"When location of wind farms and transmission are picked to make best use of large-scale meteorological patterns, there is a dramatic improvement in how steady we found the produced electric power to be," Willett Kempton of the University of Delaware told environmentalresearchweb. "In five years of data, there was never an hour with no power production."

Not only would such wind farm interconnection reduce variability and remove periods of zero power production, it would also mean that any remaining variations in power output happened more slowly. This would give electricity suppliers more time to ramp up or down alternative sources or transmission links to meet consumer demand.

Together with colleagues at Stony Brook University, US, the Delaware team analysed wind-speed data from eleven meteorological stations off the east coast, stretching from the tip of Florida in the south to Maine in the north. The researchers used this data to estimate power output from turbine arrays for the period 1998 to 2002.

Currently power supply operators use existing systems such as reserve generators and redundant power-line routes to manage the variability in wind-power output; as well as combining remote wind farms via electrical transmission, as discussed here, it's also possible to employ energy storage, either at a central location or at distributed sites, for example through home heaters or plug-in cars, to smooth supply.

There are plans for wind-turbine arrays with a total power capacity of about 2,500 MW off the eastern coast of the US. This amount, which is roughly equivalent to the output of a large coal or nuclear power plant, is only 0.1% of the available wind resource in the region.

The researchers say that to connect these turbines would require 350 miles of submarine cable, which would add around $1.4 bn, less than 15%, to the estimated $10.5 bn installation costs. This supplement is roughly equivalent to levelling wind output via existing generation, which as a rule of thumb adds 10% of the wind-power cost for an electricity mix containing up to 20% wind power, and more for higher levels of wind power. And it's much cheaper than smoothing output via energy storage schemes, such as pumping water into higher-level reservoirs – these have capital costs roughly equal to those of generation.

"If this [interconnection] design were followed, wind power would still fluctuate but more moderately and more slowly than the individual wind farms being built today," said Kempton. "In the US East, this is the largest and most practical path to be able to achieve very large reductions in carbon dioxide. This paper shows that offshore wind can be inexpensively leveled, which will make it possible to use wind for a larger fraction of our energy needs."

The researchers propose the set-up of an Atlantic Independent System Operator to manage the bulk power market along such an offshore transmission cable.

Other studies in the US and Europe have also found smoothing resulted from connecting geographically distributed wind farms. But a study in Energy Policy by Oswald et al found that distributed generation would not help much in the UK as most of the region experienced the same wind conditions. Kempton and colleagues believe that it may be necessary to have a distributed grid larger than about 1000 km in size to encounter different weather. The orientation of the grid with respect to prevailing weather patterns may also be important.

Now the team would like to investigate how to optimize steady wind power supply in the region by picking locations more systematically – in this initial study they simply chose eleven evenly spaced sites.

The researchers reported their work in PNAS.