Face it: When you watch old Dallas reruns – or, better yet, catch that cool scene in “Giant” where
James Dean swaggers on the screen covered in oil – you get a sense that drilling
for “Texas tea” is a pretty cool gig.
But the truth is that this messy
business involves a lot of upfront investment, guesswork and ultimately
failure. Geoscientists must pinpoint where the “black gold” exists beneath a
dusty plain, a task that requires much data crunching. They've traditionally
resorted to scouring endless printed geological surveys to find where organic
matter exists underneath that has the greatest probability of containing
hydrocarbons that produces oil. They need to figure out where there are rich
deposits of sedimentary rock, the kind that's likely to hold such deposits.
After the spots are picked and
drilling starts, researchers need to analyze time-lapse seismic data from
subsurface rock formations, well and lab information, and sound waves covering
wide spaces between the wells. Sometimes, it's not easy to reach consensus on
what the data is saying.
“The information is not always
sufficient enough to resolve all possibilities,” says Ulisses Mello, manager of
Petroleum and Energy Analytics at IBM.
“People may have different views and try to make their best recommendations
based upon what they know. But it can be tricky.”
Given this scenario, Mello and his
IBM team are leading a partnership effort with Shell Oil Company to reduce
the risk factor of production drilling – specifically with respect to the
scientific modeling of these wells and the reservoirs. The IBM-Shell
Joint Seismic Inversion project is intended to reformulate and automate the
task of reconciling the various data collected to create a clearer picture,
and, thus, a more accurately predictive reservoir model. The IBM team is
taking extensive algorithms, simulation software, and advanced analytics
capabilities and other data and using this to provide the well models. The
effort will launch at first from an offshore North Sea field operated
by Shell called Draugen off the coast of Norway.
“We're seeking to minimize the
difference between the prediction and the reality,” Mello says. “We also hope
to reduce the time required to produce a model from six months to a matter of
weeks.”
Shell is seeking to blend
traditional and novel types of data into new information to better evaluate
complex, geological settings, says Hans Potters, manager of Shell's Reservoir Surveillance
Technologies. “We need to do this faster than we can at the moment,” he says.
“With this collaboration, we break new ground in this area, and we expect this
will increase oil and natural gas production in the future.”
Such an outcome wouldn't simply
save Shell on operational costs. It would benefit the environment as well.
“Anytime you increase efficiencies,” Mello says, “you reduce impact on the
environment and save energy too. Ultimately, we seek to drill less wells to get
the same results.”
