Sioshansi F.P. Smart Grid: Integrating Renewable, Distributed & Efficient Energy
Подождите немного. Документ загружается.
Figure 6.2
Ranges of power system frequency during normal operations and following the
sudden loss of generation.
Source: Use of Frequency Response Metrics to Assess the Planning and Operating Requirements for
Reliable Integration of Variable Renewable Generation, December 2010, Ernest Orlando Lawrence
Berkeley National Laboratory, LBNL-4142E
The mechanisms for maintaining system balance are best
understood in terms of three time frames for the system to
respond to disturbances or imbalances. The first and most
critical time frame is what occurs in the first 30 seconds after
a disturbance in which a large amount of generation is
removed. In this time period inertia and governor response are
the primary frequency control mechanisms deployed to arrest
frequency deviation. The second time period is what happens
between 30 seconds to ten minutes. In this time frame
secondary frequency controls such as regulation are deployed
to control frequency. For the third period, starting at five
minutes, economic dispatch of resources through the
five-minute real-time market provides the tertiary control
mechanism deployed to maintain balance between supply and
demand based on forecasted conditions.
331
While maintaining balance under normal conditions is a
challenge, maintaining balance after a disturbance or a
contingency in which a large resource or group of resources
trip off-line simultaneously can have more far reaching
reliability impacts. Figure 6.3 illustrates how primary,
secondary, and tertiary frequency controls act to restore
frequency to its normal operating level after a disturbance.
After a disturbance, primary frequency response measures are
designed to deploy and automatically arrest frequency. While
frequency is decaying, demand actually increases as motor
load increases as an inverse function of frequency. Primary
frequency controls such as inertia and governor control will
automatically activate to first arrest frequency decay and then
recover frequency. Inertia is the result of large spinning
masses continuing to spin, creating a resistance to the system
slowing down due to the imbalance. Thus inertia becomes a
shock absorber or resistance to change in frequency after a
disturbance.
332
Figure 6.3
The sequential actions of primary, secondary, and tertiary frequency controls
following the loss of generation, and their impacts on system frequency.
Source: Use of Frequency Response Metrics to Assess the Planning and Operating Requirements for
Reliable Integration of Variable Renewable Generation, December 2010, Ernest Orlando Lawrence
Berkeley National Laboratory, LBNL-4142E
Governor control on the other hand is an independent control
mechanism designed into generators that automatically
increases the output of the generators as the frequency begins
to drop. The more frequency drops, the more the generator
increases its output up to the resource's maximum output
capability or fuel supply availability. For example, a steam
turbine whose governor responds due to a frequency
disturbance will open up the steam valves and release steam
333
into the turbines, increasing the generator's output. If the
steam pressure in the turbine is not built up sufficiently,
however, the steam output may only be able to increase the
generator's output for a limited amount of energy or for a
limited time. Moreover, not all generators have governors,
and in some cases generators with governors may have their
governors blocked and will not respond to a frequency drop.
This is where smart grid technologies such as synchrophasors
offer tremendous promise for improving the grid operator's
visibility to overstressed operating conditions and thereby
enabling the operators to take timely corrective actions.
The challenge of arresting frequency decay after a disturbance
event increases with high penetration of renewable resources
such as wind and solar photovoltaic, as these resource types
are unable to provide frequency control services. In the case
of wind, there is generally little inertia response resulting
from the spinning generators and no governor response to
increase output in response to a disturbance event. In the case
of solar photovoltaic there is no spinning mass providing
inertia and no governor response. Any response would be
from the inverter converting the DC power from the solar
panels to the AC current of the grid, and currently inverters
are not programmed to provide any response. As technologies
develop it may be possible for inverters to assist with
frequency response, but this will require implementation of
smart grid monitoring and communication capabilities. To the
extent that solar photovoltaic resources are distributed
resources, the communication and control issues are even
more complicated. In the case of solar thermal there is some
inertia and potentially some governor response as well, since
a solar thermal plant has converted solar energy into steam
that turns a steam turbine.
334
The above discussion explains a primary operational concern
with higher penetration levels of variable energy resources
such as wind and solar photovoltaic. To the extent these
resources displace conventional resources on the transmission
system, there can be a significant loss of the inertia and
governor response needed to recover from frequency
disturbance events. If the system does not have sufficient
amounts of these primary frequency response services, the
system may not be able to adequately arrest frequency after a
disturbance, requiring the system operator to resort to the
undesirable backstop of shedding load.
Maintaining system balance under normal non-disturbance
events is the responsibility of the secondary and tertiary
frequency control measures. Automatic generation control
(AGC) is the first form of the secondary frequency control.
AGC is a centralized control system that monitors the
frequency within the balancing authority area and the actual
interchange transfers between it and its adjoining neighbors.
This measurement is called the area control error (ACE). If
either frequency or interchange deviates from expected or
scheduled levels ACE will deviate from a balanced zero
condition, and AGC will detect the ACE deviation and send
dispatch signals to generators to raise or lower energy output
depending on the direction of the imbalance. A resource that
is certified capable and agrees to be available to respond to
AGC signals is said to provide regulation service. The CAISO
procures regulation service from certified resources that offer
to provide the service on an hourly basis through its
day-ahead and real-time market processes. Every four
seconds, the AGC system monitors for ACE deviations and
sends dispatch signals to those resources providing regulation
service.
335
Tertiary control for balancing the system is performed by
committing and dispatching resources over a longer interval
to meet the expected forecast of demand, taking into account
any forecast deviations of supply. The CAISO performs this
balancing function through its real-time economic dispatch
market, which can commit short-start and quick-start
resources every 15 minutes and issue dispatch instructions to
resources every 5 minutes. This tertiary control is also
sometimes referred to as real-time balancing or load
following. The latter term, though used traditionally to refer
to real-time balancing, is no longer really accurate in the
context of large amounts of variable renewable resources,
however, as the real-time dispatch will actually follow load
net of variable resource production.
Figure 6.4 illustrates the relationship between the secondary
and tertiary mechanisms—regulation and load following,
respectively—in balancing the system. Load following entails
the use of five-minute dispatch instructions to make up the
difference between the hourly schedule and the average net
load forecast for each five-minute interval. Regulation
provides the more granular adjustments necessary to meet the
difference between the response of the dispatched resources
to the CAISO's instructions and the actual net load as it varies
within each five-minute interval.
336
Figure 6.4
Regulation and load following.
Source: California ISO, Integration of Renewable Resources: Operational Requirements and
Generation Fleet Capability at 20% RPS, August 31, 2010
Increased levels of variable renewable generation raise yet
another complication for system balancing. The process
described above for arresting the frequency decay when a
contingency event occurs—when a generation resource is
lost, for example—is supplemented by the dispatch of
contingency reserves and the use of emergency ratings of grid
facilities, which help the system respond to disturbances. The
potentially large swings in the output of intermittent
renewable resources are not classified as contingencies,
however, so contingency reserves and emergency ratings are
not available to the grid operator, even though such swings
may be as large as the loss of a generation unit. This means
that the system must be able to respond to these changes
without using resources designated as contingency reserves or
337
any emergency line ratings, and may even have to respond to
a contingency event while it is responding to the
intermittency of the renewable generation.
Describing the primary, secondary, and tertiary balancing
mechanisms this way helps to convey how the variability and
uncertainty of renewable resources can challenge grid
operations. Before the variability and uncertainty associated
with renewable resources became a concern, supply resources
were all highly controllable and predictable, and therefore the
largest contributor to real-time imbalance was the variability
and uncertainty of the load. With increasing penetration of
renewable resources, the supply side of the operational
balance is also variable and uncertain, with supply variation
offsetting load variation in some instances and adding to it in
others in a manner that is not easy to predict.
For example, during the morning when load is naturally
increasing, wind production in California is typically
decreasing, thereby increasing the need for capacity that the
CAISO can dispatch up to follow net load. At the same time,
solar production will typically increase as the morning sun
rises and can somewhat offset decreasing wind production,
although there can be a temporal gap between the fall-off of
wind production and the upswing of solar. Obviously, the
complementary problem can occur during the evening ramp,
the exact nature of which tends to vary with weather
conditions and the daylight savings time regime.
Other operational challenges expected to increase with more
renewable generation on the grid include voltage control and
congestion management. Conventional resources are required
to be capable of providing a certain amount of reactive power
or voltage control, but currently no such requirements exist
338
for wind and solar resources. In order to impose such
requirements the CAISO will need to establish the operational
need, which the CAISO is currently assessing through its
integration studies. Until such requirements are imposed on
interconnecting variable renewable generation, the CAISO
expects the operational challenges of maintaining voltage and
avoiding cascading voltage collapse conditions to increase.
Congestion occurs when there is insufficient transmission
capacity to support total transfer of energy from areas with
supply to areas with demand. Although congestion is not a
new phenomenon created by renewable resources, it is
expected to become more challenging with the increase in
variable renewable generation.
• First, the frequency and magnitude of congestion will
increase in areas high in wind or solar development
potential to the extent renewable resources are allowed to
interconnect to the grid prior to the completion of sufficient
transmission upgrades to transfer all their energy output to
demand.
7
7
This highlights a further complication that the CAISO is addressing
through its initiatives on transmission planning and generation
interconnection. The variable nature of renewable generation means
that these resources will have lower transmission utilization rates than
conventional generators. A conventional generator with a capacity of
500 MW will likely use a 500 MW capacity transmission line often,
whereas a wind or solar generator with a 500 MW capacity will likely
use the full 500 MW infrequently. The planning and interconnection
processes need different approaches for determining the most
cost-effective way to build transmission to wind and solar regions
than have been used traditionally for meeting the transmission needs
of conventional resources.
339
• Second, the high level of wind output during off-peak
hours means that there will often be insufficient
dispatchable conventional resources on line to manage
congestion.
• Finally, the potential for large rapid swings in output from
wind and solar resources in response to changing weather
conditions can create situations where congested areas
quickly switch from having too much generation to too
little generation on-line. In these cases the grid operators
may have to take manual or out-of-market actions to
resolve congestion, including curtailment of the renewable
resources.
While operationally necessary at times, curtailment of
renewable resources is counter to the objective of the
renewables portfolio standard to maximize the amount of
energy supplied by renewable resources. As described in a
number of chapters in this volume, smart grid technologies,
particularly in conjunction with storage facilities participating
in the CAISO market, will enable more efficient utilization of
the grid in achieving the state's renewable energy goals; for
example, by storing renewable energy when transmission is
congested or there is excess supply and delivering it at times
when the energy can be utilized.
These real-time operational challenges have a planning aspect
as well, as the CAISO must consider whether the generation
fleet will be capable of managing the variability of the new
renewable resources in each year of the planning horizon,
given the current uncertainty about future retirement,
repowering, and construction of new dispatchable resources.
State policy has introduced another issue here. California has
determined that power plants can no longer rely on a cooling
340