the film forum library tutorial contact

How to Grow Wind Energy? Faster, More Powerful Data Crunching by Staff Greenbang, October 24, 2011

The most effective, cutting-edge wind turbine in the world can't generate energy if it doesn't get a good steady wind. So placing turbines in the right places can make the difference between a really good investment in renewable energy, and a losing bet.

But accounting for all the variables that together determine the "right place" hasn't been easy. That's because there are just so many variables: changing weather patterns, prevailing wind directions, turbulence, the presence (or lack of) trees in the area, even aesthetics. Plus, all these factors tend to interact in complex ways, meaning that if one changes, many of the others will also change.

Running those kinds of data through wind-turbine computer models used to take weeks. But IBM says it can now help wind-energy companies do those calculations in an hour or less.

In fact, Big Blue has just started working with Danish wind-energy giant Vestas to do just that: crunch myriad facts and figures fast to help the company make the best siting decisions possible for its turbines.

Coupled with its Firestorm supercomputer, IBM is providing Vestas with speedier turbine location guidance using its InfoSphere BigInsights software. The software took 200-plus IBM research scientists and four years of effort to develop, and can process and store terabyte- or petabyte-levels of data to create "what-if" scenarios for turbine placements.

Vestas is using that to analyze huge volumes of data: weather reports, tidal phases, geospatial and sensor data, satellite images, deforestation maps and weather modeling research.

"Vestas turbines operate for decades and clients demand to know how much energy they will produce and what their return on investment will be before they are installed," said Lars Christian Christensen, vice president of plant siting and forecasting for Vestas Technology R&D. The new software-supercomputing approach, he said, will both help the company "quickly to identify new markets for wind energy and help our clients meet aggressive renewable energy goals."

The wind-energy company is running IBM's software on 1,222 connected, workload-optimized servers that together make up the Firestorm supercomputer. The servers can handle 150 trillion calculations per second, the equivalent of 30 million calculations a second for every person in Denmark.

Once each turbine is installed, Vestas engineers plan to use the IBM technology to predict turbine performance, analyze how each blade reacts to weather changes, and determine the best times to schedule maintenance. Over the next four years, the wind-energy company expects to crunch over 20 petabytes of data in an effort to continue fine-tuning its siting strategies.

"Vestas shows how large organizations can tap big data analytics and ever more powerful computers to make smarter business decisions that can substantially accelerate growth while tackling some of the world's most pressing issues," said Arvind Krishna, general manager, IBM Information Management. That kind of data insight, he added, "has the potential to transform entire industries."