With energy consumption on the rise around the world, interest in renewable energy sources has taken off. Wind power is a major component of the US energy strategy – it’s known for being affordable, efficient and abundant, as well as being pollution-free. Over the last decade, wind turbine farms have become a common feature, dotting landscapes across the nation, and today such massive operations comprise 4 percent of the total electricity generated in the US.

While wind power has many positive attributes, its main downside is its sporadic nature. In fact, actual power production is correlated with a range of atmospheric variables, such as wind speed and turbulence, as well as spatial and temporal scales.

Getting the most energy from these mechanical giants is thus a complex endeavor, but research teams are working hard to reduce the uncertainty that affects wind power forecasts. One of the main sites dedicated to optimizing wind power in the US is Lawrence Livermore National Laboratory. The lab has about a dozen atmospheric scientists, mechanical and computational engineers, and statisticians using fieldwork, advanced simulation, and statistical analysis to boost wind power production. High-performance computing is integral to the effort.

Jeff Roberts, Program Leader for Renewable Energy and Energy Systems, recently published a letter describing the lab’s role in developing this valuable resource.

“We must reduce our dependence on imported fossil fuels while ensuring plentiful clean energy with renewable sources,” Roberts writes. “The wind, however, is an intermittent resource that is challenging to predict, sometimes varying significantly from minute to minute. What’s more, complex atmospheric factors, such as turbulence, and topographical features, such as hills, modify the wind speed and direction and hence the power that can be extracted by wind turbines. Turbulence also plays an important role in the reliability and life span of turbine components.”

These simulations can be extraordinarily complicated, says Roberts. The complexity is owed to length scales that can an span eight orders of magnitude – from millimeters in the rotor-blade boundary layer to 100 kilometers for large atmospheric weather patterns.

“Simulating wind change and its effects on turbines is challenging because of the complex forces driving wind,” explains Livermore mechanical engineer Wayne Miller, associate program leader for wind and solar power. “We’re essentially simulating a fluid flow in an environment where factors such as aerosols, clouds, humidity, surface–atmosphere energy exchange, and terrain influence to varying degrees both the complexity of the flow and how much power can be extracted by a spinning turbine.”

The computational challenges are numerous, especially when simulating farms of more than 100 turbines. Terrain variations can significantly alter output from one turbine to the next and there are wakes coming from the spinning turbine blades that diminish power from turbines downstream. To negotiate these complexities, scientists are expanding the applicability of the Weather Research Forecasting (WRF) modeling system to be suitable for wind farm scale. Developed primarily for larger-scale weather applications, WRF is maintained by more than 10,000 users and contributors worldwide.

The model was modified for use at smaller scales and to satisfy the multiscale requirements of wind power forecasting. For example, a job may start out with a simulation of the western US to capture the dominant weather patterns. Then a combination of smaller grid spacing and models developed at Livermore are pulled in to accurately capture the smaller-scale features that affect wind farms.

The project seeks to blend WRF atmospheric simulation with scales of motion that are typically the purview of computational fluid dynamics (CFD) codes. To more expertly capture the complex interplay of variables, Livermore scientists have brought in a number of codes, such as WRF-GAD, immersed boundary method (IBM), as well as CGWind and HPCMP CREATE-AV Helios (aka HELIOS), which are used for even smaller-scale simulations that are outside the range of WRF.

A team of scientists from Livermore and University of Wyoming employed the WRF model and HELIOS to perform the first-ever simulation of a 50-turbine wind farm that takes into account individual spinning turbine blades using turbulent winds. This degree of precision and realism is helping researchers to understand why real wind farms fall short of their theoretical counterparts.

Atmospheric scientist Jeff Mirocha is one of the project leads exploring ways of studying phenomena that are specific to a wind farm environment. “The simulation framework we are developing will provide advanced tools to address these knowledge gaps,” he says, “leading to improved operations, longer component life spans, and ultimately cheaper electricity.”

President Barack Obama’s administration has set a goal for the nation to obtain 20 percent of its electricity from wind energy by 2030. The LLNL team thinks that’s a reasonable goal given the current high rate of wind turbine deployment nationwide. From 2008 to 2012, wind power capacity has expanded by 167 percent.

With precision models like the ones LLNL and its parters are developing, wind farm developers and operators have the information they need to select ideal wind farm locations and run the sites more efficiently.

“It’s a big team effort,” says Livermore’s Miller. Other collaborators include National Renewable Energy Laboratory, National Center for Atmospheric Research, University of Colorado at Boulder, Sandia and Pacific Northwest national laboratories, University of Wyoming, University of Oklahoma, University of California at Berkeley, U.S. Army, and other wind power industry stakeholders. Funding comes from the Department of Energy’s (DOE’s) Office of Energy Efficiency and Renewable Energy, as well as Livermore’s Laboratory Directed Research and Development (LDRD) Program, and industrial partnerships.