Big Data and 150 Trillion Calculations Per Second at Vestas (#IODGC)

IBM highlighted Danish energy company Vestas at their Information On Demand (IOD) 2011 conference in Las Vegas today. Vestas uses IBM big data analytics software to improve its wind turbine placement. The placements of wind turbines is a significant challenge for Vestas. During the IOD general session, Lars Christensen, VP of Plant Siting and Forecasting for Vestas, likened choosing the wrong location for a wind turbine to “moving the Hoover Dam.” That did a great job of describing the challenge!

Vestas employs 23,000 folks and is the world leader in providing high-tech wind power systems. The company has supplied over 44,500 wind turbines in 67 countries. Vestas’ wind library contains more than 2.8 petabytes of information tracking wind hour by hour over the past 12 years. That’s a lot of data. But the company relies on it to properly place its turbines. Obviously, they want to put them in windy places to generate energy.

Vestas is using IBM BigInsights software and an IBM “Firestorm” supercomputer to analyze this structured and unstructured data. In addition to wind data, the company also tracks tidal phases, geospatial and sensor data, satellite images, deforestation maps, and weather modeling research. The analysis of all of this data, which used to take weeks, can be done in less than an hour using the IBM system.

“Using IBM software and systems, we can now answer these questions quickly to identify new markets for wind energy and help our clients meet aggressive renewable energy goals,” said Christensen.

The software, IBM InfoSphere BigInsights, is the result of a four year effort of more than 200 IBM Research scientists. It is powered by Hadoop. BigInsights provides the framework for large scale parallel processing and scalable storage for terabyte to petabyte-level data plus the ability to enable “what-if” scenarios.

Vestas is running BigInsights on 1,222 connected, workload optimized SYstem x DataPlex servers that collectively make up the “Firestorm” supercomputer… which can perform 150 trillion calculations per second.

This is a real-life scenario of Big Data and how directing the right resources and solutions at the data can solve big problems.