Forecasting Wind

Forecasting Wind

September 09, 2009

Ken Silverstein
EnergyBiz Insider
Editor-in-Chief

Wind power's potential could be improved if developers had better forecasting tools. Knowing about when, where and how fast the wind blows is one thing. But being able to more accurately make those assessments is clearly a cut above.

As it stands now, if the estimates are off and the wind takes a hiatus, utilities must crank up their backup generation and their production costs will rise as a result. The issue will invariably compound itself as state mandates force power companies to provide more wind power. By taking years of real weather data and modeling it to fit a specific location, however, forecasters say that they can come within a few percentage points of what the actual results will be -- even years down the road.

"Whether harnessing the power of wind or the sun, the diversity (of our) user base illustrates the vastness of the renewable energy ecosystem and the tremendous value that accurate, affordable information plays in driving renewable energy adoption," says Kenneth Westrick, chief executive of 3TIER. The technology "shortens the process needed to make sound decisions about where to invest in renewable power."

Seattle-based 3TIER's tools that can be accessed over the Internet allow wind developers to analyze hourly wind speeds throughout the year and around the country. The same tools can also be used to examine daily wind patterns for each month of the year. National governments around the world including that of the United States, meanwhile, are funding similar research. All understand that more accurate forecasting would increase wind's value and help developers get the financing they need.

Palo Alto, Calif.-based EPRI is working on same-day and next-day forecast algorithms. The objective is to work with transmission system operators to refine the technology, the method of delivery and the visual display of real-time forecasts. The organization furthermore says that studies have indicated that the cost for an imperfect next-day wind energy forecast ranges between $0.4/MWh at 10 percent error to $1.4/MWh at 50 percent forecast error -- enough to compensate for some public subsidies.

During the second quarter of this year, the U.S. wind energy industry completed a total of 1,210 MW in 10 states, says the American Wind Energy Association. These new installations nudge total U.S. wind power generating capacity to 29,440 MW, it adds. The U.S. wind power generating fleet now offsets an average of 54 million tons of carbon annually, reducing carbon emissions from the electricity sector by 2 percent or the equivalent of taking 9 million cars off the road.

Ultimately, proponents of wind power say that the fuel source can capture as much as 10 percent of the electric generation market in the United States within 20 years. The most optimistic projections say that 20 percent is possible under the right conditions.

Variable Nature

The expectation is that wind forecasting would become more of a perfect science. The Bonneville Power Administration, for example, is sponsoring a competition between two such forecasting firms, AWS Truewind and Energy & Meteo Systems. The business that makes the most accurate predictions will then be selected in 2010 to develop a full-scale model for the utility's entire fleet of wind generators located in the Northwest -- all necessary to help reduce the risks in a region with lots of variable winds.

Meantime, Ontario, Canada, the country's most populated province and that nation's leader in wind with 1,200 MW, is actively pursuing the technology. Its transmission network operator has already implemented a web-based system to illustrate both the hourly output and hourly costs from the province's wind plants.

"Due to the variable nature of wind generation, accurate wind forecasting is essential to the operation of an efficient and reliable power system," says Paul Murphy, chief executive of Ontario's Independent Electricity System Operator. The goal, he adds, is to increase reliability and to subsequently build more wind generation.

Through 3D modeling, companies can now use accurate wind histories to fit prescribed locations and they no longer have to extrapolate from a finite amount of data. Scheduling power therefore becomes easier and more predictable.

Forecasters have in essence leveraged information accumulated by weather services. The number one variable is wind speed, because the amount of energy generated by a plant goes up significantly even with small increases in velocity. So, the key is to find a good site with adequate wind, all before developers go through a complex process to get land rights. It's now common, though, to buy wind rights from land owners. And then comes the construction and financing cost along with the transmission interconnect and transmission rights issues. Lastly, developers must negotiate with turbine vendors.

Bankers, understandably, are cautious. They loan money based on what a developer's worst year may be. If simulated models can produce accurate weather forecasts and provide assurances of what cash flows will be, financing becomes easier and likely cheaper. State mandates -- and potentially federal ones -- are prompting more research and development into this field.

"By using three-dimensional modeling, utilities can get a comprehensive understanding of an entire project in three to four weeks," says Mark Ahlstrom, chief executive of WindLogics in St. Paul, Minn., in a prior talk with this writer.

If wind energy is to reach its high-mark, then utilities must have the tools to better grasp such key variables as how often and how fast winds will blow. New forecasting technologies aim to do just that and thereby enable developers to reduce their risks