Sioshansi F.P. Smart Grid: Integrating Renewable, Distributed & Efficient Energy
Подождите немного. Документ загружается.
(DLC) DR can become the glue of a decarbonizing grid by firming VERs in ancillary service markets
This chapter categorizes DR's ability to increase the grid's flexibility through a topological comparison
of wind integration costs and the capacity potential of specific loads on various ancillary timescales
Renewable energy integration, direct load control, demand response
Introduction
As described in other chapters of this volume, the smart grid
is a combination of enabling technologies that span the
structural topology of the power grid. Its mandate includes a
broad range of societal benefits that broadly encompass the
notion of grid reliability, security, versatility, and resilience.
While this points to a large class of technologies that can be
leveraged toward a broad range of applications, from the
point of view of carbon emission reductions, renewables
integration is of particular interest. The carbon reductions
attributed to smart grid deployment can be attributed to four
broad categories:
• Increased deployment of renewable energy through
increased grid flexibility.
• Decreased carbon intensity of energy goods and services
through energy efficiency.
• Reduced electric consumption in response to market
signals (conservation).
• Displacement of petroleum consumption through vehicle
electrification.
The impact of the smart grid on the cost of integrating
renewable energy is often described in qualitative terms, but
rarely quantified in a rigorous way. Despite the fact that
enabling variable energy resources (VERs) integration is
often described as the largest source of carbon emission
abatement from the smart grid (Figure 9.1), this quantity is
usually estimated in terms of an arbitrary increase in
451
renewables deployment [1] or by estimating incremental
revenue diverted to other sources of abatement as a proxy [2].
However, these approaches do not address the mechanisms by
which smart grid technology will facilitate increased
deployment of renewables.
Figure 9.1
Power sector greenhouse gas (GHG) emissions reductions from smart grid
deployment.
Source: Pratt et al. [2]
The purpose of this chapter is to look at the role of load as a
source of flexibility that can facilitate increased deployment
of variable generation from renewable sources. How do VERs
affect reliability and how can those effects be quantified in
ancillary services markets? What types of DR exist, and
which are better suited to address the needs of renewables
452
integration? Finally, is the DR resource big enough to make a
dent in the demands that wind and solar energy will place on
the system?
In Chapter 10 Hanser et al. focus on the related issue of how
to determine the incremental demand for ancillary services
stemming from VER deployment and propose a methodology
for estimating the optimal sizing of critical peak pricing
programs to serve that need.
In this chapter we focus on the supply side of renewables
integration, the full portfolio of DR resources, and the size of
the resource as it relates to specific requirements of the grid.
We begin with a description of the variable resource
integration challenge and a short description of grid support
services at various characteristic timescales. We then identify
various classes of VER integration costs. In the following
section, we develop a topology of DR resources that provides
the basis for analyzing the ancillary service requirements
relative to DR resource potential. Finally, we discuss DR
deployment potential, identify the resources best suited for
renewables integration, and quantify their potential relative to
the demands of integrating large quantities of variable
generation onto the grid.
Integrating Temporally Variable and
Geographically Heterogeneous Resources
Implicit within high renewable generation scenarios is the
heterogeneous geographic distribution of variable energy
resource (VERs) such as wind and solar. Economics drive
deployment of utility-scale renewable technologies to regions
with the highest quality wind and solar resources. The
geographic constraints of resource distribution and the
453
temporal constraints of resources variability pose unique
challenges to reliable system operation as the share of
generation from renewables grows. The path to the grid's
decarbonization invariably flows through the confluence of
grid reliability and efficient capital outlay. In that context, a
flexible yet resilient power grid will be a key enabler of
cost-effective renewables integration and rapid carbon
reductions.
The Eastern Wind Integration and Transmission Study
(EWITS) exemplifies this type of regional heterogeneity in its
analysis of 20% and 30% wind penetration scenarios for the
Eastern United States [3]. Achieving these penetration
milestones by the year 2024 represents an aggressive
trajectory for renewable energy deployment and requires
significant new infrastructure upgrades. A comparison of
regional peak wind capacities with regional peak loads is
shown in Figure 9.2. It is notable that certain regions such as
Southwest Power Pool (SPP) are projected to install new
wind capacity that exceeds total current system peak demand
in order to reach a penetration of 20% across the East.
454
Figure 9.2
Regional wind capacities in EWITS scenarios relative to 2006 peak load by
region.
Source: NREL [3]
Because a large fraction of the generation that wind capacity
produces will necessarily be consumed outside the region,
such high-penetration scenarios imply a strong role for
transmission and other flexibility resources to manage
resource variability. The take-away is that in scenarios aimed
at large-scale deployment of renewables, VERs are likely to
develop in regionally clustered pockets in order to access the
highest quality, most economic resources.
An analysis of wind capacity from the North American
Electric Reliability Corporation (NERC) Long Term
Reliability Assessment (LTRA) illustrates a similar trend [4],
with relative wind capacities highest in those areas with the
strongest wind resource (Figure 9.3). This wind capacity
455
represents actual projects that are in the interconnection
queue, as opposed to the hypothetical projections developed
in EWITS. The LTRA data translate to 26% wind relative
wind capacity by 2018. Figure 9.2 and Figure 9.3 demonstrate
how resource heterogeneity drives uneven development of the
resources. In high wind penetration scenarios, this type of
development trend can amplify the operation impacts of
integrating variable generation.
Figure 9.3
NERC region 2018 wind capacities in 2009 LTRA relative to 2018 peak load
by region.
Source: NERC [4]
The challenges associated with resource distribution and grid
integration described above for the Eastern United States are
mirrored elsewhere, including the Western United States,
Europe, and China [5], [6] and [7]. Therefore, while the focus
456
of this analysis will be on the demand-side flexibility
resources in the United States, the operational issues
addressed here will have strong relevance for all the regions
of the world currently engaged in large-scale VER
deployment.
Although the output of wind and solar generators cannot be
controlled directly, the system's ability to accommodate the
variability inherent in load is often sufficient to address the
impact of VERs at modest penetration levels. However, as the
quantity of variable generation grows, the operational impact
of VERs will increase. Combining weakly correlated
resources can help smooth out the fluctuations, but when the
deployment of VERs is highly clustered as indicated above, it
will be more difficult to leverage this type of geographic
diversity. In this case the variations in wind output will
instead be highly correlated in regions with the greatest
density of high-quality resources, making resource integration
that much more challenging.
Reliably integrating large quantities of variable generation
will ultimately require a portfolio of solutions that facilitate
resource sharing across regions and enhance system
flexibility. In the long term, building high-capacity
transmission lines that allow for increased power flow will
ameliorate the supply and demand imbalances associated with
the geographic heterogeneity of relative wind capacities. In
the short and long term, a portfolio of reliability-focused tools
will be needed to firm wind capacity on a variety of temporal
scales. Aggregating balancing authorities, high-fidelity wind
forecasting, utility-scale storage facilities, new transmission,
flexible natural gas and hydroelectric generation, and demand
response (DR) all augment the grid's flexibility. Quantifying
457
the relative value of each flexibility resource in a flexibility
supply curve [8] would enable an economically efficient path
to significant VER integration.
Direct load control (DLC) DR can be an important
component of an integration portfolio. The dispatchable
demand-side flexibility that DLC provides can increase the
grid's reliability and ensure meaningful carbon reductions
through the accelerated low-cost integration of VERs.
Demand Response: Past, Present, and Future
A historical perspective of DR is necessary to properly
understand its future potential as an important component in
the portfolio of VER integration solutions. Historically,
system operators used manual load shedding to maintain grid
reliability during emergency situations. Although such blunt
DR measures are necessary to prevent catastrophic grid
failure, they are not the best way to protect the integrity of the
system. Manual load shedding does not represent Pareto
optimality. Automatic load shedding caused by
under-frequency and under-voltage switches are in some
ways a brute-force form of DR as well.
Today, more sophisticated DR programs differ from historical
demand response in that utility customers are given the choice
to shut down during times of peak demand and are
incentivized to do so accordingly. The incentive structure,
driven by the lucrative economics of peaking generation,
provides participating customers with payments exceeding
their operational value for short durations a few times a year.
Current DR programs represent a Pareto improvement
because both utilities and customers profit from the exchange.
458
For example, Baltimore Gas and Electric Company estimates
that the capital cost of DR, at US$165/kW, is three to four
times cheaper than the cost of installing new peaking
generation, which is around US$600–800/kW [9]. This
insight is critical to the assessment of the relative value of
flexibility resources. In this example, DR significantly trumps
peaking units as the flexibility resource of choice. In addition,
DR is more efficient than peaking plants or batteries in that it
involves no energy conversion and therefore sidesteps any
losses associated with thermodynamic or electrochemical
processes. Covino et al. discuss the relative economics of DR
and peaking capacity in the PJM context in Chapter 17 of this
volume.
The historical focus on DR as a last line of defense against
cascading system failure has recently been superseded with an
offensive strategy of incentivizing reliability-focused peak
reduction. Incentive-based peak load reduction DR programs
have experienced tremendous growth in recent years. From
2006 to 2008 the number of entities offering DR programs
rose from 126 to 271, a 117% increase. Also, the size of peak
load reduction from existing DR resources, relative to the
national peak, rose from 5% in 2006 to 5.8% in 2008, a 16%
increase.
The New England Independent System Operator's (ISO-NE)
DR programs increased from 0.4% of forecasted system peak
demand in 2002 to 3.9% in 2007: a 57% levelized compound
annual growth rate (CAGR) [10]. Enrollment in PJM's DR
programs sustained a 17% levelized CAGR, growing from
2100 MW in 2002 to 4600 MW in 2007. Figure 9.4 illustrates
the DR's growth in the largest U.S. market, PJM, post-2007.
Of note is the depreciating clearing price that results from
459
DR's growth relative to peak capacity. Thus, a curtailment
service provider's business model suffers from revenue
cannibalization as a result of successful market growth.
Figure 9.4
Growth of demand response penetration in the PJM interconnection.
Source: Bloomberg New Energy Finance [36]
Future DR programs are anticipated to expand from the
commercial and industrial (C&I) sectors into the residential
with dynamic pricing programs and advanced metering
infrastructure (AMI). Advanced metering infrastructure and
the IT overlay of smart grid technology will enable
low-latency communication between loads and system
operators (or third-party aggregators) that will facilitate the
real time operational control of loads.
The Federal Energy Regulatory Commission's (FERC)
biennial staff report on AMI deployment states that AMI
penetration increased by more than a factor of 6 in just two
460