Okinawa Institute of
Science and Technology Graduate University - OIST
The amount of energy
generated by renewables fluctuates depending on the natural variability of
resources at any given time. The sun isn't always shining, nor is the wind
always blowing, so traditional power plants must be kept running, ready to fill
the energy gap at a moment's notice.
Because the grid has no storage, and
unlike coal or nuclear, there is no control over the fluctuating production of
renewable energy, the energy they produce has to be consumed straight away, or
risk collapsing the electrical grid.
On particularly windy days, for example,
surges in power generated by wind turbines have been known to overwhelm the
electrical grid, causing power outages.
To avoid this, operators of large power
plants sometimes resort to paying consumers to use electricity on particularly
sunny and windy days when there is too much excess power in the system, in
order to balance the supply and demand of energy at the grid.
Dealing with the peaks
and troughs of intermittent renewable energy will become increasingly
challenging as governments try to phase out of more stable coal-powered energy
sources in the coming decades.
In order to mitigate or manage these fluctuations in renewable energy, we need to understand the nature of these fluctuations better. Professor Mahesh Bandi, head of the Collective Interactions Unit at the Okinawa Institute of Science and Technology Graduate University (OIST) has used turbulence theory combined with experimental wind plant data to explain the statistical nature of wind power fluctuations in a single-author paper published in Physical Review Letters.
Wind speed patterns
can be depicted as a wind speed spectrum on a graph. In 1941, Russian physicist
Andrei Kolmogorov worked out the spectrum of wind speed fluctuations.
Subsequently, it was shown that the spectrum for wind power follows the exact
same pattern.
However, until now, it was simply assumed that these spectra were
identical due to the relationship between power and speed, where power equals
wind speed cubed. But this proved to be a red herring. Professor Bandi has
shown for the first time that the spectrum of wind power fluctuations follows
the same pattern as wind speed fluctuations for a different reason.
Kolmogorov's 1941
result applies to measurements of wind speed made at several distributed points
in space at the same time. But wind power fluctuations at a turbine are
measured at a fixed location over an extended time period.
The two measurements
are fundamentally different, and by carefully accounting for this difference,
Professor Bandi was able to explain the spectrum of wind power fluctuations for
an individual turbine.
We can think of
turbulence as a ball of air, or an 'eddy', of fluctuating wind speed. Long
time-scale, low frequency eddies can span hundreds of kilometres. Inside these
large eddies are shorter time-scale, high frequency eddies that might span a
few kilometres.
Therefore, if all of the turbines in the same wind plant fall
within the same short and long time-scale eddies, the energy they produce
fluctuates as if the entire plant were one giant turbine. This is exactly what
Professor Bandi found when he looked at the wind power fluctuations of all of
the turbines in a wind plant in Texas.
In fact, even
geographically dispersed wind plants can exhibit correlated fluctuations in
power if they fall within the same short and long time-scale eddies.
However,
as the distance between wind plants increases, their power fluctuations start
to decouple from each other. Two geographically dispersed wind plants might encounter
the same long time-scale wind speed fluctuations whilst encountering completely
distinct shorter time-scale wind speed fluctuations.
In the past, some
scientists have underestimated the problem of turbulence, arguing that the
power produced by geographically dispersed wind turbines in windy and calm
locations at any one point in time will average out when they reach a
centralised grid.
However, Professor Bandi's findings show for the first time,
that this phenomenon, known as 'geographic smoothing', only works to a certain
extent.
The power generated by
geographically dispersed turbine plants averages at high frequencies, because
while one plant might fall within the short time-scale eddy, the other might
not. In other words, the surge in power output at one plant is averaged out by
a trough in power output from another, far-away plant at high frequencies.
But
because the plants still fall within the same long time-scale eddy, the power
they produce will have correlated fluctuations at low frequencies, which
generate the most power. A surge in power at one wind turbine plant will
coincide with the surge at a far-away plant within the same long time-scale
eddy, meaning that the power they feed to the grid cannot be averaged out.
This
means that there is a natural limit to how much one can average fluctuations in
wind power; a limit beyond which fluctuations can continue to wreak havoc on
the grid. Using data from 20 wind plants in Texas and 224 wind farms in Ireland
Professor Bandi showed that this limit exists in reality.
"Understanding
the nature of fluctuations in wind turbine power has immediate implications for
economic and political decision making," says Professor Bandi.
Due to the variability
of renewables, coal-fired power plants providing back-up energy are kept
running in case of sudden power outages, meaning that more energy is produced
than needed.
This means that 'green' energy is still contributing to carbon
emissions, and there is an associated cost of maintaining reserve energy, that
will only increase as the proportion of renewables increases in the years to
come.
The discovery of a limit in geographical smoothing, articulated by
Professor Bandi, will enable better estimates of the operative amount of
reserves that needs to be maintained.
This discovery will
also impact environmental policy. By considering the limit for averaging
fluctuations of power, combined with the availability of different renewable
resources such as sun, wind and waves in a particular area, policy-makers will
be better equipped to work out optimal combinations of different energy sources
for specific regions
"Understanding
the nature of fluctuations for wind turbines could also open up other avenues
of research in other fluctuating systems," says Professor Bandi.