Looking for oil and gas is becoming a lot more like a video game.
This is the unique intersection where Houston-based Bluwarestands. The software company is among a handful of first movers in the upstream industry that see the use of GPUs, or graphics processing units, as the way to eliminate all the wasted time spent waiting for seismic data to be visualized.
GPUs have been around for decades, but only in the past few years have they become one of the most talked-about pieces of computing hardware in the digital tech scene.
Their uptake was initially driven by the need to quickly render graphically rich video games, especially first-person shooters, and enable players sitting around the world to co-play online without lag. Others have since awoken to the potential of GPU technology, which is now accelerating the development of driverless cars and represents the computing muscle behind cryptocurrency markets.
In the modern video game setting, any significant latency in processing might mean game over. “Whereas in oil and gas, if you have to wait another half an hour to get a seismic line to load, you just go get a cup of coffee,” Bluware Chief Executive Officer Dan Piette said.
If that time is given back, those coffee breaks may be the thing put on hold as seismic processing ceases to be a major bottleneck. Bluware says on its website that it can shave several months off the typical seismic project.
Piette said this could translate into operators gaining the inside track on bidding for offshore blocks that their competitors might bypass simply because of using slower interpretation methods that rely on traditional hardware.
More Seismic, Fewer Eyeballs
And even though the industry downturn hit offshore exploration budgets as hard as any other part of the service business, Bluware sees plenty of opportunities for its cloud platform that also incorporates machine-learning tools to help automate key parts of subsurface characterization workflows.
The company estimates that global producers are sitting on at least 1.5 exabytes of seismic datasets that can be reprocessed with modern technology to find more oil and gas. To put the volume of this data into perspective, an exabyte is equal to more than 1,000 petabytes while a high-resolution offshore survey in the Gulf of Mexico might sail home with between 1-2 petabytes.
Piette argues that this inventory of data is only getting bigger in size, and that the lasting effects of the downturn mean there are not enough technical professionals to interpret it all. “You can’t have 50 geophysicists looking at one petabyte of data and come up with the answer any more—you need one geophysicist looking at 50 petabytes of data,” he said.
Earlier this month, Bluware’s cloud-platform attracted an unspecified investment from the Stavanger-headquartered EV Private Equity firm and Shell’s venture capital group in Houston. Kirk Coburn, an investment director with Shell Ventures and newly appointed member of Bluware’s board, said in a statement that the software company’s innovation, “delivers a real-time interactive experience for activities we typically expect to take hours and days.”
More Agility In The Cloud
The investment comes about a year after Bluware merged with Hue and Headwave, two other geoscience software firms that were already closely affiliated. This new trifecta is positioned against the biggest names in the seismic software business, including Halliburton, Schlumberger, and Paradigm (acquired by Emerson last year).
Historically, seismic software developers have used techniques to help load seismic data faster, such as “bricking,” which paints the seismic picture block by block across a screen—similar to the experience of loading websites in the days of dial-up internet. But among the first things you notice when seeing Bluware’s seismic software is how users can begin with a regional view of seismic data and then zoom down to the block-level and onto the well-level as fast as their mouse will allow.
Piette explained such agility is in part thanks to the company’s data architecture, which lends its self to the cloud-storage approach known as containerization. “This is a technique that came from the gaming industry, because they needed to be able to move and access data really fast,” he said. “By changing the format ourselves, to a virtualized data source, you can use a map or coordinates and then go in and cookie-cutter the data out.”
Bluware knows that not every seismic expert will make the leap to its machine learning-based interpretations, and so it is using transcoding software that makes it possible to integrate its platform with outputs that can be used in competing legacy software.
Piette characterized this feature as “a transitional service” that will allow companies to move forward in their digital initiatives while allowing the most experienced practitioners to work within familiar environments.
“We see what’s going to happen 5 years from now,” he noted, “but we also know that people have to find oil between now and then.”
