Jean-Jacques Royer (Centre National de la Recherche Scientifique, CNRS), Regina van den Boogaart and Peggy Melzer (Geoscience Mathematics and Informatics Group), and Helmut Schaeben (Department of Geosciences, TU Bergakademie Freiberg). (L) = Lecture; (E) = Practical TrainingExercise Sep 28, 2009 Registration & Welcome (L) The gOcad – Common Earth Model (L) Using gOcad – (DSI), GMaps (E) gOcad’s user interface, major tools, visualisation, scripts, workflows, wizards etc. Sep 29, 2009 (E) Object oriented geomodeling (E) Importing data into gOcad from external database; georeferencing (E) Processing data in gOcad (E) Generating a structural model: - Pickingcontacts, faults from seismic cubes Sep 30, 2009 (E) Generating a structural model: - Buildingsurfaces and faults - Solid model, Sgrid, Voxet - Manipulatingproperties - Visualization tools Oct 1, 2009 (L) Introduction to geostatistics (L) Kriging and simulation (E) Geostatistics with gOcad – statistical tools (E) Geostatistics with gOcad – variography Oct 2, 2009 (L) Geostatistical case study: Mining case study (E) Geostatistics with gOcad – estimating ore in place Closing DiscussionVENUE & TRA VEL DIRECTIONSThe course will be held at TU Bergakademie, Department of Geology, Bernhard-von-CottaStr.2, AvHumboldt – Bau, PC-Pool second floor, D - 09599 Freiberg, Germany. By plane to Dresden Airport, take urban railway to Dresden Main Station, and then train to Freiberg. By train via Dresden Main Station to Freiberg. By car via highway A4, exit Siebenlehn (No. 75), continue via B101 to Freiberg. The site map of Bergakademie can be found at: http://campusgis.geo.tu-freiberg.de/webgis/frontend/ACCOMMODATIONUntil Sep 14, 2009, a limited number of rooms have been reserved at special rates at: Hotel Blaue Blume (www.blaue-blume.de/) Hotel Auberge Mistral (/hotel-mistral/mistral/index.htm) Reservations should be made directly by the participant (quote reference: 'Schaeben') For further assistance do not hesitate to contact Helmut Schaeben on schaeben@ geo.tu-freiberg.de or Regina.Boogaart@ geo.tu-freiberg.deCOURSE INFORMATIONThe course will be conducted in English. Due to PC pool capacity, the number of participants is limited to a maximum of 15. Participants will learn to use gOcad to build structural geological models and to apply geostatistical tools within the properly modelled geometry and topology.COURSE FEES & REGISTRATIONThe course is offered free to ProMine partners. To register, please send an email before Sep 4, 2009 to: schaeben@geo.tu-freiberg.de with copy to Regina.Boogaart@geo.tu-freiberg.dePictures courtesy of the gOcad Consortium Members。
GOCAD软件总体介绍
1、GOCAD软件总体介绍GOCAD（Geological Object Computer Aided Design）软件是一款功能强大的三维地质建模软件，在地质工程、地球物理勘探、矿业开发、石油工程、水利工程中有广泛的应用。
GOCAD软件的界面GOCAD软件具有强大的三维建模、可视化、地质解译和分析的功能。
它既可以进行表面建模，又可以进行实体建模；既可以设计空间几何对象，也可以表现空间属性分布。
并且，该软件的空间分析功能强大，信息表现方式灵活多样。
2、GOCAD联合体GOCAD研究联合体成立于1989年，该组织由致力于发展地质建模科学的高校和企业组成。
联合体的目标是开发出一整套适用于石油、气藏、矿山和环境工程领域的地质建模方案。
如今，GOCAD研究联合体已经形成旗下具有22家公司和87家高校的规模，这些都是油藏和气藏领域内将GOCAD作为勘探和生产主导产品性质的单位。
GOCAD技术研究联合体经过十多年的共同攻关，于1997年正式推出了采用独特专利技术的勘探开发一体化三维综合地质建模及虚拟现实技术软件---GOCAD！3、GOCAD软件综合建模技术特色要建模的地质目标，千姿百态，既要描述其几何形态，也要描述其所包含的地质属性特征。
但是无论多么复杂的地质体，归纳起来都可用点、线、面、体等四种类型的数据来描述。
基于这种观点，GOCAD中描述地质目标的数据定义有：•点集：描述离散数据；•线集：描述断层线、钻孔轨迹、测井曲线和河道等线状数据；•面集：描述层面、断面等面状数据；•体集：地震数据、遥感数据、地层网格、盐丘、封闭体等数据体。
3.1 GOCAD软件的对象GOCAD中的对象包括PointSet、Curve、Surface、Solid、Voxet、SGrid、Well、Group、Channel、2D-Grid、X-Section、Frame、Model3d等类型。
3.2 GOCAD软件中对象的属性GOCAD中不同类型对象包含的属性不同。
GOCAD(外文资料) (35)
Landesamt für Umwelt und Geologie
S. Görne, O. Krentz, F. Horna, C. Block, M. Hübschmann
Abteilung Geologie
Gliederung
• Einleitung • Stand der Arbeiten innerhalb der Konzeption zur Einführung
S. Görne, O. Krentz, F. Horna, C. Block, M. Hübschmann
Abteilung Geologie
Gocad-Wizard zur ModellKombination
→ Kombination benachbarter Modelle zu einem
•
geologisch-tektonische Daten sind räumliche Informationen 3D = Abbild der realen Welt Visualisierung komplexer Zusammenhänge Datengrundlage für angewandte geowissenschaftliche Disziplinen Modellierungsergebnisse für: - Projektierung von Bohransatzpunkten - Bergbaufolgen-Management - Veranschaulichung geologischer Sachverhalte für Dritte
Abteilung Geologie
57. Berg- und Hüttenmännischer Tag Landesamt für Umwelt und Geologie
GOCAD 软件三维地质建模方法
GOCAD 软件三维地质建模方法1建模方法GOCAD 三维地质建模主要包括两类：一类是构造模型（structural modeling）建模，一类是三维储层栅格结构（3D Reservoir Grid Construction）建模。
（1）构造模型（structural modeling）建模建立地质体构造模型具有非常重要的意义。
通过建立构造模型能够模拟地层面、断层面的形态、位置和相互关系；结合反映地质体的各种属性模型的可视化图形，还能够用于辅助设计钻井轨迹。
此外，构造模型还是地震勘探过程中地震反演的重要手段。
（2）三维储层栅格结构（3D Reservoir Grid Construction）建模根据建立的构造模型，在3D Reservoir Grid Construction 中可以建立其体模型；同时地质体含有多种反映岩层岩性、资源分布等特性的参数，如岩层的孔隙度、渗透率等，可对这些物性参数进行计算和综合分析，得到地质体的物性参数模型。
当采样值在地质体内密集、规则分布时，可以直接建立采样值到应用模型的映射关系，把对采样值的处理转化为对物性参数的处理，这样可以充分利用计算机的存储量大、计算速度快的特点。
当采样值呈散乱分布，并且数据量有限时，需要采用数学插值方法，拟合出连续的数据分布，充分利用由采样值所隐含的数据场的内部联系，精确的模拟模型中属性场的分布。
图1-1孔隙度参数模型分布图2 建模流程2.1数据分析（1）钻孔、测井分布及数据分析支持三维建模的数据主要为钻孔和测井。
由于对区域范围和建立三维地质建模的精度要求不同，得对所得到的钻孔、测井的分布和根据其取得的数据进行分析和处理是的必要。
根据钻孔、测井的分布范围和稠密程度可以大致确定地层的分布界限，对钻孔较少区域采取补充钻探或者采用其它方法进行处理。
图2-1由二维地质剖面图形成的三维连井剖面图（2）地质剖面对于建立三维地质模型，只根据钻孔和测井是不够的，在长期的地质勘探中形成的地质剖面图，对建立三维地质模型具有重要的作用。
GOCAD 操作手册
GOCAD综合地质与储层建模软件简易操作手册美国PST油藏技术公司PetroSolution Tech,Inc.目录第一节 GOCAD综合地质与储层建模软件简介┉┉┉┉┉┉┉┉┉┉┉┉┉┉1一、GOCAD特点┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉1二、GOCAD主要模块┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉1 第二节 GOCAD安装、启动操作┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉2一、GOCAD的安装┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉2二、GOCAD的启动┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉3 第三节 GOCAD数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉5一、井数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉5二、层数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉11三、断层数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉11四、层面、断层面加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉12五、地震数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉12 第四节 GOCAD构造建模┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉13一、准备工作┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉13二、构造建模操作流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉14三、构造建模流程总结┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉40 第五节建立GOCAD三维地质模型网格┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉41一、新建三维地质模型网格流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉41二、三维地质模型网格流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉41三、三维地质模型网格流程总结┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉47 第六节 GOCAD储层属性建模┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉48一、建立属性建模新流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉48二、属性建模操作流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉48三、属性建模后期处理┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉66四、网格粗化┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉74 第七节 GOCAD地质解释和分析┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉78GOCAD综合地质与储层建模软件操作手册第一节GOCAD综合地质与储层建模软件简介Gocad是国际上公认的主流建模软件，在众多油公司和服务公司得到了广泛的应用。
GOCAD功能简介及应用实例
GOCAD软件简介Gocad地质建模软件是国际上公认的主流建模软件，在众多油公司和服务公司得到了广泛的应用，包括Exxon、Mobile，Chevron、Texaco、BP、Halliburton，Schlumberger等世界著名的油公司和服务公司。
Gocad是以工作流程为核心的新一代地质建模软件，达到了半智能化建模的世界最高水平，具有功能强，界面友好，易学易用，并能在几乎所有硬件平台上（Sun, SGI, PC-Linux, PC-Windows）运行的特点。
Gocad研究思想是1988年法国Nancy大学的J.L. Mallet教授提出的，目的是要开发一种新的地质建模软件，以适应地质、地球物理和油藏工程的需要，为多学科的综合研究提供技术支撑。
在Gocad软件研发中除采用J.L. Mallet 教授提出的离散光滑插值技术（DSI）, 还采用了适应能力很强的三角剖分和四面体剖分技术，并独立地开发了软件中的地质统计学部分。
自1990年软件诞生后，得到了国外的许多油公司和地球物理公司的支持，取得了飞速的发展。
从最初的简单构造建模，发展到今天复杂构造建模、复杂三维模型网格生成、储层岩石物理属性模型、岩相模型等，以Gocad为代表的先进地质建模软件大大提高了地质建模的效率和精度，可以满足对复杂地质区域的建模要求。
Gocad建模思想是建立在工作流程之上的，是以地质建模的内在规律和程序为基本框架，为地质师和油藏工程师提供充分的发挥想象力的空间，使人的地质思想得以准确的融合到地质建模过程中，使整个建模过程始终以地质为本。
Gocad构造建模能处理任何复杂的构造模型，并能方便地对三维构造模型和三地质网格模型进行编辑及更新，使用Gocad软件用户能方便地对油藏进行动态跟踪。
Gocad基于流程的属性体建模方法包含几十种地质统计算法，功能强大，使用灵活，并能方便地引入约束条件，将地质家和油藏工程师的认识、大量油藏信息引入到属性体建模中，此外 Gocad建模的许多模块是EDS公司与Chevron、Texaco等油公司联合研发，因此具有很高的实用性。
GOCAD(外文资料) (30)
3D TGI Williston Basin – Gocad/Geocando projectNotes and IssuesIMPORTANT:In order for the project paths to function correctly (relative paths have not yet been implemented in Geocando) the zip file must be extracted to C:\The TGI Williston Basin 3D Geological model is based on a database of unit tops picked by geologists from selected drillholes penetrating the majority of the post-Precambrian stratigraphy and mapped unit edges derived from legacy mapping and drillhole intercepts. On average, five wells per township were selected from areas with sufficient drillhole density. The 3D surfaces were constructed using these picked tops from a total of 9012 wells, which includes 5046 wells from Saskatchewan, 2606 wells from Manitoba, and in order to reduce edge-effects, 771 wells from North Dakota and 589 from Montana. Of the 60 geological units in the TGI Williston Basin project, 42 were selected and modeled. The model was produced using Gocad earth modeling software.This model is entirely data driven with minimal human interpretation. Many fringe areas have a low data density, especially those areas close to unit edges. Because of this shortcoming, the expression of the unit edge (ie. escarpments) isn’t always accurately predicted.The TGI team was dedicated to maintaining accuracy and consistency in the picks and modeled information presented, however due to the size of the project, some errors may have been made. The TGI team is not liable for these errors. The user of this information accepts all responsibility for any work done on their part that uses all or part of this data. The data files included in this zip file have been exported from Gocad modeling software. These files can be opened using Gocad modeling software or the free Gocad viewer, Geocando. For ease of use, a Geocando project file has been included(Gocad_TGI_Model_Geocando.gtp). To use this file, open Geocando and navigate to File Æ Open Project…* Please see the ‘important issues’ section below for further information.This Geocando project includes:-42 modeled surfaces from Precambrian to rock surface-SRTM DEM ground surface-Original data pointsets used as the basis for model creationo Edge and tie points are not included-Cultural data in the form of borders and roadsGEOCANDO SOFTWARE FEATURESVertical ExaggerationWhen a project is first loaded into Geocando, the vertical exaggeration is set to 1x. In order to modify the vertical exaggeration, navigate to:1) View Æ Z-Scaling…2) Change the value to 75x for best resultsQuery toolSince the TGI Williston Basin 3D model is in real-world coordinates, you can query the model to extract coordinates from a point, and you can measure distances.Get Coordinates1) Info Æ Get Coordinates2) Click anywhere on the model. Coordinates will be displayed in the bottom left of the screenDistance tool1) Info Æ Measure Distances2) Click anywhere on the model to set the first measurement point. Click again in another location to set the second measurement point.* These tools can also be accessed using the shortcut buttons on the toolbar at the top of the screen.IMPORTANT ISSUESIn order for the project paths to function correctly (relative paths have not yet been implemented in Geocando) the zip file must be extracted to C:\If you wish to extract the file to another drive/directory, the Geocando project file (.gtp) must be modified to reflect the new path. Each surface has an associated path. The current path is: C:/3D_TGI_Williston_Basin.This issue should disappear as soon as Geocando supports relative paths.NOTE1)Many surfaces in the model are very close (vertically) to one another. Geocandorenders surfaces differently than Gocad, as a result of this difference, Geocandomay show the surfaces as interleaving. If you open the surfaces in Gocad you will not see this issue.2)Surfaces appear ‘rough’ in Geocando. Gocad’s interpolation engine renders thesurfaces with a much smoother appearance.。
GOCAD 操作手册
GOCAD综合地质与储层建模软件简易操作手册美国PST油藏技术公司PetroSolution Tech,Inc.目录第一节 GOCAD综合地质与储层建模软件简介┉┉┉┉┉┉┉┉┉┉┉┉┉┉1一、GOCAD特点┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉1二、GOCAD主要模块┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉1 第二节 GOCAD安装、启动操作┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉2一、GOCAD的安装┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉2二、GOCAD的启动┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉3 第三节 GOCAD数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉5一、井数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉5二、层数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉11三、断层数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉11四、层面、断层面加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉12五、地震数据加载┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉12 第四节 GOCAD构造建模┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉13一、准备工作┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉13二、构造建模操作流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉14三、构造建模流程总结┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉40 第五节建立GOCAD三维地质模型网格┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉41一、新建三维地质模型网格流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉41二、三维地质模型网格流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉41三、三维地质模型网格流程总结┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉47 第六节 GOCAD储层属性建模┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉48一、建立属性建模新流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉48二、属性建模操作流程┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉48三、属性建模后期处理┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉66四、网格粗化┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉74 第七节 GOCAD地质解释和分析┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉78GOCAD综合地质与储层建模软件操作手册第一节GOCAD综合地质与储层建模软件简介Gocad是国际上公认的主流建模软件，在众多油公司和服务公司得到了广泛的应用。
GOCAD(外文资料) (13)
Using Earthquake Data to Map Faults in 3-D with Gocad: Examples at Different ScalesSara CarenaDept. für Geo- und Umweltwissenschaften, Sektion Geologie, Luisenstr. 37, 80333 München, Germany. Email: scarena@geophysik.uni-muenchen.de.Eine Reihe neuer Techniken zur 3-D Kartierung von Störungen wurde entwickelt, deren Anwendung von Modellen einzelner Störungsoberflächen oder kleiner Gruppen von Störungen bis hin zu regiona-len tektonischen Modellierung reicht. Ein Beispiel zur Anwendung dieser Techniken an einzelnen, kleinen Strukturen, ist die Northridge Überschiebung in der westlichen Transverse Ranges von Südka-lifornien. Die 3-D Geometrie der Störung, die das M 6.8 1994 Northridge Erdbeben erzeugte, wurde durch Nachbeben bestimmt. Auch größere Störungssysteme können modelliert werden. Ein Beispiel hierfür ist das San Andreas Störung System nahe des San Gorgonio Passes in der östlichen Transverse Ranges von Südkalifornien. Eine Studie der 3-D Krustenstruktur im zentralen Taiwan ist ein Beispiel für die Modellierung regionaler Tektonik und Gebirgsbildung. Zum ersten Mal wurde hier die große Sohlfläche unter Taiwan abgebildet.Several new techniques in 3-D fault mapping have been developed, whose applications range from models of single fault surfaces or small fault networks, to regional tectonic models. An example of how these techniques can be applied to single structures is that of the Northridge thrust, western Transverse Ranges, southern California. The 3-D geometry of the fault that generated the M 6.8, 1994 Northridge earthquake was determined from the aftershocks of this event. Larger fault systems can be modeled too. An example of this is the San Andreas fault system near San Gorgonio Pass, eastern Transverse Ranges, southern California. A study of the 3-D structure of the crust in central Taiwan is instead an example of modeling applied to regional tectonics, and mountain building in particular. For the first time the large detachment beneath Taiwan was imaged.1 Earthquake data and 3-D Fault Mapping1.1 IntroductionTraditionally, the problem of determining the detailed 3-D geometry of active faults has been ad-dressed either by collecting geological and structural information at the surface of the Earth and ex-trapolating the geometry at depth using various theories and assumptions, or by direct imaging with seismic reflection techniques and wells, as it is routinely done in oil exploration. Earthquake data are often an under-utilized source of information concerning 3-D fault geometry. S HAW & S HEARER (1999) and S HAW ET AL. (2002) showed the benefits of combining earthquake data with seismic re-flection data in 3-D fault modeling and in better defining regional earthquake hazards. Hopefully, this kind of approach will become more common.1.2 Main Issues: Earthquake Size, Aftershocks, Location Quality, FocalPlane SolutionsHypocenter locations and focal mechanisms can provide direct information about the geometry of active faults at depths far greater than any other method. There are several key issues however that need to be addressed before using earthquake data to model fault geometry, namely earthquake size, density and distribution of earthquakes on faults, and data quality.In seismology, large earthquakes have often been the preferred means of finding out the orientation of the fault plane, by determining the earthquake focal solution. This approach usually results in fault geometries that are exceedingly simple, when not outright wrong. For example, early models of the Chelungpu thrust in Taiwan (e.g. L EE & M A 2000, K IKUCHI ET AL. 2000) that show a uniform 30°58. Berg- und Hüttenmännischer Tagdipping plane to depths of 20 km are incorrect, because they assume that the fault keeps a constant strike and dip with increasing depth. In this case however the shallow main-shock focal mechanism is not representative of the entire geometrically complex fault, which flattens to a sub-horizontal de-tachment at 5-6 km depth (Y UE ET AL. 2005). In other instances, the slip direction in a large earth-quake does not coincide with the long-term slip direction (see the Northridge thrust, C ARENA & S UPPE 2002).Often most aftershocks are not located on the main fault that ruptured, but on nearby ones (L IU ET AL. 2003). Ambiguities could be avoided by imaging the entire fault network in the area using background seismicity and previous earthquake sequences. This approach can help us to recognize the principal plane by matching it with previous events with similar focal mechanisms. Microearthquakes (M≤3) are particularly useful for this purpose due to their abundance. Networks that can reliably record mi-croseismicity already exist in a few regions of the world, and more are being added.The most important issue to consider is the quality of the earthquake data. In recent years, several authors have developed new methods of improving earthquake hypocenter locations significantly (G OT ET AL. 1994, J ONES & S TEWART 1997, R UBIN ET AL. 1999, N ICHOLSON ET AL. 2000; R ICHARDS-D INGER & S HEARER 2000, W ALDHAUSER & E LLSWORTH 2002). In our work we gave preference to the clustering method of J ONES & S TEWART (1997) because it can be quickly applied to large earth-quake catalogs (hundreds of thousands of events), because the only information needed about the earthquakes are their location and location uncertainties, and because it does not result in the exclusion of some events. Of course, the more details we want to extract about the fault 3-D geometry, the more a need arises for combining different methods in order to get the best possible hypocenter locations. In fact, even when using the clustering method, the ideal approach is to start with an earthquake catalog that has already been relocated to minimize systematic errors as much as possible (J ONES & S TEWART 1997).Focal mechanisms are extremely useful in 3-D fault modeling because, besides providing confirmation of fault geometry when there is an abundance of hypocenter locations, in certain cases they provide enough additional information to image even faults with a small number of events on them. Unlike event locations, that only provide information about fault shape, focal mechanisms allow us to deter-mine fault type and direction of motion. Thus we fitted fault surfaces simultaneously to both earth-quake hypocenter locations and principal planes extracted from focal mechanisms whenever possible. The main problem with focal mechanisms is their reliability, which decreases with decreasing earth-quake size, as the event is recorded only at a few stations.1.3 Integration of Other Data TypesEarthquakes are not the only kind of data that can be successfully used in 3-D fault modeling, and earthquakes do not occur everywhere. Therefore other types of data should be integrated in the model whenever available to fill any gaps and to better constrain it. S HAW & S HEARER (1999) and S HAW ET AL. (2002), for instance, have shown how combining seismic reflection profiles with relocated earth-quake data can significantly improve results when imaging blind faults.Surface geological data, like strike and dip measurements, stratigraphy, and the detailed location of fault traces and earthquake surface breaks are also key data, because seldom large numbers of earth-quakes occur at depths shallower than a few kilometers. For example, the aftershocks of the Northridge earthquake occurred mostly below 3-5 km depth and did not provide any information about fault geometry, or even the presence of any faulting, at shallower depths (C ARENA & S UPPE 2002). In the case of the Chelungpu thrust in Taiwan, very few aftershocks of the 1999 Chi-Chi earthquake oc-curred on the thrust itself; most of them were deeper and occurred on nearby faults. Nevertheless, Y UE ET AL. (2005) were able to generate a detailed 3-D fault model on the basis of strike and dip data, inte-grating them with surface breaks and several seismic profiles, and then connecting their model with the deeper fault network imaged from seismicity.Digital elevation models are necessary to transform a map of fault traces or surface breaks into a 3-D trace that can be directly incorporated into the fault model. Fault depth measurements in wells are also a very good constraint when available, as they are yet another type of data that can give us information about fault geometry at shallow depths.Sara Carena Using Earthquake Data to Map Faults in 3-D with Gocad: Examples at Different Scales1.4 Some Techniques in 3-D Surface BuildingRealistic fault geometry is first of all a geometry that allows the fault to slip. Large jogs perpendicular to the long-term slip direction and large bumps are unrealistic, because not only they would impede slip itself, but also because slip on the fault would quickly destroy them. On the other hand, corruga-tions parallel to the slip direction can certainly exist and have been used by some authors to constrain the 3-D fault geometry when modeling faults. Based on faults whose geometry is well known due to direct observation of the fault surface or to numerous wells intersecting it,T HIBAUT ET AL. (1996) developed the concept of faults as “thread surfaces”. They improve the results of interpolating be-tween scattered points (a common characteristic of fault data) by assuming that faults behave in a way similar to a nut-and-bolt system, with the contact surface between the two blocks dented by grooves (threads) produced by slip on the fault. They obtain the orientation of the threads (which are lineations on the fault plane at any scale, from striae visible in the field, to corrugations of the fault surface with wavelengths up to 10 km or greater) from areas with denser data, and apply their thread criterion to the interpolation in areas of low data density, thus obtaining a consistent fault geometry everywhere. A similar approach is that of M ALLET ET AL. (1999) and M ALLET (2002), who show how fault corruga-tions imaged from actual data can be combined with the assumption that long-term slip must occur parallel to the corrugations to further improve a 3-D fault model, removing bumps on the surface that would impede slip and which are most likely the result of scattering or gaps in the data.2 A New Approach to 3-D Fault MappingOur approach requires handling up to several hundred thousands of earthquakes at once, and the abil-ity to integrate different data types into a single model. This is why Gocad was chosen as a modeling toolbox. While the main focus of Gocad is geomodeling, it was not designed specifically with han-dling earthquakes in mind, so we had to devise some procedures to be able to use it to manipulate earthquake data efficiently.We start from hypocenter locations that have been relocated if possible. The hypocenter locations are then clustered using the clustering method developed and described by J ONES & S TEWART (1997) and modified by N ICHOLSON ET AL. (2000). Clustering results in tighter earthquake distributions (a “sharper” image, see fig. 1), which makes the process of selecting subsets of earthquakes much easier.Fig. 1. Example of earthquake hypocenters (aftershocks of M 7.1, 1986, Loma Prieta, California, earth-quake) before (a) and after (b) application of the clustering algorithm. (c) and (d) show the mis-fit to the 3-D fault surface.58. Berg- und Hüttenmännischer TagFault geometry can be constrained by focal mechanisms as well. The comparison between earthquake hypocenter distribution and focal mechanisms in 3-D allows us to [1] distinguish between principal and auxiliary nodal planes, thus making it possible to select only the principal planes and vectors in the data set, [2] identify and map faults which have only a few events associated with them, [3] deter-mine the current slip direction on faults. Nodal planes and slip vectors are imported into Gocad as surfaces and lines. Once one of the two nodal planes has been identified as the principal plane, focal mechanisms can then be transformed into point sets and used directly in fault surface building.A third type of constraint that can be applied to fault geometry is surface traces, either in the form of known breaks caused by a specific earthquake (C ARENA & S UPPE 2002), or as mapped fault traces. Surface traces constrain the position of the top of the fault and may disclose a near-surface change in fault dip that could have gone undetected due to the general lack of earthquakes at shallow depths. Once earthquake hypocenter locations have been relocated and/or clustered, and all the other data types have been transformed into a suitable format, we separate clusters of earthquake hypocenters that illuminate different faults. This is the most subjective part of the procedure, as it has to be done manually, but in most cases the clusters are fairly obvious features when viewed in 3-D. Different operators performing the selection could include or exclude a few different hypocenters, but in the majority of cases where there is a recognizable cluster, these differences are limited to the outer edges of the cluster. Because the steps that follow the selection always include some form of averaging of the hypocenter locations, small differences in the initial selection of hypocenters will not have any appreciable influence on the final fault geometry.Occasionally, it is not possible to separate clusters adequately because they intersect forming X- or complex junctions. In such cases only better earthquake locations might solve the problem, or at least reduce uncertainties (for example, once earthquakes are relocated, an X-junction might turn out to have one through-going fault, and one slightly offset fault). At other times, faults come together in T- or Y- junctions. The earthquake clusters can usually be separated at such junctions, but there will be uncertainty as to where exactly the truncated fault stops. Also, at the junction earthquake density is often higher and there can be more scattering than far from it, thus when separating the clusters many earthquakes of the truncated fault might be included in the through-going fault cluster. Situations where T- and Y- junctions exist can of course be improved too by better relocation of the earthquakes. Another kind of useful information in separating these types of clusters is their timing: if the two faults are active in different periods of time, then the clusters can be separated by accounting for both posi-tion and timing of the events.As previously mentioned, any other data points relative to fault position should also be considered, for example the depth of a fault in oil wells, or seismic lines on which a fault has been identified (for an excellent example, see S HAW & S HEARER 1999, and S HAW ET AL. 2002). These additional data types are especially important: they are usually available for depths up to a few km below the surface, but this is precisely the depth range where often very few earthquakes occur, and allow us to image the shallow geometry of the fault.For each fault, all different types of data should be merged so that they can be used simultaneously to build a first approximation of the fault surface. In our case we transform all data into sets of points. We choose points as our basic data type because the majority of our data are earthquake hypocenter locations, which are already discrete points. We then build our surfaces directly from the set of points, setting constraints as needed and then smoothing the resulting surface with the Gocad Discrete Smooth Interpolator (DSI, M ALLET 2002). Several iterations of the DSI are usually necessary before all the irregularities in the surface that have a size below the minimum resolution of the data are smoothed out.2 Applications and ResultsWe have applied the techniques described above to several cases at different scales and in different tectonic settings, and three of these are briefly described below: [1] a thrust fault associated with a restraining bend (Northridge thrust, southern California), [2] a network of strike-slip and thrust faults (San Andreas fault and surrounding faults near San Gorgonio Pass, southern California), and [3] an entire orogen in a subduction margin setting (Taiwan).Sara Carena Using Earthquake Data to Map Faults in 3-D with Gocad: Examples at Different Scales2.1 Northridge Thrust, Southern CaliforniaThe aftershocks of the 1994 Northridge earthquake (M 6.8) illuminate the structure beneath the San Fernando Valley (northwestern Los Angeles) in 3-D. We combined aftershocks and geological data to build an image of the 3-D geometry of the north-vergent Northridge blind thrust to a depth of 21 km. The most striking feature of the imaged fault is mega-corrugations oriented parallel to the mean after-shock slip vector, with most of the 1994 slip confined to west of the largest corrugation (lateral ramp, fig. 2). We also imaged the partially overlying south-vergent San Fernando thrust, which broke to the surface in a complex rupture in 1971 (M 7.1). Both thrusts produce fault-related folding because of either fault propagation or fault bends (S UPPE 1983). This deeper folding however is masked by over-lying complex deformation in the cover, which is one reason why the Northridge thrust was not identi-fied until it ruptured in 1994. We used trishear fold modeling (E RSLEV 1991) based on our 3-D fault geometry to evaluate possible folding due to slip on the Northridge thrust as well as its interaction with the overlapping San Fernando thrust and with shallow structures in the cover. This example illus-trates the importance of earthquake data to structural geology and the value of its 3-D integration with surface and near-surface geological data.Fig. 2. 3-D model of the Northridge thrust. Coseismic slip distribution shown on the left, fault corruga-tions indicated on the right.2.2 San Andreas Fault Near San Gorgonio Pass, Southern California The 1200 km long San Andreas fault (SAF) loses its apparent continuity in southern California near San Gorgonio Pass (A LLEN 1957). This fact raises significant questions, given the dominant role of this fault in active California tectonics. What is the fundamental 3-D geometry and kinematics of the San Andreas fault system in this complex region? Is a through-going, San Andreas rupture from the Mojave desert to the Coachella valley possible? We explored the issue of 3-D continuity by mapping over 60 faults in this region to depths of 15-20 km from hypocenter locations and focal mechanisms. We were able to constrain the 3-D geometry of the SAF near San Gorgonio Pass from the 3-D geome-try of the fault network surrounding it, as the San Andreas itself appears to be aseismic here. The most likely configuration is for the San Andreas fault to merge into the shallow-dipping San Gorgonio Pass thrust northwest of Indio (fig. 3). We concluded that there is no direct continuity at present, but rather a network of faults, and the only kind of rupture possible for the SAF in this region is a complex rup-ture, involving both strike-slip and reverse faulting. GPS measurements also suggest that, despite the fact large motions must have occurred in the past, only minor ones are occurring today in this area (Y ULE & S IEH 2001, M EADE ET AL. 2002, Y ULE & S IEH 2003). Applying our findings about the fault geometry, we explored several simple earthquake scenarios, following K ING ET AL. (1994), to58. Berg- und Hüttenmännischer Tagdetermine the most favorable conditions for a through-going rupture of the San Andreas fault system from the Mojave desert to the Coachella valley (C ARENA ET AL. 2004).Fig. 3. Two different views of the detailed 3-D geometry of the San Andreas fault system between Cajon Pass and Indio. SGPT=San Gorgonio Pass thrust. F1, F2, F3 are tear faults. The gray surface below the faults represents the base of the seismogenic crust in this region.2.3 Taiwan OrogenActive deformation in the upper crust beneath central Taiwan is illuminated by 110,000 small (M=1 to M=4) earthquakes, including both background seismicity and aftershock swarms from larger events. When viewed in 3-D, it becomes clear that the seismicity is dominated by a major sub-horizontal band of events at about 10-15 km depth. The zone steepens below eastern Taiwan to 30°–90° and reaches depths of 30–60 km. We interpreted this feature as the Main Detachment of the mountain belt. Other planes of seismicity above and below abut against this detachment, indicating its through-going na-ture. Although the availability of focal mechanisms is limited because of the small size of most earth-quakes, the available mechanisms consistently show oblique slip with reverse component on the dip-ping part of the detachment. The imaged 3-D shape of the Main Detachment in relation to surface topography allows us a straightforward test of critical taper wedge mechanics and suggests that the first-order topography of Taiwan is controlled by the shape of the detachment. In fact, the reversal of topography at the crest of the mountain belt corresponds to the inflection of the Main Detachment under eastern Taiwan (fig. 4).Fig. 4. Relationship between topography and the Main Detachment.Sara Carena Using Earthquake Data to Map Faults in 3-D with Gocad: Examples at Different Scales This geometry is consistent with critical-taper wedge mechanics (D AVIS ET AL. 1983), and in particu-lar with homogeneous mechanical properties for the shallow brittle part of the wedge. The geometry of the detachment also indicates that it is very weak: the effective coefficient of friction on the detach-ment itself (µb*) that fits the data is 0.08 (C ARENA ET AL. 2002). Some of the most dangerous faults that break the surface, like the Chelungpu thrust, are connected to the Main Detachment at depth (Y UE ET AL. 2005).3 AcknowledgmentsI am grateful to Honn Kao and Robert Jones for their help with earthquake clustering. Many thanks also to Tony Dahlen for constructive discussion about critical taper wedge mechanics, and to John Suppe for the many insightful suggestions.The original earthquake catalog data used in this work are from the Southern California Earthquake Center (SCEC), and from the Taiwan Central Weather Bureau (CWB).4 ReferencesA LLEN, C.R. (1957): San Andreas fault zone in San Gorgonio Pass, southern California. –Geol. Soc. Am. Bull., 68: 315—350.C ARENA, S., & S UPPE, J. (2002): Three-dimensional imaging of active structures using earthquake aftershocks: the Northridge thrust, California. – J. Struct. Geol., 24: 887—904.C ARENA, S., S UPPE, J., AND K AO, H. (2002): Active detachment of Taiwan illuminated by small earthquakes and its control of first-order topography. – Geology, 30 (10): 935—938.C ARENA, S., S UPPE, J., AND K AO, H. (2004): Lack of continuity of the San Andreas fault in southern California: fault models and earthquake scenarios. – J. Geophys. Res.,109, B04313, doi:10.1029/2003JB002643.D AVIS, D., S UPPE, J., AND D AHLEN, F.A. (1983): Mechanics of fold-and-thrust belts and accretionary wedges. – J Geophys. Res., 88: 1153—1172.E RSLEV, E.A. (1991): Trishear fault-propagation folding. – Geology, 19 (6): 617—620.G OT, J.L., F RECHERT, J., K LEIN, F.W. (1994): Deep fault plane geometry inferred from multiplet relative reloca-tion beneath the south flank of Kilauea. – J Geophys. Res., 99: 15,375—15,386.J ONES, R.H., & S TEWART, R.C., 1997, A method for determining significant structures in a cloud of earthquakes. – J. Geophys. Res., 102: 8245—8254.K IKUCHI M.; Y AGI Y.; AND Y AMANAKA Y. (2000): Source Process of Chi-Chi, Taiwan Earthquake of September 21, 1999 Inferred from Teleseismic Body waves. – Bull. Earthq. Res. Inst. Univ. Tokyo, 75:1—13.K ING, G.C.P., S TEIN, R.S., AND L IN, J. (1994): Static stress changes and the triggering of earthquakes. – In: L ANGSTON, C.A. (Ed.): The 1992 Landers, California, earthquake sequence. – Bull. Seismol. Soc. Am., 84 (3): 935—953.L EE, S.-J., & M A, K.-F. (2000): Rupture process of the 1999 Chi-Chi, Taiwan, earthquake from the inversion of Teleseismic data. – Terrestrial, Atmospheric and Oceanic Sciences, 11 (3): 591— 608.L IU, J., S IEH, K., AND H AUKSSON, E. (2003), A structural interpretation of the aftershock "cloud" of the 1992 Mw7.3 Landers, California, earthquake. – Seism. Soc. Am. Bull., 93: 1333 — 1344.M ALLET, J.L. (2002): Geomodeling. New York, Oxford University Press: 624 p.M ALLET, J.L., M ASSOT, J., AND C OGNOT, R. (1999): Fault characterization. – Proceedings of the 19th Gocad Meeting, Nancy School of Geology, Nancy, France, June 14-17, 1999. ASGA (Association Scientifique pour la Géologie et ses Applications).M EADE, B., H AGER, B., AND K ING R. (2002): Block models of present day deformation in southern California constrained by geodetic measurements – Eos Trans. AGU, 83 (47), Fall Meet. Suppl., Abstract T62F-05.N ICHOLSON, T., S AMBRIDGE, M., AND G UDMUNDSSON, O. (2000): On entropy and clustering in earthquake hypocentre distributions. – Geophys. J. Int., 14: 37—51.R ICHARDS-D INGER, K., & S HEARER, P.M.. (2000): Earthquake Locations in Southern California Obtained Using Source Specific Station Terms. – J. Geophys. Res. 105 (5): 10939—10960.58. Berg- und Hüttenmännischer TagR UBIN, A.M., G ILLARD, D., AND G OT, J.L. (1999): Streaks of microearthquakes along creeping faults. – Nature, 400: 635—641.S HAW, J.H., P LESCH, A., D OLAN, J.F., P RATT, T.L., AND F IORE, F. (2002): Puente Hills blind-thrust system, Los Angeles, California. – Bull. Seismol. Soc. Am., 92 (8): 2946—2960.S HAW, J.H., S HEARER, P.M. (1999): An elusive blind-thrust fault beneath metropolitan Los Angeles. – Science, 283: 1516—1518.S UPPE, J. (1983): Geometry and kinematics of fault-bend folding. – Am. J. Sci., 283: 684—721.T HIBAUT, M., G RATIER, J.P., L EGER, M., M ORVAN, J.M.. (1996): An inverse method for determining three-dimensional fault geometry with thread criterion: application to strike-slip and thrust faults (Western Alps and California). – J. Struct. Geol., 18: 1127—1138.W ALDHAUSER, F., AND E LLSWORTH, W.L. (2002): Fault structure and mechanics of the Hayward Fault, Califor-nia, from double-difference earthquake locations. – J. Geophys. Res., 107: doi:10.1029/2000JB000084.Y UE, L.-F., S UPPE, J., AND H UNG, J.-H. (2005): Structural geology of a classic thrust belt earthquake: the 1999 Chi-Chi earthquake Taiwan (Mw = 7.6). – J. Struct. Geol., 27 (11): 2058—2083.Y ULE, D., & S IEH, K. (2001): The paleoseismic record at Burro Flats: evidence for a 300-year average recur-rence for large earthquakes on the San Andreas fault in San Gorgonio Pass, southern California. – GSA Cor-dilleran Section 97th Annual Meeting, and Pacific Section AAPG, Universal City, CA, 2001.Y ULE, D., & S IEH, K. (2003): Complexities of the San Andreas Fault near San Gorgonio Pass: implications for large earthquakes. – J. Geophys. Res., 108 (11): 2584, doi:10.1029/2001JB000451.。
GOCAD(外文资料) (4)
gocad
cubit callenges:
topology book-keeping in gocad is very different from bookkeeping in most FE meshers
each TSurf in gocad “owns” its own vertexes and borders, adjoining TSurfs do not reference common nodes
examples of lithospheric structure constraints:
active source seismology: wide-angle reflection/refraction lines (onshore&offshore), MCS lines (non-commercial products)
the CBM format (as I understood in 06/15/2006):
the CBM format (as I understood in 06/15/2006):
the CBM format (as I understood in 06/15/2006):
each volume is closed (and if parts were merged, each side of the volume shares vertexes along the edges), but adjoining volumes do not share a common surface, border surface is duplicated…
?
gocad
cubit
(FEA)
GOCAD(外文资料) (32)
SIDETRACK OPTIMIZATION IN GOCADKim Touysinhthiphonexay and Joseph Bradley,University of Colorado, BP Center for VisualizationAngus Jamieson and Tom Southren, Tech21 LimitedAbstractSidetrack program optimization is a time consuming process using current well planning tools. First, planners must examine the areas of the reservoir within practical reach of each well and calculate the cost options of sidetracking from different depths and hole sizes. Then they must select targets within the reservoir and design the optimum sidetrack program based on criteria such as drilling cost, well production, trajectory constraints, and collision risk. So far as we are aware, there is no software currently available to view the options rapidly and to plan the sidetrack trajectories for maximum production at minimum cost.This year, a major effort of the University of Colorado’s Drilling Visualization Research Consortium has been to use 3D visualization and automated design algorithms to optimize sidetracking programs for mature fields. The aim has been to produce prototype software which can allow the reachable areas from any well to be clearly visualized on a reservoir surface while interactively changing the dogleg severity, sidetrack depth, target inclination, etc. Thereafter, based on various cost and engineering criteria, the user would be able to identify a target, pick a well from which to sidetrack, and interactively ‘drag’ the sidetrack point up and down the well to view the required sidetrack trajectory in 3D, along with its costs.The sidetrack optimization functionality has been developed and is being incorporated into the Immersive Drilling Planner, which is an interactive Gocad plug-in for planning and editing well paths and platforms. This enhancement will enable multidisciplinary teams to use 3D visualization technology, state-of-the-art planning algorithms, and integrated subsurface data to make rapid decisions for sidetrack options in mature fields. Integrated planning and visualization could reduce weeks of planning and cost analysis to days, resulting in earlier production and cash flow, and additionally could help minimize surprises from undrillable well designs, collision risk, or difficult geology.IntroductionAs major oil and gas fields mature, sidetrack programs are commonly used to enhance production and extend the fields’ economic life. Updated 3D seismic data or new reservoir models may indicate field or pool extensions, bypassed pay zones, or new locations for water injection. Selective zone well testing may indicate regions of higher permeability because of better porosity or fracture development. In many fields a well-designed program of drilling horizontal sidetrack wells and fracture stimulations has significantly increased production, improved sweep efficiency, and led to additional reservoir recovery.The keys to success are a multidisciplinary approach, integration and careful analysis of diverse geological and engineering information, and a practical way to rapidly generate and evaluate alternative drilling trajectories. Current methods of sidetrack program optimization are not fully integrated and can take several weeks of work for a large field. The University of Colorado’s Drilling Visualization Research Consortium (DVRC) is focused on developing 3D visualization and optimization tools for drilling. This year, the DVRC is developing a new Gocad plug-in that should significantly improve and speed up the process of optimizing sidetrack programs for mature fields. Once undrained reservoir targets have been identified, new interactive functions display reach constraints from existing wells and rapidly generate sidetrack trajectory options which balance lost production and drilling costs against potential new production gains. Teams can then quickly evaluate and rank alternative plans by cost and drillability. Additionally, integrating sidetrack planning with the 3D visualization of geology and existing wells helps teams optimize sidetrack placement within the reservoir while minimizing well collision risks.Sidetrack Optimization ProcessAs with any decision with multiple criteria and multivariate data to consider, the problem of sidetrack program optimization can best be solved by the use of interactive visualization tools. The first step is to import and display relevant geologic and engineering data in Gocad. As a minimum, this data should include the target reservoir horizon(s), other geological structures of interest, and existing wells. Other relevant data (displayed as properties or additional objects) could include target geobodies, well logs, high porosity/permeability/fracture zones, fault surfaces, and depth-converted seismic volumes. The data for the existing wells should include the depths and sizes to which casing has been run and the current production for each well. This information is stored in a database that is used by the sidetrack cost model.The second step is to examine areasof the reservoir within practical reachof each well.After defining somebasic drilling criteria, such as desiredinclination at target and maximumdogleg severity, the user simplyplaces a 3d cursor at any point on awell. The software derives anddisplays a locus of reachable pointson the reservoir surface that meet thedrilling criteria. The user can movethe cursor up and down the well pathand/or adjust the drilling constraintsFigure 1. Reservoir area reachable by sidetrack from selectedto see what areas can realistically be reached from the chosen well.Using Gocad’s 3D visualization capabilities, the third step is to identify specific target locations within the sidetrack reach locus. Currently these target locations are stored for each potential sidetrack well as a target set of one or more points. Estimated reservoir production is stored as a property of each target set.The final step allows the user to generate and evaluate multiple sidetrack options interactively to drill to a selected target. The decision of which well to sidetrack has several criteria:1. How much production will be lost by sidetracking the depleted well?2. How much production will be gained by penetrating the new target?3. How much will the sidetrack cost?4. What are the chances of successfully and safely drilling to the target? Normally, the drilling engineer will choose the well with least current production; sacrificing known production in favor of uncertain production will always require some courage. The benefits of this choice, however, might be outweighed by unacceptable risks of collision with other wells, the risks of drilling in hostile geology, or the risk of missing the production sweet spot.The user selects a target and places a 3D cursor at any point on an existing well to generate and display the sidetrack trajectory in 3D. In the background, the sidetrack cost is calculated, based on a basic cost model involving casing section, milling costs, deployment costs, drilling footage, and well path complexity. Simply moving the cursor up and down the well path updates the sidetrack trajectory and cost for the current depth and hole size.Figure 2. Rapid analysis of alternative sidetrack options: a) which well?, b) which depth?3D visualization and integration with other data immediately shows sidetrack proximity to geologic features and to other wells. Should further adjustments to the initial sidetrack trajectory be necessary, the user can make these interactively and again see the cost effect immediately.ConclusionThe sidetrack optimization functionality described in this paper has been developed and is being incorporated into the Immersive Drilling Planner, which is an interactive Gocad plug-in for planning and editing well paths and platforms.This enhancement will enable multidisciplinary teams to generate optimum sidetrack trajectories for a mature field in a matter of hours, rather than the several weeks it now takes to do such complex work using existing tools.The Sidetrack Optimization tool is designed to take the mathematical effort out of the sidetrack planning process and to present the user with all the visual and numeric data required to make planning decisions. In addition, Gocad’s data integration and 3D visualization capabilities allow users to demonstrate clearly and unambiguously the justification for these million dollar decisions. The ultimate goal of this tool is to integrate geoscience and engineering data with 3D visualization and design technology in order to enable teams to deliver quickly the most cost-effective conceptual sidetrack design that will safely satisfy the well and field objectives.。
GOCAD(外文资料) (37)
Report16Forward Modeling GOCADModulesZhaojun LiuABSTRACTWe are developing GOCAD modules for the UTAM forward modeling codes.At this time,GOCAD module have been created for the2-D elastic code PSVR4,2-D acoustic code PP4,and3-D viscoelastic adaptive code VE3DADA.The inputfile and velocity model can be created in GOCAD,the forward modeling codes can be launched and the seismograms can be displayed with the GOCAD interface.I will present a short tutorial on their implementation at the sponsor meeting.INTRODUCTIONGOCAD is a powerful rendering and visualization software package that can con-struct3-D earth models for Geophysical,Geological and Reservoir Engineering appli-cations.It also offers a user-friendly interface that can be easily adapted to the users specific needs.I have rewritten the PSVR4,PP4and VE3DADA code so that they can read the velocity,density or/and Vs/Vp ratio values at each nodal point of a model created by GOCAD.The user can modify thefinite-difference parameters directly in the dialog window.An online help about utility setting the parameters has been implemented. The seismograms are displayed as a property”Amplitude”of a GOCAD Voxet or SGrid,and the snapshots are displayed as a property”pressure”or”particle velocity”of a GOCAD SGrid.355356REPORT16.FORWARD MODELING GOCAD MODULESINSTRUCTIONS FOR INSTALLATIONIf you want to use UTAM modules as an independent part of your GOCAD modules, you should:1.Create a parallel tree for utam and four parallel libs for wave,xwave, processing,xtools.2.Download utam.tar.gz from our web site and extract thefiles to you own disc.3.Run Install to installfile.4.Recompile the codes.INSTRUCTIONS FOR USAGE Figure16.1shows the GOCAD interface for the UTAM module.It will pop up when utam is executed.There are now four sub-menus under the UTAM module:”2D Acoustic”,”2D Elastic”,”3D Elastic”and”Display Results”.Each of the modules has several functions such as”Create Indat File”,”Create Vel File”and”Run Model”.I introduced the”2D Elastic”module in the midyear report(Liu,1997),so I show how to use the”2D Acoustic”module here.I will demo all three forward modeling modules at the sponsors’meeting.Atfirst,you should build a Voxet model with GOCAD,and update the properties velocity(vp)and density(ρ)for this model(see Figure16.2).You will also need to generate the voxets for the source and receiver locations.When you click the button”Create Indat File”under the module”2D Acoustic”, it will pop up a dialog window named”Create2D Acoustic Parameter File”,as shown in Figure16.3.You can choose the model object,source and receiver objects,and modify the parameters through this window.If you click on the button”OK”or ”Apply”,the parameterfile”indat”will be created.Then you can create thefile containing the P wave velocity and density values for the model by clicking on the button”Create Vel File”.The dialog window”Create2D Acoustic Model”is shown in Figure16.4.When you click on the button”Run Model”,you can choose two different ways to run the PP4code:one is to run on a xterm window,and the other is to process to a logfile.The function of”Display Seismogram on New Voxet”is to create a Voxet,the property of which is the amplitute of the seismic wavefield(Figure16.5).The source or receiver voxet should be chosen as a reference voxet to display the common receiver gather or the common shot gather.You can also choose”Display Seismogram on New SGrid”to display the seismograms on the GOCAD SGrid(Figure16.6),and you can choose different AGC windows to apply AGC gain to the seismogram.If you selected recording snapshot in executing the forward modeling code,you can click the button”Display Snapshot on New SGrid”(see Figure16.7)to display the snapshot(Figure16.8)at any recording time you selected.The modules for the2-D elastic code and the3-D elastic code can also be easyly used.357FUTURE WORKIn the future,we will modify other UTAM codes to be executable with GOCAD modules.These modules are still under development and thefinal form will depend on the feedback from our sponsors.ACKNOWLEDGEMENTSI would like to thank Marc Vallee and John McGaughey of NORANDA Inc.for their help,and their donation of some GOCAD subroutines.REFERENCESLiu,Z.,1997,Forward Modeling GOCAD Modules:1997Midyear Report of Utah Tomography and Modeling/Migration Development Project,p.201-206358REPORT16.FORWARD MODELING GOCAD MODULESFigure16.1:The main menu for UTAM module.359Figure16.2:A2-D model created in GOCAD.360REPORT16.FORWARD MODELING GOCAD MODULESFigure16.3:The dialog window for the function”Create Indat File”.361Figure16.4:The dialog window for the function”Create Vel File”.362REPORT16.FORWARD MODELING GOCAD MODULESFigure16.5:The dialog window for the function”Display Seismogram on New Voxet”and”Display Seismogram on New SGrid”.363Figure16.6:The seimograms which are displayed on the GOCAD SGrid with AGC=0.2sec.364REPORT16.FORWARD MODELING GOCAD MODULESFigure16.7:The dialog window for the function”Display Snapshot on New SGrid”.365Figure16.8:The10th and20th recording snapshots displayed on the GOCAD SGrid.366REPORT16.FORWARD MODELING GOCAD MODULES。
GOCAD(外文资料)(1)
Seismic Acoustic Impedance Inversion in Reservoir Characterization Utilizing gOcad By Steven Clawson and Hai-Zui (“Hai-Ray”) Meng, Presented at the 2000 gOcad Users MeetingIntroductionWorkflows for utilizing seismic data inverted to acoustic impedance data in reservoir characterization will be shown. We are using the public domain 3D seismic dataset at Boonsville Field in North-Central Texas for our example. This public domain dataset is fairly complete with seismic, well, and production data:•5.5 sq. Miles of 3D seismic data•Vertical seismic profile (VSP) near center of survey•Digital well logs from 38 wells•Well markers for the bend conglomerate group•Perforations, reservoir pressures, production and Petrophysical data for the 38wellsWe acknowledge Oxy USA, Inc., Enserch, Arch Petroleum, Bureau of Economic Geology, GRI and the DOE as contributing members for making this dataset available. This data is made publicly available as part of the technology transfer activities of the Secondary Gas Recovery (SGR) program funded by the U. S. Department of Energy and the Gas Research Institute.Boonsville Field is in the Fort Worth Basin in North-Central Texas.The main productive interval are clastic sandstones in the Pennsylvanian Atokan Bend Conglomerate Group.A type log shows the interbedded sandstones and shales over about 1300 feet of section. The Bend Conglomerate is underlain by the Marble Falls Limestone, a platform carbonate. The Bend Conglomerates were sourced from the northwest on the Red River Arch as the Fort Worth Basin was forming during the Oachita orogeny. These Bend Conglomerate sandstones then pinchout to the southeast, outside of this project area as they become distal to the source, prograding into the Fort Worth Basin.Historical gas production has been from the lower most sequence in the Vineyard. Additional potential is expected in the middle sequences of the Runaway and Vineyardintervals.Conglomerate Group.This example seismic line shows the Bend Conglomerate Group structure. Most striking are the karst collapse features from dissolution of the underlying Ellenburger Limestone, some 2000 feet below the Atoka. These collapse features are seen to causecompartmentalization in the Bend Conglomerate sand bodies.Previous conclusions from the Bureau of Economic Geology’s GRI study are:1) Karsting from Ellenburger carbonates cause collapse features compartmentalizing thereservoir. Large range of compartment sizes exist.2) Need 3D seismic to image the collapse features.3) Seismic attributes can sometimes predict the reservoir faciesUpper Caddo: AmplitudeLower Caddo: Inst. FrequencyLower Bend Conglomerate sequences not definitive4) Reservoirs often exist as stacked compartments of genetic sequences.The utility of the seismic attributes derived from the amplitude data are limited and typically very dependant on the particular interval analyzed. In this project we integrate the well log data in with the seismic for a better defined reservoir model. This integration is accomplished by inverting the seismic amplitude data to acoustic impedance (AI) properties and depth converting the seismic so correlation with the well logs is possible. In this presentation I will only highlight the features of Structural Framework and Rock Property modeling in the overall Reservoir Modeling workflows:Structural Framework => Stratigraphic Gridding => Litholgy and Facies Mapping => Pressure Field => Rock Properties => Fracture Network and Stress Field =>Reservoir Fluids and Dynamic Response.Motivation for Reservoir Modeling include:1) Integration of all relevant and available data.2) Merge data of different scales:(Cores, Well logs, Seismic and Production).3) Dynamically update the model as new information becomes available.4) Measurement of errors and uncertainty as well as expected value.The specific workflows used are dependant on number and type of data available. In this case there is substantial well control and the seismic data is of high resolution (80Hz). Structural Framework WorkflowThe Structural Framework Workflow is shown below:Obtaining the Structural Framework from the seismic gives a much better description than from the well control alone. The karst features were not known until the 3D seismic data was acquired.Integration of the well marker tops and the seismic time horizons proceeds by 2 pathways:1) A reference horizon (the Caddo Limestone) was an excellent reflector that also tied the well tops. This is depth converted by a co-located co-Kriging method.2) Time horizons below this reference did not exactly tie the associated well markers due to tuning effects of the thin bedded Bend Conglomerate Group. For these horizons a velocity field was constructed from interpolating the sonic logs, calibrated to the seismic and checkshot survey. The depth was then created by the time and velocity relationship.3)The fault network will be incorporated in the future using a seismic continuity analysis. Depth conversion of the reference horizon is accomplished thru the strong correlation between the time and depth relationship at the well locations.Co-located co-Kriging of the seismic time and well marker depths produces a very accurate depth structure for the Caddo Limestone.Interpolating the sonic logs in the survey a interval velocity field is produced. Converting these interval velocities to average velocities (inverse Dix’s equation) provides the information on depth converting the intervening horizons.And here are the depth converted intervening seismic horizons.Rock Properties WorkflowThis rock property modeling workflow utilizes the seismic information obtained via inversion to acoustic impedance to better control the well log interpolation of rock properties. This is also accomplished with the accurate structural information that the seismic provides. This workflow is necessarily iterative due to the dependency of one data on another and the iteration between time (on the seismic data) and depth (for the log data) referencing.Seismic to Log Calibration is the first step in integrating the seismic amplitude data with the log properties. Starting off one may not know other than by qualitative correlation what the seismic wavelet is. In this case a reserse polarity wavelet is assumed. The synthetic is then tied to the seismic data, performing a constrained stretching and/or squeezing to fit major events. This stretching/squeezing is primarily due to dispersionbetween seismic velocities and sonic log velocities.A final seismic wavelet is then extracted. Always use more than a single seismic to log calibration tie. In this case 4 well ties were averaged for a consistent wavelet showing that the seismic wavelet is nearly –90degrees out of phase and slightly ringy. The ringing suggests that the deconvolution was not sufficient to collapse the source wavelet. Theseismic bandwidth is very good (20-80Hz).A background acoustic impedance model is needed to supply the low frequency component missing from the seismic trace data in the inversion. This first iteration uses asimple gridding of the 4 sonic logs in the survey.A model based inversion using Hampson-Russell Software’s Strata program shows the transform of the qualitative amplitude data into rock property information. The result is very dependant on the background model used and later we’ll see an improvedbackground model for a better result.Checking this inversion at our key well: B Yates 18D the seismic inverted acoustic impedance ties well qualitatively with well log acoustic impedance. Depth converting thisAI volume is also compared to the well log for quality control.Now that we ha ve seismically derived rock properties from the seismic in depth, let’s see how they correlate to the well logs. In general we see that:1) Low AI relates to shales from the gamma ray log.2) High AI relates to resisitve sandstones from the RT log.3) Correlation of AI to the porosity is more complicated since the shales measure a highporosity with low AI and the more porous sandstones are in an intermediate range of AI,while the tight sandstones are resistive and also high AI.properties show a rather low correlation coefficient.An observation of the relative scales of information is needed. The well logs of course are of higher resolution than the seismic data as shown in the lower variance of AI derived by the seismic data than that represented in the well log data. Smoothing the log curves is required to be able to statistically correlate the respective information. This correlation is also stongly influenced by the exact depth conversion of the seismic information to tie the wells. Due to the thin bedded nature of the Bend Conglomerate Group a mistie of only afew feet will severely effect the correlation.Cross-plotting the seismically derived AI to the smoothed well logs (20 feet averaging) increases the correlation, as now the data are on a more equal sample support resolution. These correlations are still low. These seismically derived AI values are also influencedby the simple background impedance model used in the inversion.in the survey.The well log acoustic impedance (AI) is highly correlated to the Log10(RT). The spatial variogram shows a fairly long range to the correlation in order to provide a goodbackground AI model for a 2nd iteration of inversion.First Kriging the Log10(RT) logs is performed. Next co_Kriging the 4 wells with acoustic impedance information is run. Spatially this new background impedance model is shown to provide spatial features not available with just the 4 wells with sonic logs. Areas near the well control have very high frequency information content. While away from well control the response is subdued towards an average from the Kriging system. Since the seismic is principally used for interpolating the interwell region this background impedance model is low pass filtered to 20Hz. This way the well control is only adding the very long wavelength trends to the inversion result. And the interwellregion should be justly controlled by the seismic data.rock properties.Qualitative correlation to the key well: B Yates 18D yields similar results as before.Now cross-plotting the seismically derived acoustic impedance and the log properties in depth shows a better correlation. These correlations are good enough to use in a co-located co-Kriging of the well log properties.Rock property models are now generated by co-located co-Kriging of the gamma ray logs for lithology discrimination and resistivity logs controlled by the seismically derived AIproperties.A reservoir model of sandstone porosity can be derived by the relationships of lithology to gamma ray and resistivity. Where these models of gamma ray and resistivity are related back to the seismically derived acoustic impedance.By segmenting the data into a sandstone region defined by where:Gamma ray is less than 90 andLog10(Resistivity) is greater than 0.8A sandstone porosity relationship is defined.Constructing the density model in the sandstone facies then is represented here.。
GOCAD(外文资料) (3)
A Short Note on the Integration of GSLIB and GOCADW.Glenn(wglenn@ualberta.ca)University of Alberta,Edmonton,Alberta,CANADAThis short note describes some of the work undertaken by Will Glenn as a summer intern supported by T-Surf to incorporate some of the latest GSLIB/CCG codes in gOcad.Please contact C.V.Deutsch or T-Surf to discuss implementation at your site.Users guide to implement GSLib dll in GocadHere are some general guidelines related to using the GSLib DLL in gocad.Note that this approach requires a C:/temp folder to hold GSLib intermediate files.Cell declustering(declus)-see below-requires a loaded Pointset with a property and assigns the declustering weights to each point in Vset.The steps are summarized by(1)display the Vset and property in the camera,(2)select declus from the GSLib algorithms menu,(3)the declustering dialog window will pop up,(4)enter the name of the Pointset,(5)select the Pointset property,(6) enter the Y and Z anisotropy variables,(7)number of cell sizes to consider in the operation,(8) minimum and maximum cell size,(9)number of origin offsets(noff),(10)select the check box for maximum declustered mean,(11)choose a suitable name for the declustered weight,(12) enter the minimum and maximum trimming limits,and(13)select the check box for Post deleting intermediate files(this is recommended if no further file analysis or file debugging is needed).Soft data declustering(sddeclus)requires a loaded Pointset with a property and a loaded Voxet with cell properties.The result is a file with a representative histogram.The steps aresummarized by(1)display the Pointset with property and Voxet with cell property in the camera,(2)select sdeclus from the GSLib algorithms menu,(3)the soft declustering dialog window will pop up,(4)enter the name of the Voxet,(5)enter the name of the Vset,(6)enter the name of the Vset property,(7)enter the name of the Voxet property,(8) enter the number of secondary classes to consider,(9)GSLib sddeclus output file,(10) minimum and maximum trimming limits of data,and(11)post deleting intermediate files. This is recommended if no further file analysis or file debugging is needed.Stepwise conditional transformation(stepcon)requires a loaded Pointset with two or more properties.The result is a file with new colums for the transformed values and another file with the output transformation table.The steps are summarized by(1)display the Pointset with property and Voxet with cell property in the camera,(2)select stepcon from the GSLib..Algorithms menu,(3)the stepcon dialog window will pop up,(4)enter the name of the pointset,(5)enter the primary property,(6)enter the secondary property,(7)if there are other properties of the pointset that must be used,select the properties accordingly,(8)enter the number of secondary classes,(9)an optional transformation table file can be considered in the algorithm by indicating its location,(10)the GSLib output file containing the original data set with additional columns with the new conditioned data,(11)the GSLib output file containing the transformation table,(12)minimum and maximum data trimming limits,and(13)post deleting intermediate files.This is recommended if no further file analysis or file debugging is needed. The reverse of the stepwise conditional transformation(backstep)requires two data files and the transformation data.The result is the back transformed values.The steps are summarized by(1) select stepcon from the GSLib algorithms menu,(2)the backstepcon dialog window will popup, (3)enter the number of variables,(4)enter data file number1,(5)enter data file number2,(6) select check box for smoothed distribution,(7)enter the input transformation table,(8)enter theunivariate transformation table file for variable1,(9)enter the univariate transformation table file for variable2,and(10)select the output file.The sequential Gaussian simulation program with self healing(sgsim_sh)requires two pointsets and generates a new SGrid properties for each cell,a simulation output file,an MVF for simulation output file.The steps are summarized by(1)display the two Poinsets with property and SGrid in the camera,(2)select sgsim_sh from the GSLib algorithms menu,(3)the sgsim_sh dialog window will pop up,(4)enter the name of the primary Pointset,(5)enter the primary property,(6)a and b:select the weight checkbox if there is a weight property in the primary Pointset to be used in sgsim_sh and enter the property name,(7)a and b:select the secondary checkbox if there is a secondary property in the primary Pointset to be used in sgsim_sh and enter the property name,(8)enter the name of the secondary Poinset,(9)enter the secondary property to be used from the secondary pointset,(10)enter the Sgrid,(11)a and b:select the checkbox to consider the reference distribution file,(12)a and b:enter the minimum and maximum Z values, (13)a and b:enter the lower tail option and parameter,(14)a and b:Enter the Upper tail option and parameter,(15)enter the simulation output file,(16)enter the mvf file for simulation output, (17)enter the number of realizations,(18)a and b:Enter the minimum and maximum original data for simulation,(19)enter the number of simulation nodes,(20)select the checkbox to assign data to nodes,(21)select the checkbox to perform a multiple grid search,(22)enter the number of multiple searches,(23)enter the maximum data per octant(0=not used),(24)a,b,and c:-maximum search radii(hmax,hmin,vert),(25)a,b,and c:-angles for search ellipsoid,(26)a,b, and c:-size of covariance lookup table,(27)kriging type,(28)colocorr,(29)varred,(30)a,b,and c:a0v,a1v,acc,(31)select checkbox to transform secondary variable,(32)enter the nugget effect,(33)select the Variogram type,(34)enter the cc parameter,(35)a,b,and c:enter the angle parameters,and(36)a,b,and c:enter the a_hmax,a_hmin,and a_vert parameters.Note1A:the SGrid must be an translation of the original XYZ grid.Surface simulation(surfsim)requires a loaded Pointset generates four output files with the surface parameters.The steps are summarized by(1)display the poinset and Sgrid in the camera,(2)select surfsim from the GSLib..Algorithms menu,(3)the surfsim dialog window will pop up,(4)enter the name of the poinset,(5)select the Sgrid to use for the algorithm,(6)enter the Maximum number of surfaces,(7)enter the total thickness,(8) select the surface type–gauss-helmet-ellipse,(9)a,b,and c:Enter the L,M,and U height of surface,(10)a,b,and c:(simple)Enter the L length,(ellipse)M long axis,and(guass) Sigma X,(12)a,b,and c:(simple)Enter the L width,(ellipse)short axis,and(gauss) Sigma Y,(13)a,b,and c:(simple)(not for ellipse&gauss)L,M,U width2,(14)select box for truncation,(15)select box for it variable,(16)a,b,and c:if trunction(14),Enter the L,M,and U of X0,(17)a,b,and c:if trunction(14),Enter the L,M,and U of Y0,(18) select box for itrans variable,(19)enter the nugget effect,(20)enter the cc variable,(21) a,b,and c:enter the L,M,and U Sigma of residual,(22)a,b,and c:enter angle_1, angle_2,and angle_3parameter,(23)a,b,and c:enter the a_hmax,a_hmin,and a_vert variables.Note2A:the SGrid must be a translation of the original XYZ grid.Create a Voxet from GSLIB takes a GSLIB data file and generates a file for GOCAD.The steps are summarized by(1)select Create_Vset_from_GSLib from the GSLib-tools menu,(2)enter an appropriate name for the Pointset,(3)select a GSLib data file from the file selector,(4)ct checkbox if GSLib data file contains a well prefix(ie.Is the first column in the GSLib data file a well number?),and(5)select checkbox if the GSLib data file contains a column for Z values(ie. Is the file in3-D or2-D format?).Load a Voxet.The steps are summarized by(1)select Load Voxet Property from the GSLib-tools menu,(2)choose the Voxet that will have the additional property,(3)give the new property a name,(4)choose the GSLib file A that contains the property,(5)select the checkbox if the GSLib file has a second property that is to be included in the Voxet,and (6)depending on step5,Enter the second property.Load a SGrid.The steps are summarized by(1)select the load_sgrid_property from the GSLib..tools menu,(2)choose the SGrid that will have the additional property,(3)enter a property name,(4)choose the GSLib file that contains the property,(5)select the checkbox if the GSLib file has a second property that is to be included in the SGrid,and (6).depending on step5,Enter the second property nameHow to integrate the GSLIB DLL in MSDEV C++:This is an example implementation of the IDECLUS function using the GSLIB.dll and GSLIB.lib.1download and install the VFRUN65AI.exe from:/fortran/kb/q1023.htmlthis exe will install the necessary dlls to run the DECLUS.dll or any other fortrandll created by visual fortran in C++code.2.Place the GSLIB.dll in the system path3.In the MSVisual Studio environment,update include the GSLIB.lib file in theProject->settings->Link(tab)->Object/library modules:linee the sample C++code to run the IDECLUS function inside theimplementation filesthe IDECLUS function requires a.dat and.par file inside a sample.cpp file:#include<string.h>…char instring[50];char outstring[50];strcpy(instring,"C:/temp/declus.par");strcpy(outstring,"C:/temp/data.out");/*Call Fortran routine-pass length of outstring explicitly*/int warnint=IDECLUS(instring,strlen(instring),outstring,50);//if warnint=0-Normal operation//if warnint=1-Missing parameter or input file//if warnint=2-Error in parameter file//if warnint=9-Extraneous Unknown error…inside a sample.h file:…extern"C"//must declare extern“C”to avoid C++name mangleing//IDECLUS being the name of the function{__declspec(dllimport)int__stdcall IDECLUS(char*STR_IN,int STR_IN_LEN,char*STR_OUT,int STR_OUT_LEN);}General Gocad Plugin informationFor future development of CCG Gocad plug-ins.The CCG Gocad plug-in was developed using the three stage implementation.Each plug-in element begins with the source stage. The source files are interfaced with the API stage.A single CLI command encapsulates the entire process of calling the API and source objects.Specific Libraries were tailored for the CCG Plug-in.the libraries are InputOutput,and Algortihms.The InputOutput library contained most of the file reading and writing.The Algortihms library contained all of the CCG_DLL.dll functionalities and GSLib parameter file creations.The source contains the GSLib function calls and extern”c”declarations.It also contains the file writing procedures for the GSLib intermediate files. The GAPI contains the calls to the source classes and performs error checks.It also prepares Gocad to receive properties and contains calls to update properties.The GAPI classes that the CCG has developed in summer2001have contained only public static functions(similar to Gocad api convention)that mainly return Boolean values as output. Inside the GAPI static functions the main source classes are called.In the CCG Gocad plug-in each library that was created,there was a corresponding GAPI file used to call the library’s code.The CLI and XCLI simplified the creation of the user interface windows.Inside the CLI commands the CCG used file selectors,Object selectors,property selectors,parameter entry fields,and Boolean variables to create a dialog to users.The CLI commands call the GAPI static functions.The menu-bars section defines the menu item names,CLI calls,and ordering of menu layout.The Gocad plugin takes care of the implementation of the menu-bar,so developers need only worry about the ordering of the menu items.Instructions on how to use each GSLib dll function in Gocad Using specific parameters and files.declus:Using this parameter file:START OF PARAMETERS:../data/cluster.dat \file with data1 2 0 3 \ columns for X, Y, Z, and variable-1.0e21 1.0e21 \ trimming limitsdeclus.sum \file for summary outputdeclus.out \file for output with data & weights1.0 1.0 \Y and Z cell anisotropy (Ysize=size*Yanis) 0 \0=look for minimum declustered mean(1=max)24 1.0 25.0 \number of cell sizes, min size, max size 5 \number of origin offsetsSteps:1.In GSLib..tools..Create VSet from GSLib,2.Enter the file cluster.dat as the GSLib file.3.make sure the check boxes are unchecked because cluster.dat does not have columns for well id orZ4.In GSLib..Algorithms..declus5.choose the Pointset previously entered and enter“Primary”property in the property boxe the default parameters(they are the same as the above)stepcon:Using this parameter file:START OF PARAMETERS:data.dat - file with data2 - number of variables to transform3 5 - columns for variable transformation-primary and Secon-1.0e21 1.0e21 - trimming limits10 - number of classes1 - input transformation table,yes=1,no=0scatsmth_k.trn - file with transformation table stepcon.out - file for outputstepcon.trn - file for output transformation tableSteps:1.In GSLib..tools..Create_Vset_from_GSLib,2.Enter the file cluster.dat as the GSLib file.3.Make sure the check boxes are unchecked because cluster.dat does not have columns for well id or Z4.In GSLib..Algorithms..stepcon5.choose the Pointset previously entered and enter“Primary”in the primary box and“Secondary”inthe secondary box.e the default parameters(they are similar to the ones above).7.Out put files should appear in the C:/temp folder.ssdeclus:Using this parameter file:START OF PARAMETERS:pairs.dat -file with paired calibration data1 2 - columns for secondary and primary-1.0 1.0e21 - trimming limitsseismic.dat -file with secondary data1 - column for secondary variable sddeclus.out -file for representative histogram50 -number of sec. classes to use (10-100)Steps:1.In GSLib..tools..Create_Vset_from_GSLib,2.Enter the file welldata.dat as the GSLib file.3.make sure the check boxes are checked because welldata.dat does have columns for well id and Z4.In Objects right click on Voxet..New..From_Objects_cage.5.give the Voxet a name,set Scale W dimention by1,set nu=81,nv=81,nw=787.Populate the Voxet nodes with properties from the file“seisimic.dat”.This file contains only onecolumn with the property“impedence”.The file data will expire before filling all Voxet nodes.8.go to Voxet..Attributes..select the Object and make the“volume”atoms visible.(this will slowthings down!)6.In GSLib..Algorithms..sddeclus7.Put in the above Voxet,Pointset,and“Porosity”property from Pointset in primary property and“impedance”in the secondary property.e the rest of the default parameters(similar to the above parameter file)uncheckPost_delete_par_file to view the intermediate GSLib files(ie.the pairs.dat file).backstep:Using this parameter file:START OF PARAMETERS:2 - number of variablessgsimscpor.out - data file number 1sgsimscper.out - data file number 21 - smoothed distribution, 1=yes, 0=noscatsmth_kfl.trn - file with input transformation table nspor.trn - univariate transformation table for variable 1 nsper.trn - univariate transformation table for variable 2 backfl.out - file for outputSteps:Not much to this Program,Just put in the files and Select where the output files go!sgsim_sh:Using this parameter file:START OF PARAMETERS:../data/cluster.dat -file with data1 2 0 3 0 4 - columns for X,Y,Z,vr,wt,sec.var.-1.0e21 1.0e21 - trimming limits1 -transform the data (0=no, 1=yes) sgsim.trn - file for output trans table0 - consider ref. dist (0=no, 1=yes) histsmth.out - file with ref. dist distribution1 2 - columns for vr and wt0.0 25.0 - zmin,zmax(tail extrapolation)1 0.0 - lower tail option, parameter1 25.0 - upper tail option, parameter1 -debugging level: 0,1,2,3GSLvm-SH.dbg -file for debugging outputGSLvm-SH.out -file for simulation outputGSLvm-SH-mvf.mvf -mvf file for simulation output mvffl 1 -number of realizations to generate25 0.5 2.0 -nx,xmn,xsiz25 0.5 2.0 -ny,ymn,ysiz1 0.5 1.0 -nz,zmn,zsiz123065 -random number seed0 8 -min and max original data for sim12 -number of simulated nodes to use1 -assign data to nodes (0=no, 1=yes)1 3 -multiple grid search (0=no, 1=yes),num 0 -maximum data per octant (0=not used) 15.0 15.0 10.0 -maximum search radii (hmax,hmin,vert)0.0 0.0 0.0 -angles for search ellipsoid31 31 1 -size of covariance lookup table2 0.774 1.0 -ktype: 0=SK,1=OK,2=LVM,3=EXDR,4=COLC1 0.1 1.0 0.02 -self healing (0=no, 1=yes)../data/ydata.dat - file with secondary variable(LVM,EXDR,COLC variable4 - column for secondary variable1 - transform secondary variable (0=no,1=yes)1 0.3 -nst, nugget effect1 0.7 0.0 0.0 0.0 -it,cc,ang1,ang2,ang312.0 12.0 12.0 -a_hmax, a_hmin, a_vertaSteps:1.In GSLib..tools..Create_Vset_from_GSLib,2.Enter the file cluster.dat as the GSLib file.3.make sure the check boxes are unchecked because cluster.dat does not have columns for well id or Z4.In GSLib..tools..Create_Vset_from_GSLib,5.Enter the file ydata.dat as the GSLib file.MAKE sure the name is not the same as above!!6.un-check check boxes because ydata.dat does not have columns for well id or Z7.create an SGrid..New..From_Step_Vectors.put orgin at0,0,-0.5to encapsulate the data,alsoset nu=51,nv=51,nw=2(these are the number of nodes)and leave the property_cell_centered8.Run the GSLib..Algorithms..sgsim_sh picking“Primary”as the pointset_vr_property,and“Secondary”as the sec_property.Also,pick the secondary pointset and its“Secondary”propertyand proper SGrid.9.DELETE the OLD properties if running again to avoid the Gocad database from rejecting theadditional properties with identical namessurfsim:Using this parameter file:START OF PARAMETERS:1 -number of realizationscond1.dat -input file of conditioning datacond1_.dat -output scaled conditioning dataW2000C1idx.out -output file of the indexW2000C1thk.out -output file of the thicknessW2000C1par.out -output of the parameters1500 0.5 1.0 -nx,xmn,xsiz20 0.5 1.0 -ny,ymn,ysiz100 0.5 1.0 -nz,zmn,zsiz100 -Maximum no. of surface200 -total thickness69069 -seed of random number3 -surface type (1-Helmet 2-Ellipse 3-gauss) 15.0, 35.0, 60.0 -L,M,U height of a surface100.0,150.,200.0 -L,M,U length (Simple), long axis (Ellipse), SigmaX (Gauss)20.0, 25., 30.0 -L,M,U width (Simple), short axis (Ellipse), SigmaY (Gauss)5.0, 7.5,10.0 -L,M,U width 2 (Simple) (not for Ellipse & Gauss)0.0,90.0,180.0 -L,M,U angle0 -truncation flag1000,1150,1300 -L,M,U of X0 if trunFlag = 010, 10, 10 -L,M,U of Y0 if trunFlag = 0-------------------------------------------0 -itrans0.00001, 0.00001, 0.00001 -L,M,U sigma of residual1 0.0001 -nst, nugget effect3 0.9999 0.0 0.0 0.0 -it,cc,ang1,ang2,ang330.0 30.0 1.0 -a_hmax, a_hmin, a_vertsteps:1.In GSLib..tools..Create_Vset_from_GSLib,2.Enter the file cond.dat as the GSLib file.3.make sure the check box for well id is unchecked because cond1.dat does not have a column forwell id4.create an SGrid from right clicking on SGrid..New..From_Step_ the SGrid andset the orgin to x=0.5,y=0.5,z=0.5,and set nu=1500,nv=20,and nw=1005.Run GSLib..Algorithms..surfsim using the default parameters。
GOCAD讲义
GOCAD地质建模Gocad是以工作流程为核心的一款地质建模软件，达到了半智能化建模的世界最高水平，具有功能强，界面友好，易学易用，并能在几乎所有平台上（Sun, SGI, PC-Linux, PC-Windows）运行的特点。
以下通过一个简单的例子，说明GOCAD地质建模的主要过程。
第一部分GOCAD的启动1、双击桌面上的Gocad启动图标，即可启动Gocad程序。
2、选择New project建立新工程，给工程起个名字，选择路径在文件夹Project中，点击保存，打开下图：3、选择Select All，点击OK即可。
4、弹出选择工程单位窗口如下。
此处选择平面上单位是英尺（Feet），深度单位也是英尺（Feet）；时间选择毫秒；深度方向选择向上为增大；z值是时间域或深度域，选择深度域（最终如下图）。
选择完后点击OK。
5、打开如下Gocad主窗口。
第二部分GOCAD数据加载一、地震数据加载以SEGY格式为例，加载三维地震数据选择File > Import objects >Seismic Data > SEG－Y 3D as Voxet。
在Data文件夹下选择文件“tornado.sgy”，点击OK即可加载，如下图。
二、井数据加载1、井位数据加载不同的数据格式有不同的加载方法。
以普通的文本格式为例，包括井名、x坐标、y坐标、补心海拔、井深等。
（1）加载方式是：File > Import objects > Well data > （path）Locations from column-based File ，在Data\Wells文件夹中选择文件“WellPaths”，接受默认，按next两次，打开下图：（2）对问题What information do you have for the path? 选择X-Y-TVDSS-MD。
下来的两个问题各选Feet和Use a Column。
GOCAD(外文资料) (14)
Class Project LibraryIndependent Import & Export of GoCAD Data ObjectsGCAP provides an independent C++class library(Software Development Toolkit – SDK) to read/write GoCAD data for creating new data bridges. Defining principle data structures from GoCAD data models with methods to access GoCAD compatible files, the library has an efficient memory model with large file scalability. Full source code is available if required with the benefit that no other 3rd party software are required. A suite of example programs, including a VTK-based visualizer is available to simplify integrating GCAP into your applications. Currently Geological/Geophysical software is faced with unsolved problems of data management and visualisation of large multi-file and multi-entity geologic models.The aims of GCAP are to offer a C++class defining principle data structures from GoCAD models,with methods to read and write this data structure to GoCAD compatible files, using “out of core” scalable techniques.Motivation:With the development of GCAP, one such problem is solved that exists in the data management and visualisation of geologic models.Such models have grown significantly in storage and logical complexity in recent years. The field of software applications available to geologists and geophysicists has not kept pace with current practices and needs,especially as the minority of software applications that can address this complexity are in themselves very complex and difficult to use. While some applications exist, none offer an easy to use interface to other existing commercial or internal research applications.Existing Practices:GoCAD software and tools define a proprietary file transport standard for complex geologic models. This is supported by the vendor's native library API. This limits the scope of potential users to those organisations participating with GoCAD tools.Most 3rd party applications surveyed appear to only support a limited subset of GoCAD specifications, and have only written the portion of the data model support they were interested in. Many smaller groups have not been able to take up the full GoCAD system, creating an opportunity to provide an alternative with different design goals to the geological and geophysicalcommunity.Aims:GCAP seeks to achieve a significant advance in the data management and data visualisation of complex geologic 3D models, especially where models are defined using GoCAD format specifications. By integrating GCAP, full access to GoCAD hierarchical and multi-segment models can be manipulated at will from within your application, with the details of “out-of-core” scalability hidden from the developer.What is GCAP ?GCAP stands for G o C AD A ccess Class P roject. It is a library of C++ classes that can be used to read, store and write GoCAD data objects.The GCAP library consists of three main modules:1.GoCAD Data Object Module2.GoCAD File Reader Module3.GoCAD File Writer Module GCAP currently supports the following object types:1.Atomic or VSet2.TSurf3.TSolid4.Grid3D or Voxet5.GSurf6.SGridA full 50-page user manual and 200-page API Reference describes how to install and use GCAP in your application.Features:•Independent C++ Class Library•Scalability using "out-of-core" methods•Dictionary system allows application to pick and choose parts•Handles attributes and properties, not just raw geometry•Comprehensive professional documentation, hyperlinked help•Both Linux and Windows supported platformsContact InformationVisual Technology Services Ltd.The Courtyard, High Street, Ascot,Berkshire SL5 7HP, UKPhone: +44 (0) 7787 517529Email: info@[ Ver. 2, 02/11/2006 ]。
GOCAD讲义
GOCAD地质建模Gocad是以工作流程为核心的一款地质建模软件，达到了半智能化建模的世界最高水平，具有功能强，界面友好，易学易用，并能在几乎所有平台上（Sun, SGI, PC-Linux, PC-Windows）运行的特点。
以下通过一个简单的例子，说明GOCAD地质建模的主要过程。
第一部分GOCAD的启动1、双击桌面上的Gocad启动图标，即可启动Gocad程序。
2、选择New project建立新工程，给工程起个名字，选择路径在文件夹Project中，点击保存，打开下图：3、选择Select All，点击OK即可。
4、弹出选择工程单位窗口如下。
此处选择平面上单位是英尺（Feet），深度单位也是英尺（Feet）；时间选择毫秒；深度方向选择向上为增大；z值是时间域或深度域，选择深度域（最终如下图）。
选择完后点击OK。
5、打开如下Gocad主窗口。
第二部分GOCAD数据加载一、地震数据加载以SEGY格式为例，加载三维地震数据选择File > Import objects >Seismic Data > SEG－Y 3D as Voxet。
在Data文件夹下选择文件“tornado.sgy”，点击OK即可加载，如下图。
二、井数据加载1、井位数据加载不同的数据格式有不同的加载方法。
以普通的文本格式为例，包括井名、x坐标、y坐标、补心海拔、井深等。
（1）加载方式是：File > Import objects > Well data > （path）Locations from column-based File ，在Data\Wells文件夹中选择文件“WellPaths”，接受默认，按next两次，打开下图：（2）对问题What information do you have for the path? 选择X-Y-TVDSS-MD。
下来的两个问题各选Feet和Use a Column。
GOCAD(外文资料) (10)
Stanford Exploration Project,Report80,May15,2001,pages1–616Iterative velocity model building for3-D depth migration byintegrating GOCAD and A VSRobert G.Clapp and Biondo Biondi1ABSTRACTWe have developed a procedure for building a3-D velocity model starting from a2-D geological model in order to image3-D poststack data.The3-D model was built using GOCAD,and it was iteratively refined by interpreting the result of3-D poststack depth migrations using A VS.At the beginning of the iterative process,when only few reflectors are detectable in the migrated cube,we interpret only one surface.From the deformations measured on the selected reflector we compute a displacementfield for every point in the model.The application of the displacementfield to the GOCAD model used for migration leads to an improved model.The migration of the data with this improved model shows better focusing of some reflectors and allows a more complete interpretation of the structure.INTRODUCTIONOne of the most crucial problems in imaging3-D seismic data is the determination of the velocity model.With the ever increasing speed of parallel supercomputers new options for defining and refining the velocity model are becoming possible.These new options take ad-vantage of the speed of these super computers to integrate time consuming operations such as migration,into an interactive and iterative processingflow(Wyatt et al.,1992;Jones,1993). Sheppard and O’Brien(1993)took advantage of these new options by creating a3-D veloc-ity model by performing depth migration,interpreting the resulting image,and refining the geologic model.In addition,improved visualization software has made3-D model building, viewing,and manipulating an easier task.Della Malva and Williams(1992)used GOCAD to combine a series of2-D velocity models into a3-D model.We blended elements of both of these methods in an attempt to obtain a good3-D velocity model for3-D poststack migration. We started with a2-D geologic model,along with the prestack and poststack3-D data sets in which initially very few reflectors were identifiable.We migrated the data with2.5-D model created in GOCAD.We then viewed the3-D cube using A VS and picked a series of surfaces on various slices of the cube parallel to the direction of the initial2-D geologic model.From this series of surfaces we created a2-D displacement net for the cube.Choosing a single surface and then extrapolating its changes in depth to the entire geologic model does not accurately 1email:not available12Clapp and Biondi SEP-80 describe geologic sedimentation and the time variant nature of the stressfield accompanying it.The methodology is not a badfirst approximation,and proved effective in improving the migrated image and the accompanying geologic model.BUILDING THE3-D GEOLOGIC MODELThefirst challenge was to convert the2-D AIMS geologic model into a3-D GOCAD model. AIMS modelAIMS creates its model by constructing a series of surfaces defined by numerous line seg-ments.These line segments are defined in an ASCIIfile by a pair of coordinates representing the beginning and end of the line segment.In addition to the coordinates,each line contains the number of the corresponding surface.The surfaces are chosen so that only one velocity is represented below them and that velocity is defined in a horizon call.An example of a portion of an AIMS2-D modelfile is presented below:HORI40,,4600.0BEG40,2310.374,1108.975END40,2331.748,1058.843BEG40,2331.748,1058.843END40,2355.868,1000.688BEG40,2355.868,1000.688END40,2378.594,956.565BEG40,2378.594,956.565END40,2499.702,889.8913-D Model in GOCADAs described in Berlioux(?)a GOCAD3-D model is defined by a series of surfacesfiles and a partitionfile.Each surfacefile is divided into two parts.Thefirst part contains a series of points(vertices in the GOCAD nomenclature)upon which the surface lies.The second section is devoted to constructing a series of triangles from the defined points in order to give the surface continuity in two-dimensions.Figure1illustrates a simple model constructed from the surfacefile defined below:SEP-80Velocity model building3Figure1:A simple GOCAD model.bob-gocad1[NR]VRTX10.000.000.00VRTX2 2.000.000.00VRTX3 2.00 2.000.00VRTX40.00 2.000.00VRTX5 1.00 1.00 1.00TRGL125TRGL235TRGL345TRGL415The partitionfile defines a series of domains bounded by the various surfaces.The domains can then be assigned properties such as velocity or density.The ConversionRather than converting the AIMSfile manually into a series of GOCADfiles it was decided to develop a procedure that would accomplish this task automatically.The conversion program was written in standard SEP-like format to better allow it to be integrated into future processing flows.The input is simply the AIMSfile,parameters defining the extension in the cross-line direction,and an option to create a free surface.The output is a series of.tsfiles(GOCAD extension for a surfacefile)corresponding to the surfaces described by AIMS.Once the initial AIMSfile has been run through the conversion routine it is a relatively simple matter to create the GOCAD model.Figure2represents the GOCAD model corre-sponding to the input AIMS geologic model.In order to run the CM-5migration program it is necessary to convert the GOCAD model to SEP format.Figure3represents a2-D slice of the resulting velocity model using the program described in Berlioux(1993).4Clapp and Biondi SEP-80Figure2:GOCAD model corresponding to input AIMS model.bob-gocad2[NR] ../bob/./Figs/gocad3.pdfFigure3:Gridding output based on initial AIMS model.SEP-80Velocity model building5REFINING THE GEOLOGIC MODELOnce the initial geologic model had been built,the task was to better describe the3-D nature of the area.The velocity model was used as input to a depth migration program on the CM-5 developed at SEP(Biondi and Palacharla,1993).The resulting cube was then viewed using A VS.A VS pickerDave Nichols has created an A VS network,and written a module for that network which enables the user to pick a polyline on a2-D image and write that polyline in a GOCAD pline format.In order to create the displacementfield,we chose a2-D slice of the data cube parallel to the geologic model(see Figure4),then chose a surface that was possible to identify on all/most of the2-D images.We picked the surface on the image and saved it in afile corresponding to the slice number.Part of the recorded plinefile for Figure4can be seen below:Figure4:Slice324of3-D migrated cube.bob-avs3241[NR]VRTX132.00000058.0000000.000000VRTX242.00000054.0000000.000000VRTX344.00000053.0000000.000000VRTX448.00000055.0000000.0000006Clapp and Biondi SEP-80VRTX551.00000058.0000000.000000VRTX656.00000060.0000000.000000VRTX759.00000063.0000000.000000VRTX862.00000065.0000000.000000VRTX965.00000068.0000000.000000VRTX1067.00000070.0000000.000000VRTX1170.00000072.0000000.000000...SEG12SEG23SEG34SEG45SEG56SEG67SEG78SEG89SEG910SEG1011...We then attempted to pick the same pline on another in-line slice in order to approximately quantify the changes in depth along the cross-line direction of the geologic model.Figure5 shows a second slice in the in-line direction,note the change in the surface approximately80% of the way down the image in comparison to the same surface on Figure4.Creating GOCAD model from series of plinesCreating the3-D GOCAD model from the series of2-D plines was a two-step process.The first step was to create a2-D displacementfield.The displacementfield was created by:first, reading in the series of vertices making up the various plines,then resampling in the in-line direction using linear interpolation between the points,andfinally,subtracting the resampled values from a baseline corresponding to the original2-D geologic model.The corresponding displacement net is specified by a user defined density parameter in the in-line direction and by the number of slices of the3-D cube in which plines were chosen in the cross-line direction. It was not necessary to resample in the cross-line direction due to the combination of GOCAD and the gridding program doing the same linear interpolation as the program performs in the in-line direction.Once the2-D displacement net has been created,Figure6,the net is used to construct the same series,and in virtually the same manner,the.tsfiles constructed in the original2.5-D GOCAD model.Figure7represents the GOCAD model derived from the displacementfield shown in Figure6.Even though the initial data quality was poor,and the re-migrated image far fro m ideal (Figure8),the image does appear to be improved compared to the corresponding slice mi-grated with the initial2.5-D velocity model,Figure4.The improvement is especially notice-able in the middle of the image where an anticline has become better focused.SEP-80Velocity model building7Figure5:Slice220of3-D migrated cube.bob-avs220[NR]../bob/./Figs/displacement1.pdfFigure6:2-D displacement net.8Clapp and Biondi SEP-80Figure7:GOCAD model constructed from initial displacementfield calculations. bob-gocad4[NR]Figure8:Slice324of migrated cube with3-D velocity model.bob-avs3242[NR]SEP-80Velocity model building9CONCLUSIONSBy creating a2-D displacement net and applying it to the geologic model,the3-D poststack migration image did show some improvement.The velocity model could be further improved by re-choosing the surface on the series of2-D slices.The improved quality of the velocity function should enable a more consistent and improved picking,further enhancing the quality of the3-D model.Further,once the surfaces become more apparent in the migrated image,ex-tending the displacementfield into the third dimension by picking two or more surfaces might help improve the results and better approximate the true nature of sedimentation variation with time.In addition velocity inside the model could be better defined by intergrating geophysical data derived from such source as velocity analysis.ACKNOWLEDGEMENTSWe would like to thank Husky Oil for providing the dataset,and in particular Fred Kierulf with Husky Oil,for his enthusiastic support.At SEP,we are indebted with Arnaud Berlioux for the numerous hours spent explaining GOCAD and his gridding program,and Dave Nichols for explaining some of the intricacies of A VS and his network.REFERENCESBerlioux,A.,1993,3-D grid with GOCAD:SEP–79,301–318.Biondi,B.,and Palacharla,G.,1993,3-D wavefield depth extrapolation by rotated McClellan filters:SEP–77,27–36.Jones,I.F.,1993,3-D velocity model building via iterative one-pass depth migration:63rd Annual Internat.Mtg.,Soc.Expl.Geophys.,Expanded Abstracts,974–977.Malva,R.D.,and Williams,M.C.,1992,3-D model construction of overthrust geology for structural imaging of seismic by migration:62nd Annual Internat.Mtg.,Soc.Expl.Geo-phys.,Expanded Abstracts,1255–1256.Sheppard,F.,and O’Brien,M.,1993,Iterative3-D post stack depth migration:An interpretive seismic imaging tool:3-D Seismology:Integrated Comprehension of Large Data V olumes, SEG Summer Research Workshop,Technical Program and Abstracts,124–.Wyatt,K.D.,Towe,S.K.,Layton,J.E.,Wyatt,S.B.,von Seggern,D.H.,and Brockmeier, C.A.,1992,Ergonomics in3-D depth migration:62rd Annual Internat.Mtg.,Soc.Expl. Geophys.,Expanded Abstracts,944–947.。
GOCAD(外文资料) (28)
Open Spirit User Meeting1, Paris June 2004.Figure 1 OS 3D Viewer displaying data from OpenWorks, SeisWorks and Finder2.SummaryAbout 40 attended the EU Open Spirit (OS) user meeting held just after the annual meeting of the EAGE in Paris. Open Spirit Corp (OSC) appears to be gaining traction for its E&P middleware (see our section ‘What is Open Spirit’ below). This has been by dint of a constant marketing and technical effort over the last several years. The list of third party application vendors using the OS framework is getting longer. One independent vendor (SMT) has even made the bold step of offering an OS data sever. A recent OS innovation is the qualification program – so far only Hampson-Russell (Veritas) has signed up for this option3.OS is a developer’s tool above all else. End users should ideally be unaware of OS’s existence. Some larger oil companies develop their own software, for instance Shell’s 123DI and Total’s Sismage. For these tools, OS offers a convenient way to plug in-house developments into vendor data stores. Larger oils, especially those in the Schlumberger camp, may be using OS technology (such as CopySync) in their IT support and data management. Elsewhere, OS end users are found in third party software vendors such as EDS and SMT who use OS to link GoCad and the Kingdom Suite to vendor data stores.Schlumberger Information Solutions’ (SIS) position regarding OS is clear. They give strong support for OS – and promise integration within existing and new product lines. The only nuance here is that the support is stronger if the OS integration is performed and supported by SIS!Landmark’s presence (Murray Roth and Janet Hicks) was significant – after all, OS is a Schlumberger affiliate. Landmark is working on two pilots using OS – one an internal test of1 Visit the OpenSpirit website for more info and presentations.2 Image courtesy Open Spirit Corp.3 Announcements from SIS and EDS are expected ‘real soon now’.OS functions and performance, the other, a funded development project for a client. While Landmark’s evaluation of OS is still ‘work in progress’, the impression one got is of lukewarm support at best. Landmark offers enough encouragement to OS so as not to offend any key clients. Landmark was also significantly touting the merits of WITSML as an alternative, real-time focused, data sharing mechanism.As ever, there are many marketing decisions lurking behind the technology announcements. Is Schlumberger really moving to all .NET? Is Landmark moving to all Java? Will OS ever get beyond the position of a souped-up GeoFrame in the marketplace? What will likely determine OS’s role outside of Schlumberger is the attitude of the major oils. Will ChevronTexaco and Shell stick with OS? Will other majors come on board? Selling middleware is a tough call, but SAP and Tibco have both turned technology into a great business – with a little help from some of the same majors that seem to have cold feet when it comes to backing this E&P-specific middleware.HighlightsSchlumberger support strongLandmark’s position ‘explained’Shell and .NETSMT’s Open Spirit serverContentsOpenSpirit on the SIS Strategic Roadmap – Dominique Pajot, SIS (3)SIS Ocean/Coral/Seabed (3)OS role in Shell .NET migration – Winand Belmans, Shell (4)Landmark’s plans for OS – Janet Hicks, Landmark (5)Version 2.6 highlights – Clay Harter, Open Spirit Corp (6)An OS Server for SMT – Tom Smith, SMT (7)End user/developer presentations (7)Total’s use of Open Spirit (7)ChevronTexaco’s SeisVu seismic viewer (7)ChevronTexaco’s PWA reservoir model builder (7)Shell reports on growing use of Petrel (7)What is Open Spirit? (8)。
- 1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。
Aktueller Stand zur Einführung der 3D-Modellierungssoftware Go-cad im LfUG am Beispiel von Zwickau-SchedewitzHenry Steinborn, Ottmar Krentz, Manfred FelixSächsisches Landesamt für Umwelt und Geologie, Abteilung 5 – Referate 52, 55, Zur Wetterwarte 11, 01109 Dresden, Email: Henry.Steinborn@lfug.smul.sachsen.deDer seit über 600 Jahren betriebene Steinkohlenbergbau im Steinkohlenrevier Zwickau hatte massive Bergbauschäden zur Folge. Diese waren und sind vor allem mit den Senkungen der Tagesoberfläche verbunden. So treten in der Stadt Zwickau gegenwärtig Absenkungstrichter des Grundwasserspiegels im flurnächsten Grundwasserleiter auf.Auf der Basis einer intensiven Datenrecherche und -aufarbeitung bot sich im Steinkohlenrevier die Möglichkeit, ein 3D-Modell zur besseren räumlichen Erfassung und Analyse der geologisch-hydrogeologisch-tektonischen Situation zu erstellen und die 3D-Technologie des LfUG an Praxisdaten zu testen.In Zusammenarbeit mit der TU Bergakademie Freiberg, die umfangreiche Erfahrungen bei der 3D-Modellierung geologischer Situationen mit der 3D-Programm Gocad besitzt, läuft dazu im Zeitraum 2002 – 2005 ein gemeinsames Forschungsprojekt mit dem Titel …Angewandte Modellierung im Schwerpunktgebiet Steinkohlenbergbaureviere Zwickau und Lugau/Oelsnitz mittels Gocad“Der Aufbau der 3D-Technologie am Landesamt für Umwelt und Geologie (LfUG) erfolgt in Abstim-mung mit den Geologischen Diensten anderer Bundesländer und wird im Vortrag vorgestellt.1 EinleitungVon 1348 bis 1978 wurde im Gebiet der Stadt Zwickau Steinkohlenbergbau betrieben. Insge-samt standen in Zwickau bis zu 11 Flöze, aus denen ca. 220 Mio. t Steinkohle gewonnen wur-den, im Abbau. Vor allem ab der Zweiten Hälfte des 19 Jahrhunderts wurde der Abbau bedingt durch den technischen Fortschritt deutlich inten-siviert, wodurch in Zwickau massive Bergbau-folgen, vor allen in Form von Senkungen der Tagesoberfläche, auftraten.Im Zuge des Bergbaus und den damit verbunden intensiven Erkundungsmaßnahmen wurde fest-gestellt, dass Zwickau regionalgeologisch im Kreuzungsbereich der Gera-Jachymov-Störungs-zone mit mehreren erzgebirgisch streichenden Störungselementen liegt.Heute hat die Stadt Zwickau in den Senkungsge-bieten mit Grundwasserproblemen zu kämpfen, so dass im Senkungsgebiet Zwickau-Bockwa ein dauerhaftes Pumpen zur Verhinderung der Über-flutung der Tagesoberfläche notwendig ist. Au-ßerdem existieren Absenkungstrichter des Grundwasserspiegels im flurnächsten Grundwas-serleiter im Raum Zwickau-Schedewitz und Zwickau-Innenstadt. Die Ursachen, welche die Wirkmechanismen und das Langzeitverhalten der Trichter bestimmen, waren strittig. Als einenmöglichen Wirkmechanismus für die Trichter wurde der Einfluss der Oberhohndorfer Haupt-verwerfung als Teil der Gera-Jachymov-Störungszone diskutiert. Eine weitere Ursache wurde in den Folgen des Bergbaus, und speziellin den mit abgesenkten Abwassersystemen gese-hen. Zur Klärung des Sachverhaltes wurde Zwi-ckau zum geologisch-hydrogeologischen Schwerpunktgebiet im LfUG erklärt.Infolge dessen erfolgte ein intensive Recherche, Aufarbeitung und Neuerfassung aller Daten im Bergbaurevier. Diese Daten bildeten die Basisfür ein geologisch-hydrogeologisch-tektonisches3D-Modell.2 Konzeption zur Einführungvon GocadDa das LfUG bisher noch keine Erfahrungen mit3D-Gocad-Modellierungen besaß, wurde für den Zeitraum 2002 bis 2005 ein Forschungsprojekt mit dem Thema …Angewandte Modellierung im Schwerpunktgebiet Steinkohlenbergbaureviere Zwickau und Lugau/Oelsnitz mittels Gocad“ zwischen der TU Bergakademie Freiberg und dem LfUG iniziiert. Die TUBAF arbeitet seit mehreren Jahren erfolgreich mit der Software Gocad und besitzt darüber hinaus die notwendi-gen Fachkenntnisse zur Geologie, Hydrogeologie und Tektonik im Raum Zwickau.Das Forschungsvorhaben untergliedert sich in insgesamt 3 Teilprojekte.Der Teil 1 des Projektes mit dem Thema …Wis-senschaftliche Konzipierung der angewandten Softwareerweiterung/-anpassung von Gocad für das LfUG“ konzentrierte sich auf die Entwick-lung und Anpassung verschiedener Filter und bestehender Programme zur Nutzung der Daten-bestände des LfUG.Die Abbildung 1 zeigt das Schema zur geplanten Arbeitstechnologie vor Beginn des Forschungs-projektes. Dabei beinhaltet die Datenebene alle im LfUG verfügbaren Datenbestände in Form von Bohrdaten (ORACLE 9), 2-dimensionale Profilschnitte (GeoDin 3.x) sowie Karten und Rasterdaten (ArcInfo 8). Die Importfilterebene wurde im Rahmen des Teilprojektes 1 neu ent-wickelt. Die Bearbeitungsebene bildet die 3D-Software Gocad. Den Abschluss des Schemas bildet die Darstellungsebene. In ihr sind die Möglichkeiten zur unabhängigen Visualisierung (3D-Viewer) sowie die Entwicklung zum 3D GIS enthalten.Teil 2 des Projektes läuft seit 04/03 bis Ende 2005 als Eigenforschungsstelle am LfUG zum Thema …Tektonik und Strukturbau des Permo-karbons Zwickau/Oelsnitz“. Ziel dieses Teilpro-jektes ist es, die entwickelten Filter und Pro-gramme an konkreten Beispielen zu testen und im Ergebnis ein bzw. mehrere unterschiedlich detaillierte geologisch-tektonische Modelle des Schwerpunktgebietes zu entwickeln. Im Ergebnis sollen aus diesen Modellen aktuelle Aussagen zur Tektonik und zum Strukturbau in der Region Zwickau und Lugau/Oelsnitz als Basis für die sachgerechte Bewertung der Bergbaufolgen ge-troffen werden.Parallel zum Teilprojekt 2 läuft an der TUBAF der Teil 3 zum Thema …Hydrogeologische Aus-sagemöglichkeiten aus dem Strukturmodell“. Bei diesem Thema soll vor allem die Anwendung von 3D-Modellen im Vordergrund stehen. Hier ist das Ziel möglichst belastbare Aussagen zur Hydrogeologie treffen zu können.3 Stand zur Einführung vonGocadAnfang 2004 standen dem LfUG alle im Teilpro-jekt 1 neu entwickelten Funktionalitäten und Konzeptionen zur Verfügung. Da das LfUG An-fang 2004 eine eigene Gocad Lizenz erworben hat, finden z. Z. intensive Tests der entwickelten Technologie statt.ORACLEDatenbankGeoDin ArcInfoIm Folgenden sollen die im Teilprojekt 1 entwi-ckelten Funktionalitäten und Konzeptionen an-hand von Auszügen aus dem Abschlussbericht "Wissenschaftliche Konzipierung der angewand-ten Softwareerweiterung/ -anpassung von Gocad für das Sächsische Landesamt für Umwelt und Geologie" vom 15.12.2003 (M. APEL; I. RUPF;H. STEINBORN) in einem kurzen Überblick dargestellt werden. Dabei gibt die Abbildung 2 den bisher erreichten Arbeitsstand wieder. Der dunkel hinterlegte Bereich ist noch nicht reali-siert.3.1 Schnittstelle ORACLE – Go-cadAufschlussdaten werden am LfUG gegenwärtig in einer von der RDS GmbH Dresden entwickel-ten Oracle-Datenbank gespeichert. Am Anfang des Projektes wurde der Wunsch klar, Interpreta-tionen/Ableitungen von den primären Schichten-verzeichnissen für 3D-Modellierungen zu nut-zen. Daher entwickelt die RDS GmbH z .Z. ein Konzept zur Speicherung von Interpretationspro-filen in einer Oracle Datenbank.Der Prototyp der Schnittstelle ermöglicht den Import von Bohrungsdaten aus einer Oracle-Beispieldatenbank direkt aus Gocad via Oracle Call Interface (OCI) über eine TCP/IP Verbin-dung. Die Schnittstelle wurde betriebssystemu-nabhängig in C++ programmiert. Die gegenwär-tige Schnittstelle ermöglicht es, sowohl mit ei-nem nutzerfreundlichen Menü als auch per SQL Bohrungen mit bestimmten Eigenschaften aus-zuwählen und in Gocad die entsprechenden Boh-rungen mit Markern und numerischen Eigen-schaften zu erzeugen. Das Auswahlmenü erlaubtdie Aufschlussauswahl nach den Kriterien: "Pro-Abb. 2: Aktueller Arbeitsstand bei der Einführung von Gocad im LfUG.jekt", "Datum", "Lage" (GK-Koordinaten und minimale Teufe) sowie als variable SQL-Abfrage, wobei diese auf jede der zu einem In-terpretationsprofil gehörenden Tabellen ange-wendet werden kann und das Programm zugehö-rige Daten anderer Tabellen ermittelt. Die erstell-ten Gocad-Well-Objekte können wahlweise di-rekt eingelesen oder in einer Datei gespeichert werden.Bei der Konversion wird der oberste Punkt (im Allgemeinen die Hangend-Schichtgrenze) einer geologischen Einheit einem Well-Marker zuge-ordnet. Zwischen zwei Markern wird eine Zone deklariert, die der Schnittlinie der Bohrung mit der geologischen Einheit entspricht. Dieser Zone wird als Name das Datenbankfeld "STRAT" zugewiesen. Die Felder "GENESE", und "GEF" werden numerisch kodiert und ebenso wie die Felder "PET" und "PROFIL_PARAM[1...5]" den Stützpunkten des Bohrpfades in den jeweiligen Teufen zugewiesen.Die Kodierung wird in einer erweiterbaren Ta-belle in einer ASCII-Datei "konv.dat" im Pro-grammverzeichnis gespeichert und zur Laufzeit vom Programm eingelesen. Mit Hilfe der Tabelle kann die Kodierung vom Nutzer leicht überprüft und geändert werden. Ein Ergänzen der Tabelle um weitere zu konvertierende Zeichenkettenei-genschaften ist möglich.3.2 Schnittstelle GeoDin – Gocad Die Firma Fugro GmbH wurde beauftragt, die Software GeoDin um folgende Funktionalitäten zu erweitern:•Implementation der Attributierung von Li-nien und Flächen als Pilotlösung für die Software GeoDin. Dabei werden für Flä-chenobjekte 4 Felder vorgesehen:o PET (Typ char[36]),o STRAT (Typ char[16]),o GENESE (Typ char[3]),o VAR(Typ char[16]).•Für Linienattribute sind folgende Felder erforderlich:o LART (Typ num[4]),o LORIENT (Typ char[1]),o LGENAU (Typ char[2]).•Konzept für die Validierung der Attribute mit Prüftabellen •Export der Attribute in ASCII-Datei, von der Ungenerate-Datei mit ID referenziert.Das Austauschformat ist wie folgt definiert: zu jedem Objekttyp (Linie, Polygon) werden je eine Geometrie- und eine Attributdatei erzeugt. Be-gonnen wird ein Objekt jeweils mit der ID. Jede ID wird unabhängig vom Objekttyp nur einmal vergeben. "END" schließt jeweils einen Geomet-rieobjekt oder Attributobjekt ab. Zusätzlich exis-tiert eine Datei, welche die Schnittlinie des Pro-fils mit der Geländeoberfläche enthält. Es ist beabsichtigt, in einer zukünftigen Version noch die Attribut-Datentypdefinition zu exportieren. Der als Gocad-Plugin in C++ entwickelte Im-port-Filter ermöglicht die Kreation von Linien (Polylinien und Polygonen) in Gocad. Dabei wird pro GeoDin-Linie ein PLine-Objekt in Go-cad erstellt, das beliebige numerische Attribute enthalten kann. Gegenwärtig ist es in Gocad nicht möglich, Zeichenketten als Attribute für Objekte zu verwenden. Ausnahmen bilden hier die "Features" = PET / LART, "Age"=STRAT sowie der Objektname.Als Problem stellte sich heraus, dass das Daten-modell von GeoDin keine Topologie beinhaltet. Damit sind in GeoDin keine Beziehungen zwi-schen den geologischen Objekten definiert. Dies ist jedoch erforderlich, um die GeoDin-Schnitte für die Erstellung sinnvoller 3D-Modelle zu nut-zen (beispielsweise kann es passieren, dass ein GeoDin-Schnitt einer Störungsfläche mit mehr-fachen geometrisch identischen Linien abgebil-det wird, was unzulässig ist).Es wurde erwogen, das Problem in Arc/Info zu lösen, indem dort eine Topologie generiert wird. Das stieß jedoch auf erhebliche Probleme, da keine geeignete Möglichkeit zum Datenimport nach Arc/Info besteht, die mehrfache Z-Werte pro XY-Koordinate erlaubt. Das Erstellen topo-logischer Beziehungen zwischen den Objekten muss folglich in Gocad geschehen. Dazu wurde eine Gocad-Funktion entwickelt, die das Filtern von Liniensegmenten verschiedener Kurven mit geometrisch gleicher Lage erlaubt. Dazu werden die Polygone in einzelne Linien zerlegt, die je-weils bis an das nächste Objekt grenzen.In Gocad werden einzelne Schichtgrenzen mo-delliert, wobei eine Grenzfläche gleichzeitig die Hangendgrenze der liegenden Schicht und Lieg-endgrenze der hangenden Schicht bildet. Dieses Konzept ist grundlegend anders als bei GeoDin, wo mit Polygonen die Schnittlinie der Schicht-grenze mit der Profilfläche abgebildet wird. PerStandardeinstellung werden die GeoDin-Attribute des Hangendpolygons der erhalten bleibenden Grenzlinie zugewiesen, die restlichen Linien werden gelöscht. Wenn Polygonlinien von Störungslinien überlagert werden, werden die Polygonlinien gelöscht. Damit bleiben Attri-butinformationen erhalten und das Erstellen to-pologischer Objektbeziehungen wird möglich. Dieser Filter wurde als Konsolenanwendung in C++ programmiert, die direkt von Gocad aufge-rufen werden kann (Menü "File - Import Objects - GeoDin"). Damit ist der GeoDin-Import tech-nisch möglich, jedoch ist der Nachbearbeitungs-aufwand aufgrund des unterschiedlichen Daten-modells sehr hoch. Daher ist eine Arbeitstechno-logie, die routinemäßigen und umfangreichen Datenaustausch zwischen GeoDin und Gocad beinhaltet, nicht zu empfehlen.Der Filter konnte im LfUG erfolgreich getestet werden. Dabei stellte sich heraus, dass es beim Export von GeoDin zu Problemen mit dem Pro-jektionsystem kommen kann. Weiterhin zeigte sich, dass die entstehenden Gocad Objekte zwar alle Objektlinien beinhalten, diese aber nochmals miteinander verbunden sind. Dadurch ist eine Zuordnung, welche Linie eine Profillinie ist, deutlich erschwert. Diese Probleme werden z. Zt. gemeinsam mit der TU BAF beseitigt.3.3 Schnittstelle ArcInfo – Gocad Für den Export von ArcInfo zu Gocad wurde von I. Rupf das Programm E002Gocad entwickelt. Das Programm konvertiert geometrische Objekte und dazugehörige numerische Attribute unkomp-rimierter ArcInfo-E00-Dateien (Version 8.0) in das Gocad-Ascii-Format.3.3.1 Geometrie-ImportJede E00-Datei beinhaltet Daten eines bestimm-ten Geometrietyps (= ArcInfo-Coverage-Typ). Das Programm verarbeitet ArcInfo-Punkt-, Poly-linien- und Polygoncoverages. Da die Art des Coverage s im Normalfall nicht beim Import be-kannt ist, wird der Geometrietyp im Laufe des Konvertierungsprozesses bestimmt. Handelt es sich dagegen um eine Datei des Sächsischen Landesamtes für Umwelt und Geologie des digi-talen geologischen Kartenwerkes 1:25 000 bzw. 1:50 000, ist der Coverage-Typ bereits an der Benennung erkennbar:•*_p.e00 (ArcInfo-Punktcoverage)•*_l.e00, *_ll.e00, *_lin.e00 (ArcInfo-Polyliniencoverage) •*_fl.e00 (ArcInfo-Polygoncoverage). Sind derartige Bezeichnungen vorhanden, kön-nen die entsprechenden Daten gezielt konvertiert werden. Ansonsten müssen zunächst alle verfüg-baren Informationen importiert werden. Letztere Variante ist weitaus zeitintensiver.Das Programm E002Gocad ermöglicht die Kon-vertierung einfacher und zusammengesetzter Objektgeometrien. Zu den einfachen Objektge-ometrien gehören Punkte, Polylinien und Poly-gone. Aus ihnen setzen sich die komplizierteren Gruppengeometrien, Polyliniengruppe und Poly-gongruppe bzw. Polygonsubgruppe zusammen. Da in einem E00 – File nur Hoch- und Rechts-werte die Lage der Objekte beschreiben, werden beim Import in Gocad auch nur die Hoch- und Rechtswertkoordinaten übertragen. Die Höhen-komponente, welche im 3-dimensionalen Raum notwendig ist wird automatisch auf 0 m ü NN gesetzt. Koordinaten in E00-Files sind entweder mit einfacher oder mit doppelter Genauigkeit (8 bzw. 15 stellige Zahlen) abgelegt. Weil Gocad jedoch ausschließlich einfach genaue Koordina-ten unterstützt, müssen entsprechende Angaben mit doppelter Genauigkeit umgewandelt werden. Dabei werden die letzten Ziffern abgeschnitten. Entsprechende Detailinformationen können bei diesem Vorgang verloren gehen. Deshalb ist vor der Konvertierung zu prüfen, ob eventuell die ArcInfo-Koordinaten einer Transformation be-dürfen (beispielsweise durch Abschneiden der ersten Ziffern der Koordinaten). Dies ist dann vor der E002Gocad-Konvertierung durchzufüh-ren.Folgende Geometrie-Abschnitte des E00-Files enthalten keine relevanten Informationen und werden nicht in den Konvertierungsprozess mit einbezogen:•SIN (funktionsloser Abschnitt)•LNK (Abschnitt mit unklarer Bedeutung) •LOG (Angaben über die Bearbeitungshisto-rie)•PAR (Abschnitt mit unklarer Bedeutung) •PRJ bzw. MTD (Abschnitt mit Metainforma-tionen, wie Koordinatensystem, Bearbeiter etc.)•TXT, TX6, TX7,... (Annotation - Text an Geometrieobjekten)•TOL (Angaben zu Toleranzen bei der Erstel-lung des Coverages)3.3.2 Attribut-ImportDer Attributimport ist derzeit nur bei den ArcIn-fo-Grundgeometrien möglich. Da Gocad aus-schließlich numerische Angaben verarbeitet, ist der Import von Zeichenkettenattributen in das Konvertierungstool nicht sinnvoll (Ausnahme: numerische Kodierung LfUG-spezifischer Zei-chenkettenattribute). Folgende numerische Da-tentypen können dagegen problemlos importiert werden:•I (integer), B (binary integer), N (number) • F (binary floating point).Die Konvertierung von Standardattributen (Att-ribute, die routinemäßig von ArcInfo automa-tisch angelegt werden) ist in der Regel wenig sinnvoll. Attribute dieser Art tragen entweder nichtrelevante ArcInfo-spezifische Eigenschaften (z.B.: System-ID's) oder sie werden in Gocad neu berechnet (z. B.: Fläche, Durchmesser etc.). Eine Ausnahme bilden die ArcInfo-Nutzer-ID's. Des Weiteren werden die landesamtsspezifischen Zeichenkettenattribute LGENAU und LORIENT numerisch kodiert und können somit in Gocad eingelesen werden.Treten entsprechende Attribute auf, wird in der Logdatei der Kodierungsschlüssel abgelegt. Derzeit können keine weiteren Attribute bzw. neue Attributuntereinträge während der Laufzeit eingelesen werdenFolgende Geometrie-Abschnitte des E00-Files enthalten keine relevanten Informationen und werden nicht in den Konvertierungsprozess mit einbezogen:•.ACODE, .PCODE (Arc Lookup Table, Polygon Lookup Table)•.BND (Angaben zur Grenzbox der Objekte des E00-Files)•.TIC (Informationen zur Lage der TIC-Koordinaten des Coverages)Eine ausführliche Beschreibung des Programms E002Gocad ist im Abschlussbericht enthalten. 3.4 3D-Visualisierungsprogramm Die Entwicklung eines Gocad-unabhängigen Visualisierungsprogramms "GeoShow" erfolgte auf der Grundlage des standardisierten internet-fähigen 3D-Formats VRML. Das Visualisie-rungsprogramm ermöglicht das Betrachten von georeferenzierten Gocad-3D-Modellen im VRML-Format.3.5 Vorliegende KonzeptionenIm Ergebnis der bisherigen Arbeiten liegen dem LfUG zwei Konzeptionen vor. Die eine Konzep-tion …Verknüpfung von 3D Modellen mit ver-schiedener räumlicher Auflösung“ zeigt Mög-lichkeiten auf zwei 3D-Modelle mit unterschied-licher räumlichen Auflösung miteinander zu verbinden. Die Grundlage dafür sind zwei Go-cad-Flächenmodelle, die eine Bounding Box -Fläche besitzen. Das detaillierte Modell ("De-tailmodell") muss im größeren Modell ("Regio-nalmodell") mindestens teilweise räumlich ent-halten sein. Es ist auch möglich, anstatt des Re-gionalmodells zwei sich wenigstens teilweise räumlich überlappende Detailmodelle zu ver-wenden. Diese Bedingungen sind erforderlich, um eine konsistente ("wasserdichte") Topologie zu gewährleisten. Es ist auch möglich, die Grenzfläche (d. h. geschlossene Schnittfläche im Sinne einer Bounding Box) zwischen beiden Modellen als unregelmäßige Fläche zu definie-ren. Das ist jedoch mit höherem Arbeitsaufwand verbunden und daher nicht sinnvoll. Die nach-einander auszuführenden Arbeitsschritte, welche zur Verknüpfung notwendig sind, wurden in einem Gocad – Wizard zusammengefasst.Die zweite Konzeption zeigt Wege zur …Spei-cherung von Gocad 3D-Modellen“ auf. Dabei wird vorgesehen die Datenspeicherung für Go-cad-Modelle in der Oracle-Datenbank des SLfUG in relationalen Tabellen zu konzipieren. 3.6 Das Strukturmodell Zwickau-SchedewitzDie Abbildung 2 zeigt einen Ausschnitt aus dem geologisch-tektonischen Model von Zwickau-Schedewitz. Die Datengrundlage für diese Model bildeten mehrere Grubenrisse des Zwickauer Steinkohlenbergbaus sowie sämtliche relevanten Bohr- und Schachtdaten aus den Archiven des LfUG.4 Weitere Arbeitsschrit-te/AusblickeDie vorliegenden 3D - Technologie soll praxis-nah getestet werden. Testobjekte sind die sächsi-schen Steinkohlenreviere von Zwickau und Lu-gau/Oelnitz. Anhand dieser Modelle sollen die entwickelten Konzeptionen zur Praxisreife ge-führt werden. Langfristiges Ziel ist es, den geo-logischen Untergrund Sachsens 3-dimensional zu erfassen und in einer 3D-Datenbank speichern. Eine interaktive Anwendbarkeit der 3-dimensionalen Daten durch alle, insbesondere angewandten Fachbereiche, wird angestrebt.Abb. 3: Ausschnitt aus dem geologisch-tektonischen Modell Zwickau-Schedewitz.。