Renewable Energy: Coping with Variability
Part 1: Supply Management
Dec 21, 2006 Roger Arnold, energypulse.net
One of the biggest issues with solar and wind power
is their variability. They produce power "when they
want to" and not necessarily when we would like them
to. There are ways to cope with this variability,
but each has some economic cost. In this three-part
article, we review current options, and suggest likely
developments for the near future.
Impact of Variability
Variability is an issue for both solar and wind,
but it affects the two sources differently. Since
availability of solar correlates reasonably well with
periods of peak power demand, its impact is usually
taken to be positive. For that reason, weâll set
it aside for now, and begin by looking at how variability
affects the economics of wind power.
Opponents argue that wind power never adds any generating
capacity to a regional power grid. The argument is
that, because wind is not reliable, the system must
always have enough capacity in other forms to meet
its highest peak demand. Otherwise, it risks rolling
blackouts during calm periods with high demand. Hence,
wind power, when it is available, merely displaces
other generating capacity that must still exist. It
conserves fuel and reduces CO2 emissions, but its
"true" economic value is limited to the
cost of the fuel its operation displaces. For a coal-fired
generating plant, in the absence of a CO2 tax, that's
a meager 1.5 cents per kilowatt-hour, not sufficient
to justify the capital cost of wind turbines under
the usual financial models.
There is some validity to that argument. However,
it mistakes the nature of wind resources and the power
grid. "Capacity" is not a hard-edged number, and in
any case the purpose of integrating wind turbines
into a system is not mainly to increase its maximum
power capacity rating. Its purpose is to reduce the
consumption of fossil fuels. Nonetheless, under the
right conditions, the energy contribution of wind
resources can have utility well above the marginal
cost of fuel saved. Part of the trick to integrating
wind resources involves strategies that enable the
"right conditions" to apply more often.
There are just three basic mechanisms for coping
with variability. One is backing generation (supply
management), another is load management, and the third
is energy storage. In this part, we'll focus
mainly on supply management. Later parts will look
more at load management and energy storage.
Load Balancing Today
Today, balancing is accomplished almost entirely
via supply management. Load management, in the form
of "demand response", is of growing interest,
but so far it has mostly been limited to emergency
curtailment of large loads during power crisis conditions.
The vehicle for load balancing is the grid within
a "regional balancing area", or RBA. The typical
RBA incorporates many individual generating units
of different types. They range from advanced units
with high capital cost but very low marginal cost
for energy generated (e.g., nuclear), to simple units
whose marginal operating costs are high, but which
don't tie up much capital when sitting idle. The former
are preferred for meeting baseload demand, while the
latter are emergency backup units and "peakers". In
between are units whose marginal operating costs are
reasonably low, and whose designs allow them to be
cycled on a daily basis without undue stress. These
units are started up or shut down as needed to follow
the daily load profile.
The ideal resource for load following is a hydroelectric
plant. A suitable plant has multiple hydro turbines
and at least a small receiving reservoir to buffer
downstream river flow. Those features allow its power
output to vary widely, according to need. Its long-term
energy output is fixed by stream flow, but there is
a lot of flexibility as to when it is generated. That
makes this type of hydro a perfect complement for
wind power. Power supplied by the wind, when it is
blowing, replaces water flow through the hydro turbines.
The water retained in the reservoir remains available
to supply power when the wind is not blowing. This
is one of the conditions in which the economic utility
of energy from variable wind resources is fully equal
to that from regular power sources.
When a suitable hydro-electric plant is not available
for load-following, then coal-fired plants with multiple
turbine-generator units are the next best choice.
Typically, groups of individual units share boilers
and condensers. They are usually scheduled to keep
at least one of the units in a group operating, so
the boiler and condenser avoid stressful thermal cycling.
Under normal circumstances, daily electrical demand
is met entirely using baseload and dispatchable intermediate
units. If it becomes necessary to draw on less efficient
peakers and backup units on a regular basis, then
it's time for utility planners to start thinking about
adding more baseload and dispatchable intermediate
capacity, or promoting energy efficiency to reduce
It's in this context that the economics of wind
power must be considered. Wind turbines have very
low marginal operating costs. When wind energy is
available, it pays to use it. That can usually be
accommodated by juggling the schedule of start-ups
and shut-downs of existing intermediate units. The
process is no different than that used to meet variable
demand over the course of a day. However, it's less
predictable a day in advance, which complicates life
for the transmission system operator (TSO).
Economics of Wind Energy Today
The industry consensus seems to be that in most
RBAs, if the level of wind penetration is below 20%
of average demand, then the variability can be accommodated
without building new backing capacity.  Although
the peak "in-feed" from a wind farm during periods
of high wind can be four times its average value,
there is usually enough intermediate capacity that
can be temporarily shut down to allow that level of
in-feed to be accepted. Conversely, when in-feed from
wind resources is low, the level of intermediate capacity
that is already installed for following the daily
load profile will usually be sufficient to take up
the slack for low wind in-feed. Occasionally, during
periods of unusually high demand and low wind, it
will be necessary to activate back-up units or invoke
demand response measures. That should be rare enough,
however, as to have only minor impact on operating
On the other hand, the inability to predict, on
a daily basis, just when power from dispatchable load-following
units will be needed can have a major impact on expenses
for a TSO. It may limit the TSO's ability to purchase
low-priced power on long-term contracts, and force
it to turn more to the high-priced spot market. That
happened, for example, to NorthWest Energy in Montana
when the Judith Gap wind project came on line.
When power must be purchased on the spot market,
it's usually good news for the owners of dispatchable
units, but bad news for ratepayers. Even if the TSO
is purchasing wind energy at rates that are low compared
to conventional sources, the higher priced spot purchases
can quickly offset any savings. As a result, the cost
of power to ratepayers goes up.
Strategies to mitigate the impact of spot market
purchases exist. They include long-term contracts
formulated to allow more flexibility to schedule power
delivery on short notice, or acquisition by the TSO
of captive load-following resources that it can draw
on for short-notice scheduling around wind availability.
However, there is one effect that can't be mitigated
by any supply-management strategy alone: an inherent
reduction in average capacity factor for the system's
The whole point of wind energy, after all, is to
displace generation that would otherwise be supplied
by dispatchable intermediate units and peakers. Fuel
consumption and CO2 emissions are reduced, but the
units are still needed for meeting peak demand. They
simply operate with lower average capacity factors.
That means that the non-fuel portion of their power
costs are amortized over fewer kilowatt-hours delivered,
raising the average cost of power.
There is a counter-effect by which wind helps to
reduce the cost of power to ratepayers. It's a hard
effect to quantify, however, and not that easy even
to explain. But I'll try.
Consequences of Fuel Saving
In the case of paired hydroelectric and wind power,
what gives wind energy its high utility is that the
fuel supply, for hydroelectric power, i.e.,
stream flow,is fixed. If more power is needed, a
hydroelectric utility can't just go out and purchase
more stream flow. But every kilowatt-hour of energy
that can be supplied by wind is a kilowatt-hour of
deferred hydro energy that can be supplied later.
If the supply of fossil fuel available to generators
within an RBA were fixed by rationing (or perhaps
by a carbon cap?) the same situation would apply.
The restricted fuel supply changes the system from
being power-limited to being energy-limited. Wind
generation and fossil-fueled generation then trade
off in the same way they do for wind and hydro power.
The variability of the wind resource becomes irrelevant,
so long as it is paired with sufficient backing generation.
At the present time, fuel supplies aren't rigidly
fixed. At least, not at the level of an RBA. Yet something
close to that situation does exist at the national
level. Supplies of natural gas are tight, with very
little elasticity. If the bid price rises, it may
prompt suppliers to sell gas from storage, but it
doesn't directly lead to higher annual gas production.
So a utility that buys gas for power generation is
ultimately buying it at auction against other would-be
gas users. To succeed, some other would-be user must
be priced out of the market.
In that situation, the reduction in fuel demand
from use of wind energy translates to a reduced market
price for fuel. The wind resource should technically
be credited with the fuel price delta for which it
is responsible, applied across all fuel purchased.
That figure can be much larger than the direct cost
of the fuel saved. However, it's diffuse, and can't
be measured directly. It can only be estimated. Whatâs
worse, it's firmly enmeshed with that most troublesome
economic notion of "the common good".
The major benefit of reduced fuel consumption for
power generation accrues not to those who paid for
construction of the wind resource, nor to the utility
that purchases its output. Rather, the benefit is
to the community of fuel users as a whole, in the
form of lower fuel prices. But there is no way for
the owners of a wind resource or those purchasing
its output to capture that benefit; reduced fuel prices
simply become their indirect "gift" to the community.
This is an example of why government subsidies can
be legitimate instruments of rational policy for the
public benefit. As much as free-market fundamentalists
may rail against them, subsidies can serve to motivate
beneficial behavior that the market alone has no means
At this point, those paying close attention will
notice that I have just argued, in effect, that construction
of wind farms will not significantly reduce total
consumption of natural gas. It will, instead, reduce
the price of natural gas, and enable uses that would
otherwise have been priced out of the market to take
up the slack. Have I just undercut the entire green
rationale for wind power?
Well, yes and no. Natural gas is still a much âgreenerâ
fuel than coal, and it's likely that most of the
additional gas usage that wind power will enable would
otherwise be served by coal. Since coal is not tightly
supply-limited, the net effect should be a reduction
in coal usage. On the other hand, the lower fuel prices
will reduce incentives for efficiency improvements.
Since efficiency improvements are unquestionably the
best long-term strategy we have for reducing our "ecological
footprint", that would be bad.
The conclusion I would draw from that, however,
is not the paradoxical suggestion that wind power
is actually an impediment to reduced consumption of
fossil fuels. My conclusion is much more pedestrian:
that the most effective policy for achieving reduced
CO2 emissions will be to tax CO2 emissions, rather
than subsidizing wind power or other non-carbon energy
sources. Surprise! The point goes to the anti-subsidy
free market crowd, after all.
Limitations of Supply Management
There are a number of important issues regarding
supply management that I did not discuss above. They
include details about the shape of the supply curve
for wind farm output, and the implications of long
distance power transmission for supply management.
These are important issues, and those interested can
read more about them in some of the references given
below. However, the most important points to take
from this discussion of supply management can be summarized
- At low levels of wind penetration, the plot of
daily demand less wind farm output is qualitatively
very similar to the plot of daily demand alone.
Any RBA that has dispatchable resources sufficient
to deal with the latter should also be able to deal
with the former.
- I.e., up to a certain level, variable wind power
can be integrated into the power supply system with
no need to add new balancing capacity. That level
will depend on the particular characteristics of
a given RBA, but is generally considered to be about
20% of average load.
- With wind shouldering a variable portion of the
load, long range forecasts of required supply from
other sources become less reliable. That can result
in higher operating costs for the TSO if it does
not control its own generating resources. A smoothly
functioning hour-ahead market is needed to mitigate
uncertainties introduced by wind supply.
- At higher levels of wind penetration (e.g., those
contemplated for much of Europe) existing mechanisms
for supply management become insufficient. At that
point, reliance on an exclusive strategy of supply
management becomes expensive, as added wind capacity
must be balanced by added balancing capacity.
A more efficient strategy for coping with variability
at high levels of wind penetration is to shift toward
load management and energy storage. "Load management",
in that case, does not mean (only) load curtailment
to reduce peak demand, but (more importantly) time
shifting of discretionary loads to match available
supply. We'll look at that in some detail next week
in part II.
Notes and References
 Long distance transmission could be considered
a fourth basic mechanism, but I prefer to view it
as a way to extend the scope of the other three mechanisms.
 Here "best" is meant from a technical and
business economic viewpoint; environmental considerations
are another matter.