To Drill or not to Drill By Gene Charleton and Lesley V. Kriewald
Petroleum engineers like information — the more, the better, usually. It helps them decide where to drill for oil. But sometimes having the right information is more valuable than having a lot of it.
Decisions, Decisions
Information is one of the most valuable tools you have when you’re drilling for oil. Decision science can help you understand how much that information is worth.
All information is not created equal. That’s a fact of Eric Bickel’s professional life.
Bickel is an engineer in Texas A&M’s Department of Industrial and Systems Engineering, and he’s an expert in decision science — using mathematics to help make complex decisions. Decision science uses the odds that something will happen, its probability, to help decide what to do in complicated situations. Few situations are more complicated than when oil producers decide where to drill new wells.
Drilling for oil is a high-risk, high-payoff proposition. Your chances of finding oil in any particular place may be low, but if you do, the payoff is high. If you’ve seen the classic movie Giant you understand how this works.
One way to improve your odds of finding oil is to use seismic imaging to get a “picture” of what the underground geography looks like. You get seismic images by setting off small explosive charges and mapping how the vibrations from the explosions move through the rock formations, or strata, under the ground. Certain strata are associated with the presence of oil.
“Geophysicists will often explore technical aspects of seismic imaging of reservoirs, attempting to predict whether or not it will be possible to detect the presence of oil,” says Richard Gibson, a specialist in seismology and associate professor in the Department of Geology and Geophysics. Or, if they’re dealing with a known oil reservoir, provide some estimates of the amount of gas or oil in the reservoir.
Bickel, Gibson and Duane McVay, an expert in reservoir management and professor in the Harold Vance Department of Petroleum Engineering, are evaluating the effectiveness of new technology developed by WesternGeco, a subsidiary of the international energy company Schlumberger that provides seismic services to oil producers. The new technology produces seismic images of the underground landscape that are more detailed and complete than those from conventional seismic technologies, says Stephen Pickering, marketing manager for WesternGeco’s reservoir seismic services.
“The question is, ‘How much value does this additional detail add to the information we can give the producers?’” Pickering says. “We think it adds quite a bit, but we’d like to be able to quantify it.”
Enter decision science
New information — such as the added detail in WesternGeco’s seismic technology — is valuable to oil producers’ decisions if it is relevant, material and economic, Bickel says.
Drilling for oil is a high-risk, high-payoff proposition. The right information can help reduce the risk.
For information to be relevant, you must be uncertain about something that’s important to your decision, and the new information must have the possibility of telling you something useful about the uncertainty. For instance, you may be uncertain about whether it’s a good idea to drill a well in a particular location.
For information to be material, it has to have the potential to affect your decision.
“If you’re going to take some particular action no matter what, new information is worthless,” Bickel says. On the other hand, if you’re considering drilling in a particular place and whether you drill or not depends on information you can get about the site, that information is material.
Finally, for information to be economic, it must be a good investment. Even if new information tells you exactly what you need to know, if you can’t afford to pay for it, it’s not economic.
By applying the mathematics of decision science to the situations oil producers who use seismic services face, Bickel, Gibson and McVay are determining how much value the additional detail adds.
“We were able to leverage Texas A&M’s energy expertise to help WesternGeco better communicate the value of its product to potential customers. In addition,” Bickel says, “the methodologies we have developed will help exploration and production companies make better use of their capital and hopefully discover more reserves.” 
Improving the Odds
Drilling for oil is one of the biggest gambles there is. New high-resolution computer models may help oil producers reap big-time payoffs.
Drilling for oil can be a multimillion-dollar gamble at long odds, but Akhil Datta-Gupta is betting he can make those long odds more favorable.
Highly detailed modeling of existing and future oil reservoirs could pay off by helping make “substantially” more oil available to U.S. producers and reducing our dependence on foreign oil.
Datta-Gupta, the LeSeur Chair in Reservoir Management in the Harold Vance Department of Petroleum Engineering, says that much of the current domestic oil and gas production comes from three sources — “mature” or partially depleted known reservoirs, geologically complex formations and ultra-deepwater reservoirs in the Gulf of Mexico.
The challenge for petroleum engineers is to identify the location and distribution of the “unswept” or bypassed oil and untapped compartments in these reservoirs. To do this, Datta-Gupta uses high-resolution fluid-flow modeling and seismic imaging techniques in combination with data assimilation methods to determine where best to drill to recover this unswept oil.
Drilling by the numbers
Petroleum engineers routinely use numerical models to understand and visualize fluid flow in the reservoir and for future performance predictions. Recent advances in seismic imaging meant that today’s geologic models consist of tens of millions of grid cells, or computational elements — so many elements, in fact, that scientists and engineers using conventional flow modeling techniques usually resort to “upscaling” or averaging schemes to reduce the number of computational elements.
Upscaling schemes, however, run the risk of losing significant geologic features of the reservoir, which can have a major impact on oil recovery. To avoid losing those features, Datta-Gupta’s team is developing “streamline-based” flow simulation techniques that model fluid flow directly at the scale of geologic models with multimillion computational elements.
The basic idea, Datta-Gupta says, is to decouple, or split, the 3-D problem into individual 1-D problems that can be solved relatively quickly, resulting in orders of magnitude savings in computational time compared to conventional flow simulation techniques. Such decoupling also makes the technique well suited for parallel computation.
And this streamline-based flow simulation technique allows Datta-Gupta to look at the interaction of fluid flow and subsurface characteristics at a highly detailed level to identify unswept or bypassed oil for targeted recovery.
“We can quickly look at multiple models and make multiple predictions to quantify uncertainty in our subsurface models and future predictions — to identify and optimize drilling locations,” Datta-Gupta says.
High-stakes game
And optimizing drilling locations is important: Drilling a single well in the deep waters of the Gulf of Mexico can cost more than $30 million, and completing the well for production can cost another $30 million.
“We cannot afford to drill too many dry holes, so before we spend that money, we need to know how much oil is down there, where it is located and how to get it out efficiently.”
In this figure, the yellow lines display stremlines indicating dominant flow patterns below the surface.
The vertical lines (magenta) show the locations of wells in the field. The green background shows the conductivity patterns below the surface.
With these highly detailed geologic and simulation models in hand, Datta-Gupta’s team then uses a variety of “dynamic” fluid flow data (such as oil, gas or water production; pressure; and time-lapse seismic images) from the reservoir to calibrate these models. Integrating the information into the geologic models allows Datta-Gupta to identify flow channels and barriers as well as any compartmentalization in the reservoir.
And with advancement in well construction and down-hole sensor technology, Datta-Gupta says the dynamic data can be available every minute. The amount of data can be simply overwhelming, he says, and the challenge is to assimilate the data in real time for on-the-spot decision making.
“We need to update the geologic model in real time to facilitate geosteering — that is, to guide the well trajectories during drilling,” Datta-Gupta says.
Datta-Gupta says that knowing the properties of the subsurface reservoir in detail is critical for designing any targeted and environmentally sensitive drilling scheme that leaves minimum “footprints” and for improved oil recovery programs.
And to put it all in perspective, Datta-Gupta says, “If we can improve domestic recovery in existing oil fields by, say, 5 percent, it will mean an extra billion barrels of oil for the United States over the economic life of the existing fields.”
Texas A&M Engineer Online
Premiere Issue
[ Back to 2006 Issue ]Business
Energy
- Biomass and clean air
- Energy 101
- Keeping the lights on
- Nonstop coast to coast
- Nuclear by the numbers
- Petroleum under pressure
- Policy + technology = security
- Tapping the trash alternative
- To drill or not to drill
