By Braden Reddall and Matt Daily

HOUSTON/NEW YORK (Reuters) - IBM thinks it can make geologists and engineers more effective at mining the fast-growing wealth of data on everything from oil reservoirs to refineries to help them find, extract and process oil.

The computer services company has spent the past three years building a team of 5,000 consultants, scattered around major oil basins worldwide, to help companies tackle the data explosion.

Just having a lot of data is not enough: As with crude, data must be refined and then shipped out to the people who can use it: those who make drilling decisions with millions of dollars on the line.

"Data management: we see a lot of effort, and by effort I mean investment, in that space," David Womack, IBM's director of strategy and business development for chemicals, petroleum and industrial products, told the Reuters Global Energy & Environment Summit on Tuesday.

"They'll give you these big data sets, and being able to manage and manipulate that is a non-trivial act," he added.

Chevron Corp has estimated its network manages 1.5 terabytes a day, or equivalent to about 1.5 million books, and says its computers store more information than the Library of Congress.

Much of that is due to the development in recent years of three-dimensional oil reservoir imaging, which is far more data-intensive than 2-D seismic imaging. Even on the other end of the oil value chain, refiners deploy thousands of small instruments to monitor various parts of their plants.

"They're about valves and transmitters, and if it creates a pile of data for somebody? They don't think about that," IBM's Womack said.

He said Royal Dutch Shell Plc hired his team to help manage the data deluge.

An equivalent expansion of processing power has been driven by "cluster computing," with multiple computers tied together to solve one huge problem. "As you get faster machines, you can beat out analytics that tell you you've got to change injection rates in this area or that area," Womack said.

"You can hit enter and get an answer back in 10 hours, instead of 10 weeks."

Womack stressed that sophisticated imaging will be crucial in locating new reservoirs now that all the "easy oil" is gone.

Three years after Chevron introduced the industry's first 3-D visualization centers in 1997, the California oil company started working with cluster computers, which are now its standard set up. It retired its last mainframe computer in 2010.

In a 2010 interview, Chevron Chief Technology Officer John McDonald explained how real time reservoir management allowed the company to start up a Gulf of Mexico project three months early. "That data has great utility if you can start making sense of it," McDonald said at the time.

Follow Reuters Summits on Twitter @Reuters_Summits

(Reporting by Braden Reddall in Houston and Matt Daily in New York)