Reservoir modelling – new iterative workflow and software enables models to remain current

Reservoir modelling has long been a best practice as a practical component of ongoing field management in the oil and gas industry. A revolutionary new dynamic workflow has been developed by Fugro-Jason, saving time over existing approaches and allowing operators to keep the reservoir model up to date while also keeping to drilling schedules. New operational data and revised interpretations, such as a new well or a recently discovered fault, are easily incorporated into the model at any point in the workflow in a matter of minutes. Real-world use has shown a 3-5x speed up in modelling and maintenance tasks. Additionally, seismic data, often limited to being used as 2D maps can now be incorporated as full 3D volumes and used to drive the modelling process, enabling greater accuracy between wells. To top it off, the new software, FastTracker® can be run on both Windows®-based as well as Linux-based computers, reducing the potential cost and learning curve.

By Graeme Eastwood, Fugro-Jason

The Need for Modelling

The objective in building a reservoir model is, of course, to be able to predict the future performance of the field. Drilling rates are at an unprecedented level, generating large volumes of data that, if incorporated into a model, could yield a greater understanding of the subsurface. This could yield improved drilling pro­grams and advanced warning of opera­tional challenges. With increased focus on efficiently exploiting existing fields, modelling is taking a lead role in identifying opportunities and answering questions such as:

1. Are there ‘hidden’ reser­voirs that were missed?

2. What impedes flow?

3. What is the best way to leverage secondary recovery techniques?

4. How can overall recovery be optimized?

5. What downstream facili­ties will be needed in the next few years?

6. What surface facilities will be needed to deal with extra water, high­er gas mix, cutover to other wells, and other operational changes?

7. Is there a risk of ‘surprise wells’?

By interpreting data from well logs, cores, seismic and other data, geolo­gists, engineers and geophysicists can develop a better understanding of how the reservoir functions, and can simulate flow over time. An early model can help determine the initial drilling program. Ongoing models that take into account each drilling cycle can drive adjustments and tune operational logistics as the field matures.

Operational Obstacles

Clearly, there are compelling arguments in favor of modelling. However, until recently, the associated techniques and technologies have not been able to keep up with the quickening pace of new data acquisition. As a result, many fields are modelled infrequently, updated rarely and operational logistics tend to be driven by reactive decision-making.

Various problems in the traditional workflow have typically prevented even the largest oil and gas companies from keeping models up to date with current data. First, significant effort has been required to ensure that flow simulations match production data. This manual intervention, usually needed to accurately model the geological heterogeneity between wells, is time consuming and can be expensive.

The second problem stems from the need to update the model with new Information. The process has simply been too time consuming or specialized to be practical. In fact, in most cases so much time and effort is involved that information from a new well cannot be added to the model before the next well must be drilled.

These workflow problems introduce a tremendous burden into the modelling process. As a result, even in large fields oper­ated by the national and major oil companies, modelling has often been infrequently applied. Some fields go as much as three years between modelling efforts. This choice, based on economic realities at the time, forces a reactive stance at the field level. When the un­expected is encoun­tered it is handled in the moment. As time goes by the unexpected is compounded and the subsurface becomes more uncertain.

Changing the Workflow

The traditional linear approach to modelling had the advantage of stepping the user through the workflow. However, it made updating the model a lengthy process. Each discipline passed its results to the next With no looking back or consistency checking. Unconnected steps make updating the simulation model quite difficult.

Figure 1: Linear workflow commonly used to generate reservoir models.

To be effective and useful to everyone, the workflow must be dynamic, allowing changes or additions to the model at any time. Now technology exists that enables this iterative approach, automatically recording the workflow as the reservoir model is initially built and providing the framework for the model to be updated when any input is modified or any modelling pa­rameter is changed. Iteration is a choice and can happen anywhere in the workflow. By taking advantage of this new approach, companies can improve their drilling programs, production yield and predictive planning accuracy.

Linear and dynamic workflows are contrasted in Figures 1 and 2. The linear workflow shown in Figure 1 moves steadily from left to right. Although iteration is technically possible in the early stages of interpretation and structural modelling, and in the late stages of flow simulation and well planning, generally the workflow assumes a one-way path.

Figure 2: Iterative workflow enabling changes at any time at any point in modelling.

 

In contrast, the iterative workflow shown in Figure 2 assumes that iteration may be desired at any stage in the workflow. This workflow allows interactive, dynamic updating of new data and ideas at any position in the workflow. This approach increases accuracy and reduces the time involved in both building and updating the model, making modelling a living workflow.

 

The most realistic models are achieved by creating a framework using both seismic and well data. Porosity and lithology information derived from the seismic data can be combined with the well data in the initial geologic model. Models built with both the seismic and well information typically require dramatically less intervention to obtain satisfactory history matches with the production data. The resulting model is kept current by including ongoing operational data as it is acquired.

The first step in modelling is gathering well logs, cores, a geologic interpretation, and other op­erational data. A geologic model is then built from this data, based on a corner point grid (CPG) or a hexahedral grid. Geologists and engineers review this model to ensure it matches conditions observed in the field.

Next, seismic data is used to provide cost effective, laterally extensive field measurements because it provides better horizontal resolution. Techniques for acquisi­tion and processing have evolved over the past few years, leading to better qual­ity seismic data.

The biggest challenge to date has been to transfer this information from the seismic grid to the right location in the corner point grid, partly because of the different scales and geometries of the grids. Technology is now available to correctly resample the seismic derived properties into the CPG. This information can then be used as a trend for modelling reservoir properties.

Seismic inversion also creates a 3D represen­tation of the stratigraphy and lithology of the field, which is then combined with the geologic model to create a new mod­el that honors all the data in the field. The resulting grid, quickly built from the combination of geologic information and seismic data, can then be updated and upscaled as required. Engineers can use the upscaled model, as shown in Figures 3 and 4, to run flow simulations or for in-fill drilling decisions. This process is now far more integrated and substantially faster than ever before.

Figure 3: Fine-scale model. Figure 4: Upscaled model.

 

Updating the Model

The seismic-to- reservoir model workflow is recorded automatically as the first version of the model is built. It can then be modified and updated in minutes, via a simple drag and drop interface, every time new information is available. Geologists can import horizons from any source, create a geologic model including full 3D seismic information, upscale the results and analyze the reservoir over time using any simulator. Any input can be modified; any parameter changed. Figure 5 shows how a fault can be added, repaired, and truncated against surfaces.

Fast updating of three elements is key to this new approach:

1. Structure . When a structural feature is added, certain characteristics are implicit and automatic. For example, if a fault is dropped into the workflow and it crosses a horizon, the fault horizon interface is automatically adjusted (Figure 5). This modularity removes much of the need for manual intervention.

Figure 5: New fault added, repaired and truncated against surfaces.

1. Data . New operational data can be added to the model at any time. The effect of the new data is essentially immediate because after adding the new data, the previously recorded workflow is automatically rerun with the additional inputs.
2. Recorded Workflow . Every step of the workflow is automatically recorded. This is not a workflow manager and requires no training. The geologist or engineer can revisit any decision or parameter in the workflow at any time, even if made by another employee. All changes are immediately reflected in the model.

Significant time-savings enable ge­ologists and engineers to run multiple scenarios and simulations, resulting in a greater understanding of the reservoir behaviour in a fraction of the time previously required.

Supporting Operational Logistics

The value of ongoing modelling to ongoing operations is tremendous. For example, preventing the drilling of one dry well or re-routing a well to target the highest porosity zones through ongoing modelling can saves millions of dollars. Anticipating the needs of surface and downstream facilities can save substantially more than that.

Consider the case of a new fault discovery. This will likely have an effect on flow and may drive the need for an additional well. If the new condition could be quickly incorporated into the reservoir model, the operator could determine the best course of action to preserve or enhance the field’s production.

A common complaint with infrequent modelling is that the realities in the field no longer match the predictions of the model. A quick way to improve field management, then, is to incorporate new field data as quickly as possible into regular modelling and simulation. In the past, models were not able to keep up with the volume or pace of data acquisition and got out of sync. With the new approach, this is far less likely to happen.

Dealing with Uncertainties

Even the most thorough model can miss an uncertainty that results in a ‘surprise well’. With the ability to update the model, geologists can quickly incorporate the ‘surprise well’ into the model, speeding reaction time from months to weeks. This same capability makes it easier to model a wide range of uncertainties, reducing the risk of surprises.

Scenario Modelling

Accurate models created and updated in a fraction of the time previously required can help improve operational logistics and overall field production. An iterative workflow opens the possibilities for in-depth analysis without significant time and effort, yielding:

• Fine scale interpolation between closely-spaced wells
• Estimation of uncertainty to assess risk
• Generation of lithotype (e.g., sandstone, shale, high and low porosity carbonate) volumes
• Estimation of porosity (crucial in carbonate fields) from seismic impedance
• Integration of high vertical resolution well data with high lateral resolution seismic
• Inputs for reservoir flow simulation

Geologists and geoscientists can then investigate various options for production wells and explore multiple models and interpretations. When dealing with large fields, the significantly reduced time investment makes a compelling case for keeping reservoir modelling in step with data acquisition.

Fugro-Jason is part of the Fugro group of companies.

Windows is a registered trademark of Microsoft Corporation in the United States and other countries. All other product names and trademarks are property of their respective owners. All rights reserved.

Contact:

Graeme Eastwood
Region Manager, Fugro-Jason
c/o Fugro Survey (Middle East) Ltd
PO Box 43088
Abu Dhabi
United Arab Emirates
T: +971 2 554 1011
F: +971 2 554 7811
E: geastwood@fugro-jason.com