By Gerard Wynn
Winds are notoriously footloose and hard to predict, but for grid operators
keen to keep their customers’ lights on and hold down costs, agreeing on a
method to measure the reliability of the fast-growing electricity source is
Policymakers and operators around the world have come up with a range of
ways to estimate wind power and a bunch of terms to measure wind power
reliability, including capacity credit, capacity value and availability factor.
Capacity value is the proportion of a power plant’s installed capacity that
can be absolutely relied upon for security of supply.
For a coal plant that figure is around 85 percent but for wind can range
from zero to 30 percent, depending on which methodology you choose.
The range partly reflects actual differences across regions – for example
in wind speed, interconnection and the share of wind power in wider generation.
But it also reflects different methods of calculation, and highlights the
need for more consistency to avoid blackouts if wind power is
over-estimated or spending too much on reserve capacity if it is
Specific measurements are vital to gauge the proportion of installed
capacity that can be reasonably relied upon at any one time, whether during
predictable demand surges or unexpected events such as an unplanned nuclear
HERE’S THE MATH
Estimating capacity value, whether for variable renewable or conventional
coal and gas-fired capacity, provides a systematic way to measure security
The risk of not meeting demand – called “loss of load” – can be expressed
in various ways.
One standard term among grid planners is “Loss of Load Expectation” (LOLE).
According to a recent report by the British energy regulator, Ofgem**, LOLE
“represents the number of hours per year in which supply is expected to be
lower than demand under normal operation of the system. Importantly, this
is before any intervention by the System Operator, so does not represent
the likelihood of customer disconnections.”
Capacity values are key in estimating the impact of new power plants on
LOLE, pinpointing how much of the installed capacity can be relied upon.
In the case of wind, it is estimated to lie in the range of 0-30 percent of
installed capacity, compared with more than 80 percent for baseload
conventional, gas, coal, nuclear and hydro power.
Ofgem last month calculated the reliability of wind power capacity in its
six-year outlook for security of supply.
Britain in the near-term faces a greater risk of limited blackouts than
historically, as the country shuts down polluting coal and ageing gas plants.
As a result, Ofgem is interested in capacity values during periods of peak
demand, for example in January when heating and lighting needs are higher.
The regulator obtained local wind speeds from a re-analysis of NASA
satellite weather data – a standard academic procedure to work around a
lack of direct observations of wind power generation.
It converted wind speeds into power generation using observations from a
sample of actual wind farms.
It then generated probability distributions for wind power output during
peak demand annually through 2019, taking into account a doubling of wind
power capacity over the period.
It added expected conventional generation and electricity demand to
calculate what it terms the “Equivalent Firm Capacity” of wind power, which
it found to be in a range from 17-24 percent of installed capacity.
The range refers to different assumptions for the amount of installed wind
Perhaps paradoxically, the greater the share of wind power in the
generation mix the smaller the proportion which can be relied upon, because
on still days with no wind there will be a greater risk of a loss of load.
Uncertainty about how to measure wind power capacity values is a concern.
Even within Britain, Ofgem calculated a very different value over the next
five years (17-24 percent) than the country’s transmission operator
National Grid for the winter of 2011/12 (8 percent).
Ofgem says the National Grid’s calculation method – direct observation of
wind power generation – contrasts with its statistical modelling of all
supply and demand, and is inappropriate for assessing security of supply.
“The large difference in these numbers reflects two very different
approaches. The (National Grid’s) Winter Outlook approach is based on
observations of the output of wind at peak times. By its nature this is a
small number of observations, and it is therefore possible that the wind
output at the time of observation could have been very different.
“We do not consider this approach appropriate for a capacity adequacy
analysis, as it represents a pessimistic estimate of the availability of
wind in isolation from the rest of the system,” Ofgem said last month.
In Europe, estimates among transmission system operators for wind power
capacity value during peak demand vary from zero (in Austria, Cyprus and
Estonia) to up to 30 percent (in France and Portugal), according to the
European Network of Transmission System Operators for Electricity.
In the United States, grid planners projected wind capacity values during
peak demand in 2019 ***, ranging from 8 percent in the Midwest coordinating
region, to 18.5 percent in the western United States and Canada region,
called WECC (Western Electricity Coordinating Council). (Chart 2)
The U.S. Energy Information Administration (EIA) illustrated the importance
of getting capacity values right, with the example of a region projecting
20 gigawatts of wind capacity by 2019.
“If it decreased its capacity value by one percentage point from 12 percent
to 11 percent, and had to replace that lost wind capacity in order to meet
its target reserve margin, it would require an additional 200 megawatts of
If gas-fired power supplied the difference, that would cost $195 million in
upfront capital, EIA estimated in 2011.
But the range is up to 30 – not one – percentage points, underlining how
important it is to agree on a consistent modelling approach for a more
accurate balancing of demand and supply, to save costs and better ensure
** See Electricity Capacity Assessment Report 2013 at