If Noah had IBM’s Deep Thunder and Websphere iLOG software, he would have been piloting a “smart” ark and wouldn’t have needed that dove to find land.

“Our effort here started with weather,” said Lloyd A. Treinish, an IBM Senior Technical Staff Member, talking about Deep Thunder’s origins. But as former IBM CEO Lou Gerstner observed, “You don’t get points for predicting rain. You get points for building arks.”

In the 2007 to 2009 period, Treinish said, they therefore extended their work on “business decisions sensitive to weather” for the Deep Thunder services to “experimental work related to wind farms.”

In 2009, another IBM team was working with iLOG software to optimize wind farm operations and maintenance (O&M). “When we started to put together the wind solution,” said Jay Mashburn, IBM’s Principle Consultant for Wind Power, “we built an O&M solution as Phase I, and as Phase II we had this optimization engine that they used for generation sources.” It occurred to them that Deep Thunder’s weather-forecasting capability would improve the iLOG optimization engine.

“Each of the individual products had been field tested,” Mashurn said. “It’s when we put them together and created an end-to-end solution for wind power -- which could also be translated to solar power -- that we decided to build a simulation farm.”

Mashburn was referring to IBM’s virtual 42-megawatt wind farm hypothetically “comprised” of 25 turbines, a control room, a maintenance shop and an interconnection substation. In it, the IBM Energy & Utilities Solution Lab in Austin, Texas, is testing the Deep Thunder/iLOG tools against real world problems faced by the three major wind industry stakeholders, wind project owner/operators, utilities, and independent transmission system operators (ISOs).

They have discovered a singularly important feature of wind farm optimization. “The dollars that could be saved are pretty significant,” Mashburn said.  “In our demonstration, from the ISO perspective, looking out 84 hours in advance and doing optimization on it, it was several hundred thousand dollars.”

They have no plans yet to create a virtual solar power plant, but are confident that “the capabilities of the wind solution are very applicable and easily translated into solar,” Mashburn said.

“There, the issue is determining the level of the incident solar radiation,” Treinish pointed out. “That involves other aspects of the weather, such as cloud physics.”

Treinish added that IBM researchers are also doing work on “base technology associated with photovoltaics and solar thermal” to see how they can “use these sorts of weather models for those purposes. It’s not work that’s yet deployed,” he said. They are studying applications “to some of the experimental solar facilities that we’ve had here in our research labs.”

The most significant real-world applications of the tools have been at wind farms outside the U.S., though the locations are proprietary information. Theoretical, virtual and real-world evaluations have all shown, according to Mashburn and Treinish, that any application is going to have some degree of uncertainty. Individual wind turbines and wind projects vary as to how well equipped they are with the meteorological and operational sensors that gather the data Deep Thunder and iLOG digest to optimize productivity.

“We take advantage of what’s available,” Mashburn said. But, Treinish added, “A lot of this is about leveraging the client’s investment.” He said they “start with what’s there. But we would also be able to identify the gaps” and let the project owner/operator know what the limitations of the existing infrastructure are and what could be obtained with further investment. “The idea,” Treinish said, “is to develop the tools that focus on that business problem.”

There is big money value to all three of the stakeholder groups that IBM hopes its Deep Thunder and iLOG tools will serve. “Three different views of the world,” Mashburn said. “They want to get different things out of the same problem.”

For wind farm owner/operators, maximizing power production maximizes their return on investment. They can commit themselves to generating more electricity in Power Purchase Agreements (PPAs) if they know it is possible.

For utilities, a deep insight into how much renewable power they can harvest from their contracted suppliers allows them to be more confident of committing themselves to variable sources.

As ISOs integrate larger percentages of wind and solar into their portfolios, they will need to have “ramping” power generation sources available, and IBM’s tools promise to help them forecast the need for ramping and to both have “spinning reserves” in readiness mode and to more readily find generation sources on short notice.

The ultimate value of all these services, Treinish said, is entirely data-contingent. “We’ve developed a stochastic (statistical) model,” he explained, “that’s driven by a physical model that’s driven by weather. The stochastic model is built from data.”

IBM research scientists had been talking for some time about the challenges to using variable renewable energies they are now solving with Deep Thunder and iLOG when IBM sales executive Neil Gerber suggested there would be a market for such a product. Both tools are now commercially available, according to Mashburn, but have only recently been brought together for Wind Power Solution Phase II and have yet to find official form as a solar solution. It may be too late for Noah, but it's just in time for the expansion of renewables.