TRIP goes deep for fossil fuels

Rice’s Bill Symes leads the way in sophisticated seismic modeling

BY MIKE WILLIAMS
Rice News staff

X-ray vision would sure be handy for Bill Symes’ clients, but he’s hot in pursuit of the next best thing.

Symes, director of The Rice Inversion Project (TRIP), and his students are looking for the holy grail of oil and gas exploration, a mathematical model that will quickly and accurately interpret seismic data and tell them where the good stuff is.

It’s no small quest. Though the cost of a barrel of oil is fluctuating beyond all reason, the necessity of finding more sources is imperative in a world that, at least for the foreseeable future, will be hungry for energy from fossil fuels.

Bill Symes, director of The Rice Inversion Project, and his students are looking for the holy grail of oil and gas exploration, a mathematical model that will quickly and accurately interpret seismic data and tell them where the good stuff is. Photo by Jeff Fitlow

Symes will speak about his team’s specialty — quality assurance in large-scale seismic simulations — at the 2009 Oil & Gas High Performance Computing Workshop to be held at Rice March 5, a daylong event at Duncan Hall’s McMurtry Auditorium. The workshop is free, but registration is required.

Symes and TRIP have quietly gone about the business of maximizing returns on industrial research since the early ’90s, and their success shows in the fact that even though the industry is in turmoil, the project continues to grow in size and stature.

”I have more students than I’ve had since the mid-’90s,” said Symes, the Noah Harding Professor of Computational and Applied Mathematics at Rice. ”The number of sponsors is at an all-time high; I now have 10 paying sponsors, plus one that contributes in kind. That’s the highest it’s ever been by far.”

The problem TRIP has been attacking all these years lies in the complexity of interpreting data from seismic research — the process of using waves to model terrain beneath the surface of the Earth. A seismic wave’s speed and how it’s absorbed or reflected tells researchers a great deal about the unseen world beneath the ground or ocean.

Symes noted the impossibility of getting measuring devices 10,000 feet below the sea floor. ”We’re never going to see that place. You have to use some kind of physical signal that actually goes through rock.” That signal can take the form of controlled explosions at the bottom of boreholes or blasts of sound from arrays towed behind seagoing labs. Filtering out bad data from good is another part of the process addressed by TRIP.

”I guess that’s what tickles me pink about the whole subject,” Symes said. ”You actually wind up seeing through rock. I think that’s cool.”

Symes’ primary research interest remains finding a way to unravel the mass of data seismic explorers collect and return useful results. ”We look at blue-sky kinds of things,” he said. The goal is to turn raw data into where-to-drill data and to do it quickly.

Since its founding in 1992, TRIP has contributed to a number of advances that started out as ”blue-sky” ideas but have become commonplace in geophysical computation. An industry standard method for computing the time it takes for a seismic wave to get from A to B is partly a product of TRIP research, as is a way to accurately simulate the attenuation and dispersion of waves traveling through a variety of materials — rock, soil, gas and oil.

Creating models to interpret data starts with the rocks, he said. ”If you look at a cut in a mountain, the rocks look really complicated, and they are. The physics that govern the way they shake, the way electromagnetic radiation gets through rock and all the types of radiation you can use to remotely sense what’s down there are pretty complicated.

”You make some simplifying assumptions so the description becomes tractable as a math/computer problem. You have certain structures in the Earth, and you try to understand what kinds of data they would cause you to measure. That’s a prediction problem, and we can figure out how to solve that.

”Then you try to run it backward. You measure the data and ask, ‘Now, what structure caused that?’ That’s the inversion part.”

Automatically updating the velocity model (a map of wave speed at various positions in the Earth) is key to getting good results, he said. The signature product of TRIP is a method called differential semblance optimization, which produces such automatic velocity updates. ”We invented differential semblance and did the basic work on the method, and now quite a few other people and organizations are evaluating its potential.”

But there’s more to do. Like most of the state-of-the-art in seismic data processing, differential semblance, as formulated up to now, separates the propagation and reflection of waves into two processes. “The bottom line is to use real physics, where the propagation and reflection of a wave are not separate phenomena but are treated coherently.

“That’s what I’ve been working on since I came to Rice, and I’m going to keep working on it until I get it right.”

Producing a “complete, coherent, physical explanation of the data” is an appealing proposition, said Symes. “Many people have been seduced by the inversion concept, but it’s almost never been done.”

Only once, he said, have all the elements come together to accurately locate a reservoir through inversion. In that instance, graduate student Susan Minkoff, now an associate professor at the University of Maryland, used raw data gathered in the vicinity of an existing commercial well as a way to prove the Rice concept.

”One of our sponsor reps helped us pick out a piece of data, premasticate it so that it was clean, helped Sue formulate the approach and figure out the important physics. Then Sue implemented all that and experimented,” he said.

The three-year project hit paydirt when Minkoff was able to analyze the data and come up with the correct results. ”There was a region of the Gulf of Mexico our contact had called the ‘mother of all gas sands.’ The sponsor was able to give us a record of the physical parameters, and we got them exactly right.

”It was a great moment, and it hasn’t been repeated by us, or by anybody else, to my knowledge. Other folks at the sponsor company have told me that when people ask, ‘Does inversion work?’ they cite Sue Minkoff’s thesis. They say it could — it did once.”

Symes and his students are positive that successful inversion will happen again, and indeed will eventually become a routine and reliable tool in the search for fossil fuels, as well as another feather in Rice’s cap. And that’ll be a trip.