Thrust Belt Imaging

The Tools for Accurate Seismic Imaging

By Tim McIntire

Gittins (left) and Vestrum (right) process seismic imagery using dual 30-inch Apple Cinema Displays, each with eight megapixels of viewable data.

The high-risk, high-reward business of oil and gas exploration relies heavily on the accurate analysis of seismic images. Drilling in the wrong place can cost anywhere from $1 million to $10 million per occurrence, so risk mitigation is taken seriously. Using an Apple Xserve cluster and 30-inch Apple Cinema Displays, Thrust Belt Imaging (TBI) has been able to implement new and improved seismic image processing techniques that drastically reduce risk. The accuracy of the new image processing algorithms — combined with efficient processing time from their eight-node Xserve cluster — enables TBI’s clients to bring oil and gas exploration back to gas-rich foothills that were once written off as too difficult to analyze.

“Anisotropic effects are traditionally ignored in seismic imaging,” TBI co-founder Rob Vestrum explains. “Thrust Belt Imaging uses an anisotropic depth migration algorithm to correct for imaging and positioning problems in seismic data, allowing image interpreters to make informed, accurate decisions.”

“It was a no-brainer to pick Apple for the interpretive workstations. Besides the availability of 30-inch displays, Mac OS X comes with Xcode, incredibly rich development libraries, 64-bit addressing, and OpenGL graphics.”

Rob Vestrum, Dale Schack, and Jon Gittins founded TBI in 2005 and have quickly built a reputation as the go-to guys when traditional imaging techniques, which ignore anisotropic effects, aren’t up to the task. Their mission is to mitigate the risk of searching for oil and gas in thrust belts found in the rocky foothills of numerous regions around the world, including parts of Canada and South America. Such regions are traditionally difficult to explore because of anisotropic effects, which make objects look different depending on the direction of measurement, leading to inaccurate drilling decisions. TBI was founded on powerful new methodologies for seismic imaging, but they knew that to go to market, they needed a stable, powerful compute platform to serve their clients, which include behemoths such as Shell and BP.

Deciding to Go Mac

The process of seismic imaging starts with data acquisition, which involves laying out thousands of evenly spaced surface charges that are detonated while sensors record the resulting vibrations. In simple terms, this method is a low-frequency ultrasound of the earth. The resulting datasets are measured in terabytes, and each image can require several hours of processing before it can be viewed in a meaningful way. TBI’s clients use the imagery to determine which areas should be drilled for oil and natural gas extraction.

With extensive backgrounds in seismic imaging using solutions such as UNIX-based workstations, Linux-based clusters, and NeXTSTEP Workstations, TBI’s founders were well positioned to make solid IT choices to launch their company. To start, they had to make two key decisions. First, they had to agree on the interpretive workstations for their hands-on image analysis. And second, they needed to choose a hardware and software solution for the back end.

The first decision was easy, says Schack. “It was a no-brainer to pick Apple for the interpretive workstations. Besides the availability of 30-inch displays, Mac OS X comes with Xcode, incredibly rich development libraries, 64-bit addressing, and OpenGL graphics.”

When TBI launched, the company needed the interactive analysis application up and running as soon as possible. With the rich library of tools and open source compilers in Xcode, the integrated development environment in Mac OS X, TBI was able to quickly set its focus on geophysics and geology rather than software development and maintenance. Xcode includes the GNU Compiler Collection (GCC), Interface Builder, and powerful debugging features, as well as the ability to compile universal binaries, which run natively on both PowerPC- and Intel-based Mac systems.

The second decision TBI had to make was what to use for the back end — the compute environment that would do the heavy lifting, pushing results and feedback to the scientists’ workstations. From prior experience, TBI had found that although Linux clusters seemed cost-effective on the surface, the time spent on system administration, hardware problems, and software development negated much of the perceived savings. With Mac OS X’s UNIX base, TBI could use Apple hardware for both cluster and desktop computing.

The benefits of running on a single, unified platform are immense. Before the availability of Mac OS X, a platform capable of running UNIX-based applications and desktop productivity software side by side, the standard solution was to have two computers at each scientist’s desk, one with an OS geared for science and one with an OS geared for productivity. TBI takes the unified OS theory a step further by using a single OS for visualization workstations, desktop productivity, and cluster computing — without compromise in any of these uses.

1 2 3 Next Page >