Elsevier

Renewable Energy

Volume 62, February 2014, Pages 331-340
Renewable Energy

Meteorologically defined limits to reduction in the variability of outputs from a coupled wind farm system in the Central US

https://doi.org/10.1016/j.renene.2013.07.022Get rights and content

Highlights

  • Surface wind is determined by transient waves and boundary layer turbulence.

  • Central Plains region is under the influence of transient waves.

  • Meteorologically defined limits to reducing the power variability is quantified.

Abstract

Studies suggest that onshore wind resources in the contiguous US could readily accommodate present and anticipated future US demand for electricity. The problem with the output from a single wind farm located in any particular region is that it is variable on time scales ranging from minutes to days posing difficulties for incorporating relevant outputs into an integrated power system. The high frequency (shorter than once per day) variability of contributions from individual wind farms is determined mainly by locally generated small scale boundary layer. The low frequency variability (longer than once per day) is associated with the passage of transient waves in the atmosphere with a characteristic time scale of several days. Using 5 years of assimilated wind data, we show that the high frequency variability of wind-generated power can be significantly reduced by coupling outputs from 5 to 10 wind farms distributed uniformly over a ten state region of the Central US in this study. More than 95% of the remaining variability of the coupled system is concentrated at time scales longer than a day, allowing operators to take advantage of multi-day weather forecasts in scheduling projected contributions from wind.

Introduction

Some 13,131 MW of wind generating capacity were added to the US electrical system in 2012, an increase of 93% compared with the same period in 2011. Total installed capacity for wind power in the U.S. amounted to 60,007 MW by the end of 2012, equal to approximately 6% of total U.S. power generating capacity. Lu et al. [1] argued that an onshore network of GE 2.5 MW turbines installed in the contiguous U.S. could supply as much as 16 times total current U.S. demand for electricity. A study by the U.S. Department of Energy concluded that wind could account economically for 20% of total U.S. demand for electricity by 2030 [2], while Short et al. [3] argued that as much as 25% of demand could be met feasibly by 2050.

The current electrical system requires an essentially instantaneous balance of supply and demand dictated largely by the latter. Opportunities for storage of electricity when supply exceeds demand are limited, while options to modulate demand are also minimal. Baseload demand is accommodated in the present system mainly by a combination of contributions from nuclear and coal with an additional contribution in some regions of the country from hydro. Gas-fired systems provide the fast response required to adjust to short and intermediate-term fluctuations in demand. The challenge posed by the need to incorporate a significant source from wind relates to the intrinsic variability of this source. Production of electricity from an individual wind farm can vary over a wide range on time scales as brief as minutes or as extensive as days or even longer [4].

A number of authors have pointed to the advantages that could be realized by combining outputs from a series of spatially separated wind farms [5], [6], [7], [8], [9], [10], [11], [12], [13], [14]. Katzenstein et al. [10] reported a frequency dependent analysis of the smoothing in output that could be obtained by coupling up to 20 spatially separated wind farms in Texas. Linking up as few as 4 of these farms resulted in a reduction of 87% in the variance of hourly output as compared to that associated with a single installation. Adding the remaining 16 facilities resulted in only a minimal reduction in the overall variance (8%). Kempton et al. [11], using 5 years of wind data from 11 meteorological stations distributed over 2500 km of the US East Coast, concluded that when outputs from an array of wind farms distributed along the coast were coupled, the output from the interconnected system was much more stable than that from any individual location. The correlation between individual station outputs decreased exponentially on a scale of 430 km as determined by properties of the related synoptic weather systems. Archer and Jacobson [12] considered the benefits of connecting wind farms from up to 19 sites in the mid west characterized by annually averaged wind speeds in excess of 6.9 m/s (class 3 or greater) at 80 m. They concluded that on average 30%, as much as 47%, of the connected output could be deployed as reliable base-load power. Hart and Jacobson [13] found that combining complementary renewable resources, such as wind, solar and hydro, can help mitigate the variability problems associated with any one of these options. Fertig et al. [14] reported that interconnecting wind plants on a large scale would reduce the most extreme hour-to-hour changes in energy output and increase the percentage of reliable power. Previous studies exploring the issue of interconnections focused on the statistical analysis of wind data and did not explicitly address the physical factors responsible for the observed variation of surface winds.

This study addresses the issue of interconnection with specific attention to the physical factors that determine the temporal variability of winds in the near surface region of the atmosphere. Surface winds are influenced by the passage of transient waves and by boundary layer turbulence triggered by these waves [15], [16], [17]. An understanding of these physical factors can help interpret the findings of the previous studies. We consider specifically how transient waves influence instantaneous power output. We show that there is a limit to the extent that the intrinsic variability of power output can be reduced, and quantify how this reduction in variability responds to different levels of wind farm coupling.

Section snippets

MERRA reanalysis data

This study was based on meteorological data from the Modern Era Retrospective-analysis for Research and Applications (MERRA) compilation covering the period Dec 2002 to Nov 2007. Boundary layer winds and geopotential heights included in this compilation were obtained on the basis of retrospective analysis of global meteorological data using Version 5.2.0 of the GEOS-5 DAS. Geopotential heights are available on a 3-hour basis with a resolution of 5/4° latitude by 5/4° longitude, while boundary

Examination of transient waves

The boundary layer wind, e.g. 100 m wind, as indicated earlier, is controlled by two factors: conditions in the free atmosphere which vary on a time scale of a few days with a spatial scale of about 1000 km, and conditions at the surface which are responsible for small scale and fast varying turbulence in the boundary layer [15], [16]. Fig. 2a shows the variation of the vertical profile of the flux kinetic energy per square meter, 1/2 ρ V3 (ρ represents the density of air at different layers of

Discussion

Fertig et al. [14] argued that fluctuations in wind power are not white noise, based on frequency-domain analysis. The high frequency variability of outputs from individual wind farms is determined, however, mainly by small scale boundary layer turbulence associated with local conditions; the low frequency variability is associated with the passage of transient waves with a characteristic time scale of several days. Fertig et al. [14] also concluded that the interconnection of wind plants

Acknowledgments

The work described here was supported by the National Science Foundation under grant ATM-1019134 to Harvard University. Junling Huang was also supported by the Harvard Graduate Consortium on Energy and Environment.

References (19)

There are more references available in the full text version of this article.

Cited by (0)

View full text