IEEE Power & Energy Magazine - July/August 2018 - 87

as an operational tool to make sure sufficient resources are
available in energy markets to cover forecast errors or meet
expected and unexpected changes in net load or generation.
under this definition, most ancillary service products can be
categorized as near-term capacity markets.
in a highly electrified, highly variable electricity system,
all capacity markets will need to increase in sophistication.
ancillary service products must accommodate increased
forecast error, more extreme generator ramping, and certain
grid services provided today by conventional generation
but not replicated by renewables such as primary frequency
response. this is an area in which rtos have been active in
recent years, and much progress has been made.
the changes necessary for long-term capacity markets
are not as well understood, and debate still revolves around
exactly what capabilities are necessary for future capacity
and how much will be required. it is often assumed that a
combination of wind, solar and 6-24 h of energy storage
will be sufficient to create a reliable low-cost decarbonized electricity system, but this does not stand up to serious scrutiny.
to illustrate these concepts, a new type of screening
curve is presented here for systems with large amounts of
inflexible supply that demonstrates the capacity required
from balancing resources of different durations. figure 8
shows the residual capacity over the modeled year 2050 for
two cases, the u.s. department of Energy's 2017 annual

Energy outlook and the high-renewables scenario from
the u.s. ddpp. residual capacity is calculated as gross
load minus all zero-carbon resources (excluding bioenergy) and includes the impact of all flexible loads and energy storage.
this residual is instructive because it is the energy that must
be met with either thermal resources or a greater deployment
of the zero carbon resources already in the system. the key
question is what the characteristics for other resources must be
to replace thermal and reduce emissions or cost. this question
is approached by analyzing the residual capacity requirement
to understand the number of hours a given capacity must be
maintained to ensure a reliable system. this result is shown
in figure 9. in the high-renewables case, capacity resources
had to be able to provide at least 422 GW for 1 h, at least
376 GW for 10 h, and at least 141 GW for 100 h. the flatter
these screening curves, the higher the likely capacity factors
for residual resources, and the longer the duration, the more
difficult it is for energy-constrained balancing resources, such
as battery storage, to meet the need.
the most striking aspect of the curve is how little the
shape of the curve changes between the two cases. despite
huge quantities of renewable generation, residual capacity
needs with significant duration remain a necessity for reliable
system operation. one important caveat is that this screening curve does not consider the need for long-duration storage resources to charge as well as discharge. thus, a storage

Hour of the Year

Scenario

AEO 2017
Reference

Residual GWh

600

400

200

High
Renewables

Residual GWh

0
600

400

200

figure 8. The residual capacity need across simulated 2050 scenarios. Residual capacity is equal to total load minus all
zero-carbon resources, including nuclear and hydroelectricity. In the modeled scenarios, this residual was met with thermal resources.
july/august 2018

ieee power & energy magazine

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - July/August 2018