--> Abstract: Second Generation 3D Gridding: Changing the Way We Think from Reservoir Modeling, by Stan Abele and Jim Thom; #90105 (2010)

Datapages, Inc.Print this page

AAPG GEO 2010 Middle East
Geoscience Conference & Exhibition
Innovative Geoscience Solutions – Meeting Hydrocarbon Demand in Changing Times
March 7-10, 2010 – Manama, Bahrain

Second Generation 3D Gridding: Changing the Way We Think from Reservoir Modeling

Stan Abele1; Jim Thom2

(1) Seismic Micro-Technology, Inc., Houston, TX.

(2) JOA Oil and Gas LLC, Houston, TX.

Making 2D maps and more recently, 3D models of the subsurface is the mainstay for establishing funding and operations of oil and gas producers worldwide. Over the past hundred years these maps and models (including the processes, data types and machines used in generating them) have evolved to attempt better understanding of the subsurface with the goal of reducing the risks associated with drilling and field development Although each of the major subsurface geoscience disciplines (geophysics, geology and engineering) have made a great deal of progress within their individual areas of expertise it seems that no one has been analyzing the problem at the 30,000 foot level. Most critical field development and drilling decisions are still made without integrating the data and the human interaction around interpretation of the data. For the most part, this failure to integrate data is due to the different scales each of the disciplines measures their data in, often having a range of 5 orders of magnitudes in difference. Each of the major disciplines are discussing reservoir problems at different scales (see figure 1). They lack the integrated software tools required to understand how the data associated with the different scales can be cross referenced to provide insight into the key reservoir issues. Most often, each of the disciplines retreat back into their specific expertise trying to find ways to resolve the problem(s), unable to use the information provided by their team members. Finally, most 3D models will not include the latest information from the field as it takes too long to incorporate it or the associated specialists needed to assimilate data into the 3D model have been seconded to a “more important” project. It is expected that due to the demographics of the Oil and Gas Industry that these same seasoned specialists will be on the endangered species list soon. As costs for field operations escalate, the industry will need to consider different ways and means to create and update 3D models more quickly, while at the same time incorporating all of the field data from all disciplines (both old and new) in order to maximize the effectiveness of their field operations.

Figure 1: Example of data types and scales regularly measured and or derived for use in field development planning and Integrated Field Management Studies (IRM)

Historically, each discipline in the 3D modeling business had people almost entirely focused on solving the problems associated with their specialty or discipline. For example, while the geophysical groups were busy modeling seismic anomalies in the time domain the flow engineers were making new grids that more effectively solved the problems associated with flow behavior. This software development environment led to “best in class” type software solutions which may have further separated the disciplines. With each discipline speaking it’s increasingly specialized and sophisticated language and with little regard as to how other disciplines could use these new types of models and data to effectively solve production and exploration problems, it is no wonder that almost all decisions are made for the most part by “gut feeling”. Indeed, the problem is analogous to business planning and business intelligence (BI) where disparate information with different quality levels (errors or emissions) within organizations needs to be combined to make more effective decisions.

Figure 2 - Geologic Grid (Pillar Grid) on left compared to Simulation Grid (Stair Step Grid) on right with targeted reservoir zones in red. Note the differences in reservoir connectivity and geometry and how the same well path (in black) penetrates different zones in the middle fault block.

A common component that all 3D modeling, mapping, and simulation software packages have in common is a gridding system. Most all of the gridding systems associated with each of the major disciplines were specifically designed to solve their unique problems. For example, Geological Pillar Gridding was developed as a means of letting geologists build 3 dimensional models quickly and efficiently. These same grids were sometimes totally incompatible or at least created significant inaccuracies with models “coming from” the geophysicist or “going to” the engineering groups or team members.

Geophysical data “comes with” its own gridding system which is typically the bin spacing associated with the seismic survey. In 3D surveys the bin size or grid in the x-y direction is typically 25m x 25m in the x-y direction and 4 milliseconds in the z direction. Windows based seismic interpretation systems have developed cost effective, advanced techniques to quickly interpret fault networks and horizons in time or depth with the intent to define the container(s) (both structural and stratigraphic) that oil and gas are contained in. Depending on the structure, there are many instances when the structural surfaces that are interpreted cannot be accurately represented in a traditional 3D geologic model. Conflicts associated with Pillar Gridding systems used in first generation modeling packages require the geomodeler to change the orientation of the faults or build multiple models that need to be “patched together” to model the entire geologic sequences of events.

Figure 3 - modeling complex fault geometries with traditional pillar gridding can lead to verticalized faults (a) or “squashed” cells that need to be manual cleaned up (b). The impact of situation (b) is that the collapsed or flattened cells make modeling of transmissibility’s through faults almost impossible.

With each discipline having their own 3D models (and purpose-made grids for that specific problem) collaborative meetings that are held for the purpose of minimizing risk and maximizing productivity per well or for the field in total can be ineffective. Typically, geophysicists bring their structural map(s) and 3D models, geologists bring their stratigraphic map(s) and columns, while the engineers bring fluid contact map(s) and flow models to each meeting. If the drilling is being done in geomechanically challenging areas a fourth discipline (drilling and completion engineers) bring their maps and concerns to the meeting as well. Collaboration and facilitation amongst the different maps, models and disciplines is at best difficult which in turn forces field development decisions to be made from a “gut feel” as opposed to a more enlightened risk analysis process.

Thanks to new second generation gridding systems and tools all of the geoscience disciplines can now define, create and more easily maintain up-to-date static and dynamic reservoir models. The approach is facilitated through hybrid gridding techniques which enable fast, accurate links to models and maps generated in purpose built applications while providing a framework to integrate the data into one consistent 3D gridding technique. The grid and associated properties can then easily be passed and utilized amongst the disciplines.

Second generation Microsoft Windows based reservoir modeling techniques solve many of the important bottlenecks in today’s multi-disciplinary reservoir modeling requirements, with emphasis on:

  • Reducing model building and update times by orders of magnitude
  • Data integration across disciplines
  • An easy and intuitive interface that all disciplines can utilize for analysis of data and modeling (both static and dynamic) assumptions
  • Handling of complex geological reservoir structures

The main driver for the development of this new software and gridding technique is the shifting E&P market - where production of complex fields and heavy hydrocarbons has become increasingly important - thus requiring up-to-date reservoir models with a fast turnaround time measured in weeks instead of months. The vision is to build higher quality reservoir models, realized by no longer focusing on ‘best in class’ for individual parts of the static and dynamic workflow, to generate better insight on the total integrated reservoir model, referred to as 4 models in 1: the geophysical structural model, the geological model, the reservoir simulation model and the geo-mechanical model when required.

Figure 4 - Second generation gridding and visualization technology allows all disciplines to add their specific data to the field model and analyze the “reservoir puzzle” to ensure their interpretation of their data fits into the field model at different scales
The solution to this problem is a new kind of pillar gridding, which combines the advantages of pillar gridding, namely lower computer memory requirements and a regular ordering of stratigraphic layers, but overcomes the disadvantages mentioned above. There are two types of pillars in this grid: the vertical pillars and the pillars along the discontinuities like faults, intrusions or unconformities. The pillars along the discontinuities connect the vertical pillars. As a result horizontal surfaces can be modeled as a discontinuity and the gridding does not require that the fault block reaches the top or the base of the reservoir. Faults can also end in the middle of a reservoir without problem. No cell distortion occurs except along the discontinuities. The gridding process requires that discontinuity surfaces are truncated properly against each other and no holes are present. This is accomplished by computing the truncation lines from the edge of the surfaces in two directions. First from the inside direction to resolve any intersecting surfaces and then in the outside direction to close any holes.

Figure 5 - Three Dimensional geologic models (including horizons and faults) can be checked (quality controlled) for correct fault orientation in conjunction with the seismic data to analyze potential sealing or non-sealing parts of the model

This hybrid gridding allows 3D geomodels to be created without unnecessary distortion of the original structural interpretation. When integrated with a geophysical interpretation system, these structurally true models enable iterative modeling and interpretation, whether both activities are performed by the same geoscientist or if working in teams. Uncertainties in the geomodeled surfaces can easily be analyzed against the original interpreted data and discrepancies can be resolved and immediately reincorporated into the model. The time to an accurate, complete model which honors the scale of each domain is minimized, allowing for more informed reservoir management decisions.

Figure 6: Structural interpretation steps can now be minimized enabling geoscientists to “model while interpreting” thereby reducing cycle time and potentially increasing interpretation quality

A fundamental problem of geosciences is the creation of accurate 3D models of the subsurface. We can send a man to the moon but we cannot send him to the subsurface... yet! That being said, we did not make a successful moon landing without the level of significant collaboration and teamwork which second generation gridding enables in reservoir modeling.

704667_A.jpg

704667_B.jpg

704667_C.jpg

704667_D.jpg

704667_E.jpg

704667_F.jpg