- Email: [email protected]
Evolution and current status of demand response (DR) in electricity markets: Insights from PJM and NYISO
Energy 35 (2010) 1553–1560
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
Evolution and current status of demand response (DR) in electricity markets: Insights from PJM and NYISO Rahul Walawalkar a, *, Stephen Fernands a, Netra Thakur a, Konda Reddy Chevva b a b
Customized Energy Solutions Ltd., 1528 Walnut St, 22nd Fl, Philadelphia, PA 19102, USA United Technologies Research Center, 411 Silver lane, E. Hartford, CT 06108, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 May 2009 Received in revised form 15 September 2009 Accepted 15 September 2009 Available online 20 February 2010
In electricity markets, traditional demand side management programs are slowly getting replaced with demand response (DR) programs. These programs have evolved since the early pilot programs launched in late 1990s. With the changes in market rules the opportunities have generally increased for DR for participating in emergency, economic and ancillary service programs. In recent times, various regulators have suggested that DR can also be used as a solution to meet supply – demand fluctuations in scenarios with significant penetration of variable renewable sources in grid. This paper provides an overview of the evolution of the DR programs in PJM and NYISO markets as well as analyzes current opportunities. Although DR participation has grown, most of the current participation is in the reliability programs, which are designed to provide load curtailment during peak days. This suggests that there is a significant gap between perception of ability of DR to mitigate variability of renewables and reality of current participation. DR in future can be scaled to play a more dynamic role in electricity markets, but that would require changes both on technology as well as policy front. Advances in building technologies and energy storage combined with appropriate price signals can lead to enhanced DR participation. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Demand Response programs Electricity markets PJM NYISO
1. Introduction When electric demand is at or near its peak level, less efficient or higher cost generating units must be utilized to meet the higher peak demand. In some cases, electricity prices in wholesale markets could fluctuate from less than 5 cents per kWh to as much as 30 cents per kWh on a significant number of days per year. During capacity shortages, prices could increase to 50 cents per kWh or higher for a few hours, reflecting the price signals that are required to match available supply to meet the demand. Under these circumstances, even a small reduction in demand through demand response (DR) programs can result in an appreciable reduction in system marginal costs of production. In competitive electricity markets, where the marginal generating unit determines market clearing price for all load, a drop in wholesale peak prices also means that non-participants in demand response also share in the benefits, as prices for everyone are held in check. These peak costs, although short in duration, add to the average cost per kWh to the consumer and hence raise the average cost of a kWh of electricity.
* Corresponding author. Tel.: þ1 215 875 9440; fax: þ1 215 875 9490. E-mail address: [email protected] (R. Walawalkar). 0360-5442/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.energy.2009.09.017
The introduction of DR into constrained electricity networks can significantly reduce volatility in wholesale electricity prices and can potentially act as a check against the exercise of market power by generators [1–4]. DR is also valuable as a tool to improve reliability of the grid [5,6], as well as increasing available transfer capacity on transmission grid [7,8]. Recent research has also indicated that historically low participation in time-differentiated pricing programs, as well as the low short-run price elasticity of demand, can result in potentially large social welfare losses in deregulated markets. The welfare losses from low demand response levels could be significantly reduced by introducing administered DR programs in concert with centralized energy spot markets. [9] Studies have also identified the need for advanced metering infrastructure (AMI) and building automation controls for enabling the potential of DR and energy efficiency [10–12]. In last decade, various research groups have tried to quantify the social benefits of DR in US markets. [13–15] A 2001 study by McKinsey & Company [13] estimates that, $10–15 billion in annual benefits can be achieved from participation of all customers in dynamic pricing programs on a wide scale across U.S., with the majority of the potential, contrary to conventional wisdom, from residential sector DR efforts. The study estimated that the infrastructure needed for dynamic pricing can be brought to the mass market, with payback periods of 5–6 years. Based on a review of
1554
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
current utility programs, Electric Power Research Institute (EPRI) [15] estimated that DR has the potential to reduce current U.S. peak demand by 45,000 MW. The U.S. Federal Energy Regulatory Commission (FERC) [16] released a cost-benefit analysis in 2002 that showed a $60 billion savings over the next 20 years if DR is incorporated into RTO market design and operations. 2. Evolution of DR programs in PJM and NYISO The ability of customers to respond to prices and reduce consumption during periods of system shortage has been a critical component of both the PJM and NYISO electricity markets since their start in 1997 and 1999 respectively. DR programs are designed to encourage consumers to modify their electric demand level and pattern of electricity usage. DR refers only to energy and load-shape modifying activities undertaken in response to economic or reliability signals provided by utilities or ISOs and not to load-shape changes arising from any normal operation. The Demand Response and Smart Grid Coalition (DRSG) defines DR as the reduction of customer energy usage at times of peak usage in order to help address system reliability, reflect market conditions and pricing, and support infrastructure optimization or deferral. Demand response programs may include dynamic pricing/tariffs, priceresponsive demand bidding, contractually obligated and voluntary curtailment, and direct load control/cycling. Based on the type of signal used to activate the DR program, these programs can be categorized as either Emergency (or Reliability based) DR programs or Economic (Price based) DR programs or Demand Side Ancillary Service programs [17]. The emergency DR programs aim to provide cost-effective capacity resources to help avoid system outages in case of severe grid stress. On the other hand economic DR programs are developed to exert a downward pressure on electricity prices, by allowing demand side participation in electricity markets. Demand Side Ancillary Service programs allow DR to participate in ancillary service markets such as frequency regulation and operating reserves. Until recently some of the energy efficiency and load shaping programs, that were part of traditional DSM initiatives, were not considered as DR. As explained later in this paper, PJM introduced a new program in 2008–09, that provides capacity credits to qualified energy efficiency projects as part of DR program. The PJM and NYISO markets have separate energy, capacity and ancillary services markets. Initially, both PJM and NYISO had mechanisms for inclusion of DR programs in the capacity markets through Emergency DR programs based on reliability criteria. This was accomplished through reducing a Load Serving Entity’s (LSE) capacity obligation but limited integration into the energy markets. Partly in response to the very high price spikes experienced in 1998 and 1999 various stakeholders realized that there was a benefit to increase the ability of customers to respond to higher prices and reduce consumption. Since most customers at that time were on fixed pricing they did not have a direct incentive to reduce consumption during high priced periods. Additionally there was no capability to put in differing bidding parameters (such as the minimum commitment period for a load to reduce). Other issues included limited availability of interval metering required for monitoring and billing customers based on their actual usage. Initially inclusion of DR in PJM’s capacity markets was done through the Active Load Management (ALM) program. This program required customers, at that time typically through their LSE, to commit prior to the summer period (June 1–September 15) that they could reduce their power consumption during at least 10 days for a period of up to 6 h per day during the summer period. In exchange the customer (or their LSE) would receive a capacity credit for the entire year. The NYISO program, called Special Case Resources, was similar in that it reduced LSE’s capacity obligation
but unlike PJM included a separate summer and winter season as well as testing requirements on participants. In both the PJM and NYISO capacity programs, the response of the participating customers was mandatory and there were various penalties associated with non-compliance. The types of DR enrolled in these programs included residential load control devices (water heaters/ air conditioners), commercial/industrial load reduction, and behind the meter generation. In practice these programs were called infrequently with 0–5 events per year being called as shown in Table 1. Capacity prices in the first few years of the markets were relatively high and were the primary source of overall revenue paid to demand side response resources. However, as capacity markets started to go down in subsequent years and energy prices started to rise, there was a move to integrate DR resources more closely into the operation of the energy markets. For example, in PJM capacity prices were $34,799/MW-year in 2001 and fell to $2091/MW-year by 2006. Since then the introduction of locational capacity market and demand curve for capacity under Reliability Pricing Mechanism (RPM) has reversed the trend, particularly in capacity constrained zones, and capacity prices varied between $35,000 and $86,000/MW-year during 2009–2010 delivery year. While many LSEs used the ability of their customers to reduce consumption during peak times for energy purposes (most often through interruptible rates), there was not an easy way to integrate these customers into the markets. In 2001 both PJM and NYISO started to develop mechanisms to allow customers to participate in the energy markets either directly through a Curtailment Service Provider (CSP) or through the customer’s LSE. For a period of time the PJM and NYISO programs leapfrogged each other with one market rolling out a component of DR and the other adopting that and building off of it. The first add-on was an emergency energy program that both NYISO and PJM added allowing customers to get paid an energy payment if they would voluntarily reduce consumption during periods of emergency. These resources were called on very rarely and there was much debate about whether or not they should set price in energy market. NYISO opted to have them set the market clearing price and PJM chose not to. In the NYISO the next DR market to be developed was the day ahead market with customers able to bid into the day ahead market in a similar manner to generators and through this process set price and be dispatched by the NYISO. In PJM both the day ahead and real time markets were opened up to DR resources. PJM initially promoted the participation in real time DR program as a voluntary effort, where customers did not face any penalties for non-compliance. It was initially believed that new customers will use the real time DR program to get familiar with the markets and then will gradually start participating in the Day Ahead DR program, where participants commit to mandatory load curtailment targets if their bid is accepted in the day ahead energy market. Even today, in PJM most of the DR participation is in the Real Time DR program, and very few customers participate in Day Ahead DR program.
Table 1 Summary of ISO/RTO initiated emergency DR events (source: PJM [18] & NYISO [19]) Year
Emergency events in PJM
Emergency events in NYISO
2000 2001 2002 2003 2004 2005 2006 2007
2 (May 8 and 9) 4 (Jul 25, Aug 8–10) 3 (Jul 3, 29, 30) None None 2 (Jul 27, Aug 4) 2 (Aug 2, 3) 1 (Aug 8)
4 (Aug 7–10) 4 (Apr 17 and 18, Jul 30 and Aug 14) 2 (Aug 15 and 16) None 1 (Jul 27) 5 (Jul 17 and 18, Aug 1–3) 2 (Jul 19 and Aug 3)
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
In both markets one of the early questions was measurement of demand response. Since the compliance of emergency DR programs were measured against the customers’ coincident peak demand, verification was much more straight forward than in the economic DR programs where various factors could affect quantity of load curtailment. The compliance verification procedures have varied over the time and have included similar day analysis using a Customer Base Line (CBL), short term load drop, separate generation metering, and statistical analysis. The latest steps in the evolution of DR in these markets, include opening up of ancillary service markets for DR and inclusion of energy efficiency as a capacity resource. PJM started allowing DR resources to participate in ancillary services markets in 2006. In 2008, FERC issued Order No. 719 which adopted additional measures to improve the operation of organized wholesale electric markets in the area of DR to further eliminate barriers and encourage the use of market prices to elicit DR. In response, NYISO also opened the ancillary service markets for DR in 2009. PJM has recently taken the next step by allowing energy efficiency projects to start receiving capacity credits for up to 4 years. 3. Current status of demand response programs in PJM and NYISO With the expansion of DR programs in PJM and NYISO, customers with demand response capabilities now have the opportunity to get paid by the ISO similar to an energy generator, making it a more level playing field. End use customers can participate in these DR programs by using either distributed generators or energy management control strategies to reduce their load in response to a price or emergency signal. NYISO: The NYISO currently operates five demand response programs [20,21]. Reliability based DR programs o Emergency DR Program (EDRP): this is a voluntary program where participants are paid the higher of $500/MWh or the real time LMP for responding during system emergency. o Special Case Resource Program (SCR): this program pays participants higher of their strike price (up to $500/MWh) or the real time LMP for mandatory response to emergency events. These resources can also sell capacity in the capacity market. o Targeted Demand Response Program (TDRP): this program was initiated in July 2007 to pay EDRP and SCR resources higher of $500/MWh or the real time LMP to respond to local reliability events within a particular zone. This avoids need to activate emergency events for entire zone to address local reliability concerns. Economic DR program: o Day Ahead DR Program (DADRP): DADRP allows customers to submit bids in day ahead energy market for curtailing load above a certain bid floor price. The bid floor price was increased in 2004 to $75 from original value of $50. Customer must respond if their bid is cleared in the day ahead market and curtailment cannot be achieved with use of backup generators. Failure to curtail results in a penalty defined by the MW curtailment shortfall multiplied by the greater of the corresponding day ahead or real time market price. Ancillary services: o Demand Side Ancillary Service Program (DSASP): DSASP was initiated in 2009 and allows DR to provide three ancillary services to NYISO markets. These include regulation, synchronous reserve and non-synchronous reserve.
1555
Fig. 1 shows that over the years customers have preferred the SCR program which is most lucrative due to the locational capacity market revenues. Participation in the EDRP has gradually decreased since 2003, when the EDRP and SCR program participation was made mutually exclusive. The figure also shows that participation in the DADRP program has not changed significantly over the years, and suggests that NYISO is not able to attract more customers to participate in the economic DR program in recent years [21]. PJM: The PJM currently operates seven DR programs [22–24]. Reliability based DR programs o Emergency DR Program: similar to the NYISO program, this is a voluntary program for reliability that offers energy payments to customers that reduce load during a system emergency. The payments are the higher of $500/MWh or the zonal LMP for the hour. There is no penalty for noncompliance, and this program is rarely utilized by PJM (only five times in past 3 years). o Interruptible Load for Reliability (ILR): this program is similar to the NYISO SCR program and allows ILR to receive capacity payments for load reduction achieved during system emergency. Load curtailment is mandatory for up to 10 events each year, lasting up to 6 h per event. To encourage higher participation of DR in capacity markets, ILR resources are allowed to register by March each year for upcoming delivery year (May–April). During 2008–2009, ILR accounted for 3489 MW out of a total of 4497 MW registered under emergency DR. Last year FERC ruled in favor of eliminating the ILR program option to eliminate the favorable treatment of DR under RPM. Thus 2011–2012 is the last delivery year for ILR participation. o Demand Resource (DR): Demand Resource is also eligible to receive capacity revenues through PJM’s Reliability Pricing Model (RPM). This resource has to participate in the forward capacity auction conducted by PJM 3 years before the delivery year along with other generation resources. So far majority of the DR resources had opted for ILR option to participate in capacity markets. Since DR will be the only option available from 2012 to 2013, all DR resources interested in receiving capacity credits in that year need to register as DR in May 2009. Some of the current ILR participants may find it difficult to accept the financial risk of penalties by committing resources 3 year before the delivery year. This may reduce total amount of DR available as a capacity resource for PJM.
Fig. 1. NYISO DR registration in MWs during 2001–2008 (source: NYISO, 2009 [21]).
1556
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
o Energy Efficiency Resource (EE): in 2009, PJM started allowing energy efficiency resource to receive capacity credit for up to 4 consecutive years. An EE Resource is a project that involves the installation of more efficient devices/equipment, or the implementation of more efficient processes/ systems, exceeding then-current building codes, appliance standards, or other relevant standards, at the time of installation. Similar to Demand Resource, an EE resource also has to participate in the forward capacity auction conducted by PJM 3 years before the delivery year. The nominated EE value is the expected average demand (MW) reduction during the hours between 2 p.m. and 6 p.m. during all days from June 1 through August 31 (PJM, 2009).
Economic DR Program: Under this program PJM pays the customer the difference between the LMP and the generation and transmission (G&T) components of the customer’s bill. PJM offers this economic DR program in both its day ahead and real time markets. o Day Ahead Demand Response Program (DADRP): this is a mandatory program, where participant submits bids in day ahead energy market and faces penalty for any non-compliance. o Real Time Demand Response Program (RTDRP): RTDRP is a voluntary program, where customer needs to notify PJM of intent to curtail load at least 1 h before load reduction is performed. Before January 2008, PJM also included an incentive payment for economic DR participants when they responded above a strike price of $75/MWh (similar to NYISO). The incentive payment equaled the G&T component, and thus till 2007, DR participants were eligible to receive full LMP value for curtailment during hours when LMP was above $75/MWh as is the case in NYISO. FERC [25] upheld the expiration of the incentive payments in an order under Docket EL08-12-000, issued on 31 December 2007. Ancillary Service DR Program: In 2006, PJM started allowing DR to provide frequency regulation, synchronized reserve and day ahead scheduling reserves in ancillary service market. DR resources are required to bid into ancillary service markets and respond to any event similarly to a generator. As of 2008, there were no DR resources providing frequency regulation in PJM. Although PJM has not released information about amount of DR providing synchronized reserves, these resources received almost $5M in payments during both 2007 and 2008 [26,27]. Emergency DR participation has increased more than four times over past 3 years, particularly as a result of substantial increase in capacity revenues under the RPM capacity markets, as compared to historical capacity revenues. This can be seen in Fig. 2. Table 2 provides a quick summary of current status of DR programs in these electricity markets.
Fig. 2. PJM DR program registration (2006–2008) [26,27,30].
wind) in the supply mix over next 3–5 years. Due to the variability of wind, the systems operators are concerned with reliability of the grid as well as price volatility. During some recent discussions, DR is being proposed as one of the solutions that could balance the variability in supply [28]. Although DR has proved to be an effective resource to meet reliability needs in both PJM and NYISO, the amount of DR as well as the type of response desired to mitigate variability of large scale wind or solar is significantly different. Current DR participation in these markets is clearly not sufficient to provide a viable solution for matching variability of resources like wind on a continuous basis, especially in scenarios with 10% or higher penetration of renewables. As mentioned earlier, NYISO [21] has failed to attract any additional load curtailment in the DADRP program since 2003. On the other hand the registration in PJM’s economic DR programs has grown substantially over the years to 3290 MW in 2008 [26]. At the same time, a distinction must be made between loads that are registered to participate in the PJM economic DR program, and the amount of load that actively participates in the market [29]. During 2006, there was 1475 MW load registered under the economic DR program and an additional 1081 MW load registered under the emergency DR program [30]. However, during the summer of 2006 when PJM registered it’s all time peak load of 144.6 GW, there was only 325 MW of DR cleared in the economic DR program [33]. This trend has continued in recent years and data released by PJM in 2009 indicates that out of over 4800 sites registered in economic Table 2 Summary comparison of NYISO and PJM NYISO
PJM
Population Served States
19.2 Million NY
Peak load Types of DR programs offered
33.9 GW Economic (DADRP), Emergency (EDRP, SCR, TDRP), Ancillary Services (DSASP) EDRP/SCR/TDRP: 100 kW DADRP/DSASP: 1 MW 1774 MW (SCR þ EDRP)
51 Million DE, IL, IN, KY, MD, MI, NC, NJ, OH, PA, TN, VA, WV and DC 144.6 GW Economic (DADRP, RTDRP), Emergency (EDRP, ILR, DR, EE), Ancillary Services
4497 MW (ILR þ DR)
320 MW
3290 MW
3.1. Challenges for economic demand response in PJM Minimum size
With the impressive growth of DR in these markets as well as plans of implementing smart grid, some policy makers and market participants are anticipating that DR will be able to play a much bigger role in electricity markets in near future. Currently most of the electricity markets, including NYISO are anticipating a significant growth in penetration of renewable resources (particularly
Registered emergency DR Registered economic DR
100 kW (for all)
Adapted from Walawalkar et al. [31], FERC [32], and ISO websites.
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
1557
Fig. 3. NPV of potential DR revenues over 5 years with and without incentive (source: Walawalkar et al. [23]).
DR programs, maximum of 301 sites actually curtailed load under DADRP/RTDRP program in any month during 2007–2008 [26,27]. This situation needs to change dramatically by encouraging more active participation of DR in energy and ancillary service markets as well as development of new DR programs that will provide appropriate price signal for customers to mitigate fluctuations in renewables from hour to hour. One of the issues facing economic DR participants in PJM is the elimination of incentive for participation in economic DR program in December 2007 [23]. As a result of the expiration of incentives for DADRP and RTDRP, instead of receiving full LMP for load curtailment, participants now receive payments from PJM based on difference between the LMP and the generation and transmission (G&T) components of the customer’s bill. Thus DR participants may have increased their strike price so that the marginal payments from DR program participation are equalized with and without the incentive payment. For instance, customers with a G&T rate of $50/ MWh may change their strike price from $75/MWh to $125/MWh to retain the same payment from PJM for curtailing load. The total payoff to any individual customer from economic DR participation depends not only on the strike price and incentive structure, but also the distribution of market prices that can change from year to year. Fig. 3 shows the results of a simulation conducted to project effect of removal of incentive on revenues for economic DR participants in PJM [23]. We have incorporated uncertainty due to fluctuations in energy prices from year to year into our analysis by performing a Monte Carlo simulation on the expected annual DR revenue stream, using energy market data from PJM during 2004– 2007 [34]. We modeled annual revenues based on participant’s strike prices, with and without the DR market incentives. For each year t and strike price k, we assumed that annual DR revenues (p) follow a triangular distribution with the minimum value equal to the sum of hourly revenues in 2004 (the lowest revenue year in our sample), the maximum equal to the sum in 2005 (the highest), and the most likely value equal to the average of the four years. We generated 1000 realizations of the discounted present value of expected annual revenues over a five year time horizon, assuming a customer with an internal discount rate of 10%; we did this for a $75 strike
price with the incentive, for a $75 strike price without the incentive and for a $125 strike price without the incentive. The simulation indicates that before expiration of incentives, any economic DR resource offering 1 MW load curtailment at $75/ MWh would receive a discounted gross revenue stream of $610,000 over five years with a 50% probability. On the other hand, without the incentive payment and customer adjusting her strike price to $125/MWh (to receive the DR payment of $75/MWh), the NPV would fall by roughly a factor of five, to $107,000 as shown in Fig. 3. The change in the potential DR revenue in economic DR program due to change in the incentive structure can also be seen in the data released by PJM as shown in Fig. 4. The figure indicates that although overall DR revenues for all participants have grown substantially over past 3 years, revenue for economic DR participation decreased from approximately $60M in 2007 to $20M in 2008. As a result although overall participation in the DR is increasing over the years considering the reliability based programs, increasing active participation in the price-responsive DR will be a challenge. PJM Board of Directors acknowledged this in a press release on 26 June 2009, which announced that PJM will reintroduce incentive payments for demand response for the top 9% of the hours from the previous year.1
4. Opportunities for growth Currently the majority of DR program participation is through use of backup generators or use of curtailable loads. The vast potential of existing building EMS systems and energy management strategies is still untapped, not to mention future possibilities as home automation and smart appliances become standardized along with deployment of smart grid infrastructure as shown in Fig. 5. Also most of the energy efficiency project developers are not
1 Statement of Terry Boston, President and CEO, on behalf of the PJM Board of Managers is available at http://pjm.com/w/media/about-pjm/newsroom/2009releases/20090626-pjm-board-statement-regarding-dr-in-pjm-markets.ashx.
1558
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
Fig. 4. PJM demand side response estimated revenues 2001–2008 (source: PJM [35]).
familiar with the electricity markets and the newly developed option of receiving capacity revenues for EE resources. In a competitive market, new generating capacity will not be added until prices rise sufficiently above the cost of new facilities. The existing energy efficiency and control technologies, however, can be a great tool in expanding DR program participation. Table 3 shows a sample of some of the control strategies that can be used by different types of customers for participating in various DR programs. Ongoing advances in the research focused on develop technologies, tools and standards to make commercial and residential building systems more energy efficient and productive can also facilitate DR participation. Some of key areas of research related to high performance buildings include new building and equipment standards, integrated building systems, and solid-state lighting. Besides making existing buildings energy efficient, there is also an increased effort to design new net zero energy buildings by integrating renewable technologies like solar, geothermal, wind, biomass and fuel-cell and hydrogen technology.
We briefly discuss some of the emerging technologies and standards that hold a lot of promise for enabling DR in residential and commercial segment. ZigBee enabled DR systems: demand response systems for monitoring energy usage in homes are amongst the most promising of new technologies. The technology that is driving a new wave of home automation is Zigbee, a low-power wireless networking standard designed for controlling and monitoring applications. On one hand, utility companies are increasingly evaluating Advanced Metering Infrastructure (AMI) networks to provide two-way communication between the home and utilities via wireless smart meters. On the other hand, wireless Home Area Networks (HANs) are being adopted by home-owners to better gain whole-house entertainment control. Zigbee is the wireless technology that helps connect AMI networks and HANs. This can allow home-owners and utilities to communicate in real time and manage energy consumption.
Fig. 5. Components of a Demand Response Program implementation (Walawalkar et al. [31]).
R. Walawalkar et al. / Energy 35 (2010) 1553–1560 Table 3 Control strategies and applicable DR programs for different customer types Customer type
Equipment/ Control building strategy component
DR programs
Residential
Air Cycling/forced conditioners demand shedding Water Cycling heaters Pool pumps Cycling Electric Scheduling stove
U
U
U
U
U
U
U
U U
U
Demand limiting U during on peak period Chillers Pre-cool bldg HVAC over nightstorage DX Forced U demand scheduling Refrigerator/ Prioritized freezer demand shedding Lighting Scheduled on/off Lighting Scheduled dimming of selected circuits
U
Emergency Economic Ancillary (capacity) (energy) (Reg & Res)
Commercial Chillers
Industrial
Chillers
Demand limiting on time schedule Electric Demand limiting furnace through heat stages Curtail (during U Electric furnace peak period) VSDs Limit output on scheduled basis Well pumps Defer during U peak Production Prioritized equipment demand on selected units
Restaurants/ HVAC shopping malls DX compressor
Chillers – demand limiting during peak Forced demand U shedding of multiple units Refrigerator/ Prioritized freezer demand shedding Electric Scheduled preSToves cooking
U
U
U
U U
U U
U
U
U
U
1559
New Building Standards: building codes have been used for nearly three decades and are a cost-effective strategy to overcome barriers to energy efficiency in buildings. Most residential buildings adopt the International Energy Conservation Code (IECC) whereas most commercial building energy codes are based on ASHRAE/IESNA Standard 90.1. According to California Energy Commission, building standards have helped save businesses and residents in California more than $15.8 billion in electricity and natural gas costs since 1975. Now with the availability of capacity revenues for EE resources from electricity markets, businesses have an added incentive for choosing newer standards that go beyond current standards. The vision for the future is that buildings are ‘‘DR enabled’’ and major appliances are ‘‘DR ready’’ from the factory. Building standards that incorporate controllable thermostats in residential and price server interface in commercial buildings will greatly help DR. On the appliance side, DR will benefit from standards that include embedded controls for HVAC and other systems. Energy storage technologies: another emerging technology area that can provide significant boost to DR capability in electricity markets is electric energy storage (EES). Currently approximately 2.5% of the total electric power delivered in the United States passes through energy storage, largely pumped hydroelectric. The percentages are significantly larger in Europe and Japan, at 10% and 15%, respectively [36]. In recent years, there has been tremendous advances made in various EES technologies, particularly in areas of advanced batteries (such as lithium ion, flow batteries and sodium sulfur), flywheel technologies that can provide frequency regulation as well as modular designs for above ground Compressed Air Energy Storage (CAES). With successful adaption of some of these technologies in commercial and industrial applications, market participants investing in EES as a behind the meter resource, can use it as DR resource.
U U
U
U
U
U
Adapted from Walawalkar et al. [31].
In-Home Energy Use Displays: in-home energy use displays provide real time feedback to the occupants about the total energy usage in the house. This information can be displayed through existing home PCs using applications such as Google PowerMeterÒ or Microsoft HohmÒ. Google PowerMeterÒ can receive information from utility smart meters and energy management devices and provides customers with access to their home electricity consumption right on their personal iGoogle homepage. On the other hand, new products are being developed that can be used to signal the need for demand response without consumers using their computers. In California, PG&E has introduced a ‘‘Demand Response Orb’’ to inform customers of Critical Peak Pricing periods and/or emergency events. The projected energy savings from the use of such displays ranges from 4% to 15%.
5. Conclusions Electricity markets have made significant progress over past decade in integrating DR into electricity market design. PJM and NYISO have been on the forefront of such developments, and currently offer a wide range of programs that allow DR to participate in these markets. In general all DR programs rely on end users deliberately altering use of equipment and systems, which generally means lifestyle or comfort changes, or changes in operating procedures. Such changes would be acceptable to end users only if the consumer has a stake in the process either through financial compensation or through improved reliability of power supply. Also it is important to realize that all DR programs are not equal. Different types of programs have specific objectives, and demand varying level of commitment from DR resources with different payoffs. Current market based DR programs are very successful in attracting participation in emergency or reliability based programs, which require relatively modest commitments from end users. If we expect DR resources to scale up to enable greater level of renewable integration, then we need to start evaluating DR technologies that are suitable for participating every day. We also need to evaluate if current programs are the right programs to provide incentives to customers for such role. For example, Economic DR programs will not provide sufficient revenues for customers to respond during off peak hours, whereas emergency programs are designed to provide response only during peak hours. On the other hand the wind energy is generally expected to have the largest share of the supply mix during night or early morning hours. We will need DR technologies such as EES that can respond in both
1560
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
direction (absorb power as well as provide fill up power), and new DR programs that provide appropriate financial incentives for customers to participate. We are witnessing great advances in technology that can help DR to scale up substantially, but technology alone is not sufficient. The process of evolution of DR programs in electricity markets is not complete, and with changing needs of the market, stakeholders in NYISO and PJM need to continue efforts for increasing DR participation.
[14]
[15]
[16]
[17]
Acknowledgements
[18]
We thank Dr. Jay Apt, Dr. Lester Lave, Dr. Rahul Tongia, Dr. Granger Morgan (Carnegie Mellon University), Dr. Bruce Colburn (EPS Capital Corp.) and Dr. Seth Blumsack (Pennsylvania Stage University) for useful comments and conversations. We acknowledge timely inputs from our colleagues Rick Mancini and Rick Gilkey at Customized Energy Solutions. We also acknowledge contribution of the reviewers of this special issue, which helped us to identify areas that required further explanation.
[19]
References
[23]
[1] Electricity markets: consumers could benefit from demand programs, but challenges remain (US-GAO-04-844), A report to Senate Committee on Governmental Affairs. U.S. Government Accountability Office. Available from:, http://www.gao.gov/new.items/d04844.pdf; 2004. [2] Violette D, Freeman R, Neil C, DRR valuation and market analysis. Overview, a report for the International Energy Agency (IEA) DRR Task XIII, vol. I; 2006. Available from: http://www.demandresponsecommittee.org/id80.htm. [3] Violette D, Freeman R, Neil C, DRR valuation and market analysis, Assessing the DRR benefits and costs, a report for the International Energy Agency (IEA) DRR Task XIII, vol. II; 2006. Available from: http://www. demandresponsecommittee.org/id81.htm. [4] Rassenti S, Smith V, Wilson B. Demand-side bidding will reduce the level and volatility of electricity prices. Independent Review(3):441–5. Available from: http://www.independent.org/pdf/tir/tir_06_3_smith.pdf, 2002;6. [5] Goel L, Wu Q, Wang P. Reliability enhancement of a deregulated power system considering demand response. IEEE Power Engineering Conference Issue 2005;2:934–8. [6] Azami R, Fard A. Impact of demand response programs on system and nodal reliability of a deregulated power system. Sustainable Energy Technologies, ICSET 2008;2008:1262–6. [7] Shayesteh E, Yousefi A, Moghaddam P. ATC enhancement using emergency demand response program. IEEE Power Systems Conference and Exposition 2009;2009. PES ’09. [8] Green R. Electricity transmission pricing: how much does it cost to get it wrong?. MIT Center for Energy and Environmental Policy Research; 2004. MIT-CEEPR (Series); 04-020WP. [9] Boisvert R, Neenan B, Social welfare implications of demand-response programs in competitive electricity markets, Lawrence Berkeley National Laboratories working paper LBNL-52530; 2003. Available from: http://eetd.lbl. gov/ea/EMS/reports/LBNL-52530.pdf. [10] Lavy R, Herter K, Wilson J, Unlocking the potential for efficiency and demandresponse through advanced metering, Lawrence Berkeley National Laboratory working paper; 2004. Available from: http://repositories.cdlib.org/lbnl/LBNL55673. [11] Piette M, Watson D, Motegi N, Bourassa N. Findings from the 2004 fully automated demand response tests in large facilities. Demand Response Research Center; 2005. LBNL Report Number 58178. [12] FERC. A national assessment of demand response potential. staff report. Federal Energy Regulatory Commission; June 2009. [13] McKinsey & Company. The benefits of demand side management and dynamic pricing programs. Mckinsey and Company. Available from: http://republicans.
[20] [21]
[22]
[24]
[25] [26]
[27]
[28] [29]
[30]
[31]
[32] [33]
[34] [35]
[36]
energycommerce.house.gov/107/hearings/06222001Hearing265/swofford.pdf; May 2001. Lovins A, Datta E, Feiler T, Rabago K, Swisher J, Lehmann A, et al. Small is profitable: the hidden economic benefits of making electrical resources the right size. Rocky Mountain Institute; 2002. The Western States power crisis: imperatives and opportunities, EPRI white paper. EPRI. Available from: http://www.pserc.wisc.edu/ecow/get/publicatio/ specialepr/gridreliab/westernstatessynthesi1.pdf; 2001. ICF Consulting. Economic assessment of RTO policy. FERC. Available from: http://www.ferc.gov/legal/maj-ord-reg/land-docs/rtostudy_final_0226.pdf; 2002. PJM. A demand response roadmap in PJM symposium on demand response II. PJM Interconnection. Available from: www.pjm.com; 2008. PJM. Summary of PJM-initiated load management events. PJM Interconnection. Available from: http://www.pjm.com/planning/res-adequacy/ downloads/20070907-alm-history.pdf; Sept 2007. NYISO. Historic EDRP and SCR activation information. Nyiso. Available from: http://www.nyiso.com/public/webdocs/products/demand_response/general_ info/event_history.pdf; 2008. NYISO. State of the market report 2008. New York ISO. Available from: www. nyiso.com; 2007. NYISO, Annual compliance report on demand side management and new generation projects submitted to FERC; January 2009. Available from: http://www.nyiso.com/public/webdocs/documents/regulatory/filings/2009/01/ DemandSideMgmtCompFiling_1_15_09.pdf. PJM, Demand response program information. Available from: http://www.pjm. com/markets-and-operations/demand-response.aspx [Information accessed on 1 May 2009]. Walawalkar R, Blumsack S, Apt J, Fernands S. An economic welfare analysis of Demand Response in PJM electricity market. The Energy Policy October 2008;36(10). Brattle Group, 2007. Quantifying demand response benefits in PJM, A report for PJM Interconnection and Mid Atlantic Distributed Resource Initiative (MADRI). Available from: http://www.energetics.com/madri/pdfs/BrattleGroup Report.pdf. Docket EL08-12-000. Federal Energy Regulatory Commission (FERC); 2007. issued on 31 December 2007. O’Neil J. DSR activity report presented to PJM Demand Response Steering Committee. PJM 2009. Available from: http://www.pjm.com/committees-andgroups/committees/drsc.aspx; 2008. Gockley B. DSR activity report presented to PJM Demand Response Steering Committee. PJM 2009. Available from: http://www.pjm.com/committees-andgroups/committees/drsc.aspx; 2007. Stadler I. Power grid balancing of energy systems with high renewable energy penetration by demand response. Utilities Policy 2007;16(2):90–8. Hopper N, Goldman C, Bharvirkar R, Engel D. The summer of 2006: a milestone in the ongoing maturation of demand response. The Electricity Journal June 2007;20(5):62–75. Kujawski D. DSR activity report presented to PJM Demand Response Steering Committee. PJM 2009. Available from: http://www.pjm.com/committees-andgroups/committees/drsc.aspx; 2006. Walawalkar R, Tongia R, Colburn B. Web-enabled metering and controls for demand response. In: Capehart BL, Capehart LC, editors. Web based enterprise energy and building automation systems. Lilburn, Georgia: The Fairmont Press, ISBN 0849382351; 2007. p. 299–314. FERC. Electric power markets: national overview Available from: http://www. ferc.gov/market-oversight/mkt-electric/overview.asp. Covino S, PJM economic load response: summer 2006 potential and performance. A presentation to the Demand Side Response Working Group; November 2006. Available from: http://www.pjm.com/committees/workinggroups/dsrwg/downloads/20061130-item-09-summer-2006-performance. pdf. PJM locational marginal pricing data. Available from: http://www.pjm.com/ markets/jsp/lmp.jsp. PJM. PJM demand side response estimated revenue, Presented to PJM Demand Response Steering Committee; 2009. Available from: http://www. pjm.com/committees-and-groups/committees/w/media/committees-groups/ committees/drsc/postings/demand-side-response-estimated-revenue.ashx. Walawalkar R, Apt J, Market analysis of emerging electric energy storage systems, A report developed for Department of Energy (DOE) and National Energy Technology Labs (NETL), DOE/NETL-2008/1330; July 2008.
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
Evolution and current status of demand response (DR) in electricity markets: Insights from PJM and NYISO Rahul Walawalkar a, *, Stephen Fernands a, Netra Thakur a, Konda Reddy Chevva b a b
Customized Energy Solutions Ltd., 1528 Walnut St, 22nd Fl, Philadelphia, PA 19102, USA United Technologies Research Center, 411 Silver lane, E. Hartford, CT 06108, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 May 2009 Received in revised form 15 September 2009 Accepted 15 September 2009 Available online 20 February 2010
In electricity markets, traditional demand side management programs are slowly getting replaced with demand response (DR) programs. These programs have evolved since the early pilot programs launched in late 1990s. With the changes in market rules the opportunities have generally increased for DR for participating in emergency, economic and ancillary service programs. In recent times, various regulators have suggested that DR can also be used as a solution to meet supply – demand fluctuations in scenarios with significant penetration of variable renewable sources in grid. This paper provides an overview of the evolution of the DR programs in PJM and NYISO markets as well as analyzes current opportunities. Although DR participation has grown, most of the current participation is in the reliability programs, which are designed to provide load curtailment during peak days. This suggests that there is a significant gap between perception of ability of DR to mitigate variability of renewables and reality of current participation. DR in future can be scaled to play a more dynamic role in electricity markets, but that would require changes both on technology as well as policy front. Advances in building technologies and energy storage combined with appropriate price signals can lead to enhanced DR participation. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Demand Response programs Electricity markets PJM NYISO
1. Introduction When electric demand is at or near its peak level, less efficient or higher cost generating units must be utilized to meet the higher peak demand. In some cases, electricity prices in wholesale markets could fluctuate from less than 5 cents per kWh to as much as 30 cents per kWh on a significant number of days per year. During capacity shortages, prices could increase to 50 cents per kWh or higher for a few hours, reflecting the price signals that are required to match available supply to meet the demand. Under these circumstances, even a small reduction in demand through demand response (DR) programs can result in an appreciable reduction in system marginal costs of production. In competitive electricity markets, where the marginal generating unit determines market clearing price for all load, a drop in wholesale peak prices also means that non-participants in demand response also share in the benefits, as prices for everyone are held in check. These peak costs, although short in duration, add to the average cost per kWh to the consumer and hence raise the average cost of a kWh of electricity.
* Corresponding author. Tel.: þ1 215 875 9440; fax: þ1 215 875 9490. E-mail address: [email protected] (R. Walawalkar). 0360-5442/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.energy.2009.09.017
The introduction of DR into constrained electricity networks can significantly reduce volatility in wholesale electricity prices and can potentially act as a check against the exercise of market power by generators [1–4]. DR is also valuable as a tool to improve reliability of the grid [5,6], as well as increasing available transfer capacity on transmission grid [7,8]. Recent research has also indicated that historically low participation in time-differentiated pricing programs, as well as the low short-run price elasticity of demand, can result in potentially large social welfare losses in deregulated markets. The welfare losses from low demand response levels could be significantly reduced by introducing administered DR programs in concert with centralized energy spot markets. [9] Studies have also identified the need for advanced metering infrastructure (AMI) and building automation controls for enabling the potential of DR and energy efficiency [10–12]. In last decade, various research groups have tried to quantify the social benefits of DR in US markets. [13–15] A 2001 study by McKinsey & Company [13] estimates that, $10–15 billion in annual benefits can be achieved from participation of all customers in dynamic pricing programs on a wide scale across U.S., with the majority of the potential, contrary to conventional wisdom, from residential sector DR efforts. The study estimated that the infrastructure needed for dynamic pricing can be brought to the mass market, with payback periods of 5–6 years. Based on a review of
1554
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
current utility programs, Electric Power Research Institute (EPRI) [15] estimated that DR has the potential to reduce current U.S. peak demand by 45,000 MW. The U.S. Federal Energy Regulatory Commission (FERC) [16] released a cost-benefit analysis in 2002 that showed a $60 billion savings over the next 20 years if DR is incorporated into RTO market design and operations. 2. Evolution of DR programs in PJM and NYISO The ability of customers to respond to prices and reduce consumption during periods of system shortage has been a critical component of both the PJM and NYISO electricity markets since their start in 1997 and 1999 respectively. DR programs are designed to encourage consumers to modify their electric demand level and pattern of electricity usage. DR refers only to energy and load-shape modifying activities undertaken in response to economic or reliability signals provided by utilities or ISOs and not to load-shape changes arising from any normal operation. The Demand Response and Smart Grid Coalition (DRSG) defines DR as the reduction of customer energy usage at times of peak usage in order to help address system reliability, reflect market conditions and pricing, and support infrastructure optimization or deferral. Demand response programs may include dynamic pricing/tariffs, priceresponsive demand bidding, contractually obligated and voluntary curtailment, and direct load control/cycling. Based on the type of signal used to activate the DR program, these programs can be categorized as either Emergency (or Reliability based) DR programs or Economic (Price based) DR programs or Demand Side Ancillary Service programs [17]. The emergency DR programs aim to provide cost-effective capacity resources to help avoid system outages in case of severe grid stress. On the other hand economic DR programs are developed to exert a downward pressure on electricity prices, by allowing demand side participation in electricity markets. Demand Side Ancillary Service programs allow DR to participate in ancillary service markets such as frequency regulation and operating reserves. Until recently some of the energy efficiency and load shaping programs, that were part of traditional DSM initiatives, were not considered as DR. As explained later in this paper, PJM introduced a new program in 2008–09, that provides capacity credits to qualified energy efficiency projects as part of DR program. The PJM and NYISO markets have separate energy, capacity and ancillary services markets. Initially, both PJM and NYISO had mechanisms for inclusion of DR programs in the capacity markets through Emergency DR programs based on reliability criteria. This was accomplished through reducing a Load Serving Entity’s (LSE) capacity obligation but limited integration into the energy markets. Partly in response to the very high price spikes experienced in 1998 and 1999 various stakeholders realized that there was a benefit to increase the ability of customers to respond to higher prices and reduce consumption. Since most customers at that time were on fixed pricing they did not have a direct incentive to reduce consumption during high priced periods. Additionally there was no capability to put in differing bidding parameters (such as the minimum commitment period for a load to reduce). Other issues included limited availability of interval metering required for monitoring and billing customers based on their actual usage. Initially inclusion of DR in PJM’s capacity markets was done through the Active Load Management (ALM) program. This program required customers, at that time typically through their LSE, to commit prior to the summer period (June 1–September 15) that they could reduce their power consumption during at least 10 days for a period of up to 6 h per day during the summer period. In exchange the customer (or their LSE) would receive a capacity credit for the entire year. The NYISO program, called Special Case Resources, was similar in that it reduced LSE’s capacity obligation
but unlike PJM included a separate summer and winter season as well as testing requirements on participants. In both the PJM and NYISO capacity programs, the response of the participating customers was mandatory and there were various penalties associated with non-compliance. The types of DR enrolled in these programs included residential load control devices (water heaters/ air conditioners), commercial/industrial load reduction, and behind the meter generation. In practice these programs were called infrequently with 0–5 events per year being called as shown in Table 1. Capacity prices in the first few years of the markets were relatively high and were the primary source of overall revenue paid to demand side response resources. However, as capacity markets started to go down in subsequent years and energy prices started to rise, there was a move to integrate DR resources more closely into the operation of the energy markets. For example, in PJM capacity prices were $34,799/MW-year in 2001 and fell to $2091/MW-year by 2006. Since then the introduction of locational capacity market and demand curve for capacity under Reliability Pricing Mechanism (RPM) has reversed the trend, particularly in capacity constrained zones, and capacity prices varied between $35,000 and $86,000/MW-year during 2009–2010 delivery year. While many LSEs used the ability of their customers to reduce consumption during peak times for energy purposes (most often through interruptible rates), there was not an easy way to integrate these customers into the markets. In 2001 both PJM and NYISO started to develop mechanisms to allow customers to participate in the energy markets either directly through a Curtailment Service Provider (CSP) or through the customer’s LSE. For a period of time the PJM and NYISO programs leapfrogged each other with one market rolling out a component of DR and the other adopting that and building off of it. The first add-on was an emergency energy program that both NYISO and PJM added allowing customers to get paid an energy payment if they would voluntarily reduce consumption during periods of emergency. These resources were called on very rarely and there was much debate about whether or not they should set price in energy market. NYISO opted to have them set the market clearing price and PJM chose not to. In the NYISO the next DR market to be developed was the day ahead market with customers able to bid into the day ahead market in a similar manner to generators and through this process set price and be dispatched by the NYISO. In PJM both the day ahead and real time markets were opened up to DR resources. PJM initially promoted the participation in real time DR program as a voluntary effort, where customers did not face any penalties for non-compliance. It was initially believed that new customers will use the real time DR program to get familiar with the markets and then will gradually start participating in the Day Ahead DR program, where participants commit to mandatory load curtailment targets if their bid is accepted in the day ahead energy market. Even today, in PJM most of the DR participation is in the Real Time DR program, and very few customers participate in Day Ahead DR program.
Table 1 Summary of ISO/RTO initiated emergency DR events (source: PJM [18] & NYISO [19]) Year
Emergency events in PJM
Emergency events in NYISO
2000 2001 2002 2003 2004 2005 2006 2007
2 (May 8 and 9) 4 (Jul 25, Aug 8–10) 3 (Jul 3, 29, 30) None None 2 (Jul 27, Aug 4) 2 (Aug 2, 3) 1 (Aug 8)
4 (Aug 7–10) 4 (Apr 17 and 18, Jul 30 and Aug 14) 2 (Aug 15 and 16) None 1 (Jul 27) 5 (Jul 17 and 18, Aug 1–3) 2 (Jul 19 and Aug 3)
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
In both markets one of the early questions was measurement of demand response. Since the compliance of emergency DR programs were measured against the customers’ coincident peak demand, verification was much more straight forward than in the economic DR programs where various factors could affect quantity of load curtailment. The compliance verification procedures have varied over the time and have included similar day analysis using a Customer Base Line (CBL), short term load drop, separate generation metering, and statistical analysis. The latest steps in the evolution of DR in these markets, include opening up of ancillary service markets for DR and inclusion of energy efficiency as a capacity resource. PJM started allowing DR resources to participate in ancillary services markets in 2006. In 2008, FERC issued Order No. 719 which adopted additional measures to improve the operation of organized wholesale electric markets in the area of DR to further eliminate barriers and encourage the use of market prices to elicit DR. In response, NYISO also opened the ancillary service markets for DR in 2009. PJM has recently taken the next step by allowing energy efficiency projects to start receiving capacity credits for up to 4 years. 3. Current status of demand response programs in PJM and NYISO With the expansion of DR programs in PJM and NYISO, customers with demand response capabilities now have the opportunity to get paid by the ISO similar to an energy generator, making it a more level playing field. End use customers can participate in these DR programs by using either distributed generators or energy management control strategies to reduce their load in response to a price or emergency signal. NYISO: The NYISO currently operates five demand response programs [20,21]. Reliability based DR programs o Emergency DR Program (EDRP): this is a voluntary program where participants are paid the higher of $500/MWh or the real time LMP for responding during system emergency. o Special Case Resource Program (SCR): this program pays participants higher of their strike price (up to $500/MWh) or the real time LMP for mandatory response to emergency events. These resources can also sell capacity in the capacity market. o Targeted Demand Response Program (TDRP): this program was initiated in July 2007 to pay EDRP and SCR resources higher of $500/MWh or the real time LMP to respond to local reliability events within a particular zone. This avoids need to activate emergency events for entire zone to address local reliability concerns. Economic DR program: o Day Ahead DR Program (DADRP): DADRP allows customers to submit bids in day ahead energy market for curtailing load above a certain bid floor price. The bid floor price was increased in 2004 to $75 from original value of $50. Customer must respond if their bid is cleared in the day ahead market and curtailment cannot be achieved with use of backup generators. Failure to curtail results in a penalty defined by the MW curtailment shortfall multiplied by the greater of the corresponding day ahead or real time market price. Ancillary services: o Demand Side Ancillary Service Program (DSASP): DSASP was initiated in 2009 and allows DR to provide three ancillary services to NYISO markets. These include regulation, synchronous reserve and non-synchronous reserve.
1555
Fig. 1 shows that over the years customers have preferred the SCR program which is most lucrative due to the locational capacity market revenues. Participation in the EDRP has gradually decreased since 2003, when the EDRP and SCR program participation was made mutually exclusive. The figure also shows that participation in the DADRP program has not changed significantly over the years, and suggests that NYISO is not able to attract more customers to participate in the economic DR program in recent years [21]. PJM: The PJM currently operates seven DR programs [22–24]. Reliability based DR programs o Emergency DR Program: similar to the NYISO program, this is a voluntary program for reliability that offers energy payments to customers that reduce load during a system emergency. The payments are the higher of $500/MWh or the zonal LMP for the hour. There is no penalty for noncompliance, and this program is rarely utilized by PJM (only five times in past 3 years). o Interruptible Load for Reliability (ILR): this program is similar to the NYISO SCR program and allows ILR to receive capacity payments for load reduction achieved during system emergency. Load curtailment is mandatory for up to 10 events each year, lasting up to 6 h per event. To encourage higher participation of DR in capacity markets, ILR resources are allowed to register by March each year for upcoming delivery year (May–April). During 2008–2009, ILR accounted for 3489 MW out of a total of 4497 MW registered under emergency DR. Last year FERC ruled in favor of eliminating the ILR program option to eliminate the favorable treatment of DR under RPM. Thus 2011–2012 is the last delivery year for ILR participation. o Demand Resource (DR): Demand Resource is also eligible to receive capacity revenues through PJM’s Reliability Pricing Model (RPM). This resource has to participate in the forward capacity auction conducted by PJM 3 years before the delivery year along with other generation resources. So far majority of the DR resources had opted for ILR option to participate in capacity markets. Since DR will be the only option available from 2012 to 2013, all DR resources interested in receiving capacity credits in that year need to register as DR in May 2009. Some of the current ILR participants may find it difficult to accept the financial risk of penalties by committing resources 3 year before the delivery year. This may reduce total amount of DR available as a capacity resource for PJM.
Fig. 1. NYISO DR registration in MWs during 2001–2008 (source: NYISO, 2009 [21]).
1556
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
o Energy Efficiency Resource (EE): in 2009, PJM started allowing energy efficiency resource to receive capacity credit for up to 4 consecutive years. An EE Resource is a project that involves the installation of more efficient devices/equipment, or the implementation of more efficient processes/ systems, exceeding then-current building codes, appliance standards, or other relevant standards, at the time of installation. Similar to Demand Resource, an EE resource also has to participate in the forward capacity auction conducted by PJM 3 years before the delivery year. The nominated EE value is the expected average demand (MW) reduction during the hours between 2 p.m. and 6 p.m. during all days from June 1 through August 31 (PJM, 2009).
Economic DR Program: Under this program PJM pays the customer the difference between the LMP and the generation and transmission (G&T) components of the customer’s bill. PJM offers this economic DR program in both its day ahead and real time markets. o Day Ahead Demand Response Program (DADRP): this is a mandatory program, where participant submits bids in day ahead energy market and faces penalty for any non-compliance. o Real Time Demand Response Program (RTDRP): RTDRP is a voluntary program, where customer needs to notify PJM of intent to curtail load at least 1 h before load reduction is performed. Before January 2008, PJM also included an incentive payment for economic DR participants when they responded above a strike price of $75/MWh (similar to NYISO). The incentive payment equaled the G&T component, and thus till 2007, DR participants were eligible to receive full LMP value for curtailment during hours when LMP was above $75/MWh as is the case in NYISO. FERC [25] upheld the expiration of the incentive payments in an order under Docket EL08-12-000, issued on 31 December 2007. Ancillary Service DR Program: In 2006, PJM started allowing DR to provide frequency regulation, synchronized reserve and day ahead scheduling reserves in ancillary service market. DR resources are required to bid into ancillary service markets and respond to any event similarly to a generator. As of 2008, there were no DR resources providing frequency regulation in PJM. Although PJM has not released information about amount of DR providing synchronized reserves, these resources received almost $5M in payments during both 2007 and 2008 [26,27]. Emergency DR participation has increased more than four times over past 3 years, particularly as a result of substantial increase in capacity revenues under the RPM capacity markets, as compared to historical capacity revenues. This can be seen in Fig. 2. Table 2 provides a quick summary of current status of DR programs in these electricity markets.
Fig. 2. PJM DR program registration (2006–2008) [26,27,30].
wind) in the supply mix over next 3–5 years. Due to the variability of wind, the systems operators are concerned with reliability of the grid as well as price volatility. During some recent discussions, DR is being proposed as one of the solutions that could balance the variability in supply [28]. Although DR has proved to be an effective resource to meet reliability needs in both PJM and NYISO, the amount of DR as well as the type of response desired to mitigate variability of large scale wind or solar is significantly different. Current DR participation in these markets is clearly not sufficient to provide a viable solution for matching variability of resources like wind on a continuous basis, especially in scenarios with 10% or higher penetration of renewables. As mentioned earlier, NYISO [21] has failed to attract any additional load curtailment in the DADRP program since 2003. On the other hand the registration in PJM’s economic DR programs has grown substantially over the years to 3290 MW in 2008 [26]. At the same time, a distinction must be made between loads that are registered to participate in the PJM economic DR program, and the amount of load that actively participates in the market [29]. During 2006, there was 1475 MW load registered under the economic DR program and an additional 1081 MW load registered under the emergency DR program [30]. However, during the summer of 2006 when PJM registered it’s all time peak load of 144.6 GW, there was only 325 MW of DR cleared in the economic DR program [33]. This trend has continued in recent years and data released by PJM in 2009 indicates that out of over 4800 sites registered in economic Table 2 Summary comparison of NYISO and PJM NYISO
PJM
Population Served States
19.2 Million NY
Peak load Types of DR programs offered
33.9 GW Economic (DADRP), Emergency (EDRP, SCR, TDRP), Ancillary Services (DSASP) EDRP/SCR/TDRP: 100 kW DADRP/DSASP: 1 MW 1774 MW (SCR þ EDRP)
51 Million DE, IL, IN, KY, MD, MI, NC, NJ, OH, PA, TN, VA, WV and DC 144.6 GW Economic (DADRP, RTDRP), Emergency (EDRP, ILR, DR, EE), Ancillary Services
4497 MW (ILR þ DR)
320 MW
3290 MW
3.1. Challenges for economic demand response in PJM Minimum size
With the impressive growth of DR in these markets as well as plans of implementing smart grid, some policy makers and market participants are anticipating that DR will be able to play a much bigger role in electricity markets in near future. Currently most of the electricity markets, including NYISO are anticipating a significant growth in penetration of renewable resources (particularly
Registered emergency DR Registered economic DR
100 kW (for all)
Adapted from Walawalkar et al. [31], FERC [32], and ISO websites.
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
1557
Fig. 3. NPV of potential DR revenues over 5 years with and without incentive (source: Walawalkar et al. [23]).
DR programs, maximum of 301 sites actually curtailed load under DADRP/RTDRP program in any month during 2007–2008 [26,27]. This situation needs to change dramatically by encouraging more active participation of DR in energy and ancillary service markets as well as development of new DR programs that will provide appropriate price signal for customers to mitigate fluctuations in renewables from hour to hour. One of the issues facing economic DR participants in PJM is the elimination of incentive for participation in economic DR program in December 2007 [23]. As a result of the expiration of incentives for DADRP and RTDRP, instead of receiving full LMP for load curtailment, participants now receive payments from PJM based on difference between the LMP and the generation and transmission (G&T) components of the customer’s bill. Thus DR participants may have increased their strike price so that the marginal payments from DR program participation are equalized with and without the incentive payment. For instance, customers with a G&T rate of $50/ MWh may change their strike price from $75/MWh to $125/MWh to retain the same payment from PJM for curtailing load. The total payoff to any individual customer from economic DR participation depends not only on the strike price and incentive structure, but also the distribution of market prices that can change from year to year. Fig. 3 shows the results of a simulation conducted to project effect of removal of incentive on revenues for economic DR participants in PJM [23]. We have incorporated uncertainty due to fluctuations in energy prices from year to year into our analysis by performing a Monte Carlo simulation on the expected annual DR revenue stream, using energy market data from PJM during 2004– 2007 [34]. We modeled annual revenues based on participant’s strike prices, with and without the DR market incentives. For each year t and strike price k, we assumed that annual DR revenues (p) follow a triangular distribution with the minimum value equal to the sum of hourly revenues in 2004 (the lowest revenue year in our sample), the maximum equal to the sum in 2005 (the highest), and the most likely value equal to the average of the four years. We generated 1000 realizations of the discounted present value of expected annual revenues over a five year time horizon, assuming a customer with an internal discount rate of 10%; we did this for a $75 strike
price with the incentive, for a $75 strike price without the incentive and for a $125 strike price without the incentive. The simulation indicates that before expiration of incentives, any economic DR resource offering 1 MW load curtailment at $75/ MWh would receive a discounted gross revenue stream of $610,000 over five years with a 50% probability. On the other hand, without the incentive payment and customer adjusting her strike price to $125/MWh (to receive the DR payment of $75/MWh), the NPV would fall by roughly a factor of five, to $107,000 as shown in Fig. 3. The change in the potential DR revenue in economic DR program due to change in the incentive structure can also be seen in the data released by PJM as shown in Fig. 4. The figure indicates that although overall DR revenues for all participants have grown substantially over past 3 years, revenue for economic DR participation decreased from approximately $60M in 2007 to $20M in 2008. As a result although overall participation in the DR is increasing over the years considering the reliability based programs, increasing active participation in the price-responsive DR will be a challenge. PJM Board of Directors acknowledged this in a press release on 26 June 2009, which announced that PJM will reintroduce incentive payments for demand response for the top 9% of the hours from the previous year.1
4. Opportunities for growth Currently the majority of DR program participation is through use of backup generators or use of curtailable loads. The vast potential of existing building EMS systems and energy management strategies is still untapped, not to mention future possibilities as home automation and smart appliances become standardized along with deployment of smart grid infrastructure as shown in Fig. 5. Also most of the energy efficiency project developers are not
1 Statement of Terry Boston, President and CEO, on behalf of the PJM Board of Managers is available at http://pjm.com/w/media/about-pjm/newsroom/2009releases/20090626-pjm-board-statement-regarding-dr-in-pjm-markets.ashx.
1558
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
Fig. 4. PJM demand side response estimated revenues 2001–2008 (source: PJM [35]).
familiar with the electricity markets and the newly developed option of receiving capacity revenues for EE resources. In a competitive market, new generating capacity will not be added until prices rise sufficiently above the cost of new facilities. The existing energy efficiency and control technologies, however, can be a great tool in expanding DR program participation. Table 3 shows a sample of some of the control strategies that can be used by different types of customers for participating in various DR programs. Ongoing advances in the research focused on develop technologies, tools and standards to make commercial and residential building systems more energy efficient and productive can also facilitate DR participation. Some of key areas of research related to high performance buildings include new building and equipment standards, integrated building systems, and solid-state lighting. Besides making existing buildings energy efficient, there is also an increased effort to design new net zero energy buildings by integrating renewable technologies like solar, geothermal, wind, biomass and fuel-cell and hydrogen technology.
We briefly discuss some of the emerging technologies and standards that hold a lot of promise for enabling DR in residential and commercial segment. ZigBee enabled DR systems: demand response systems for monitoring energy usage in homes are amongst the most promising of new technologies. The technology that is driving a new wave of home automation is Zigbee, a low-power wireless networking standard designed for controlling and monitoring applications. On one hand, utility companies are increasingly evaluating Advanced Metering Infrastructure (AMI) networks to provide two-way communication between the home and utilities via wireless smart meters. On the other hand, wireless Home Area Networks (HANs) are being adopted by home-owners to better gain whole-house entertainment control. Zigbee is the wireless technology that helps connect AMI networks and HANs. This can allow home-owners and utilities to communicate in real time and manage energy consumption.
Fig. 5. Components of a Demand Response Program implementation (Walawalkar et al. [31]).
R. Walawalkar et al. / Energy 35 (2010) 1553–1560 Table 3 Control strategies and applicable DR programs for different customer types Customer type
Equipment/ Control building strategy component
DR programs
Residential
Air Cycling/forced conditioners demand shedding Water Cycling heaters Pool pumps Cycling Electric Scheduling stove
U
U
U
U
U
U
U
U U
U
Demand limiting U during on peak period Chillers Pre-cool bldg HVAC over nightstorage DX Forced U demand scheduling Refrigerator/ Prioritized freezer demand shedding Lighting Scheduled on/off Lighting Scheduled dimming of selected circuits
U
Emergency Economic Ancillary (capacity) (energy) (Reg & Res)
Commercial Chillers
Industrial
Chillers
Demand limiting on time schedule Electric Demand limiting furnace through heat stages Curtail (during U Electric furnace peak period) VSDs Limit output on scheduled basis Well pumps Defer during U peak Production Prioritized equipment demand on selected units
Restaurants/ HVAC shopping malls DX compressor
Chillers – demand limiting during peak Forced demand U shedding of multiple units Refrigerator/ Prioritized freezer demand shedding Electric Scheduled preSToves cooking
U
U
U
U U
U U
U
U
U
U
1559
New Building Standards: building codes have been used for nearly three decades and are a cost-effective strategy to overcome barriers to energy efficiency in buildings. Most residential buildings adopt the International Energy Conservation Code (IECC) whereas most commercial building energy codes are based on ASHRAE/IESNA Standard 90.1. According to California Energy Commission, building standards have helped save businesses and residents in California more than $15.8 billion in electricity and natural gas costs since 1975. Now with the availability of capacity revenues for EE resources from electricity markets, businesses have an added incentive for choosing newer standards that go beyond current standards. The vision for the future is that buildings are ‘‘DR enabled’’ and major appliances are ‘‘DR ready’’ from the factory. Building standards that incorporate controllable thermostats in residential and price server interface in commercial buildings will greatly help DR. On the appliance side, DR will benefit from standards that include embedded controls for HVAC and other systems. Energy storage technologies: another emerging technology area that can provide significant boost to DR capability in electricity markets is electric energy storage (EES). Currently approximately 2.5% of the total electric power delivered in the United States passes through energy storage, largely pumped hydroelectric. The percentages are significantly larger in Europe and Japan, at 10% and 15%, respectively [36]. In recent years, there has been tremendous advances made in various EES technologies, particularly in areas of advanced batteries (such as lithium ion, flow batteries and sodium sulfur), flywheel technologies that can provide frequency regulation as well as modular designs for above ground Compressed Air Energy Storage (CAES). With successful adaption of some of these technologies in commercial and industrial applications, market participants investing in EES as a behind the meter resource, can use it as DR resource.
U U
U
U
U
U
Adapted from Walawalkar et al. [31].
In-Home Energy Use Displays: in-home energy use displays provide real time feedback to the occupants about the total energy usage in the house. This information can be displayed through existing home PCs using applications such as Google PowerMeterÒ or Microsoft HohmÒ. Google PowerMeterÒ can receive information from utility smart meters and energy management devices and provides customers with access to their home electricity consumption right on their personal iGoogle homepage. On the other hand, new products are being developed that can be used to signal the need for demand response without consumers using their computers. In California, PG&E has introduced a ‘‘Demand Response Orb’’ to inform customers of Critical Peak Pricing periods and/or emergency events. The projected energy savings from the use of such displays ranges from 4% to 15%.
5. Conclusions Electricity markets have made significant progress over past decade in integrating DR into electricity market design. PJM and NYISO have been on the forefront of such developments, and currently offer a wide range of programs that allow DR to participate in these markets. In general all DR programs rely on end users deliberately altering use of equipment and systems, which generally means lifestyle or comfort changes, or changes in operating procedures. Such changes would be acceptable to end users only if the consumer has a stake in the process either through financial compensation or through improved reliability of power supply. Also it is important to realize that all DR programs are not equal. Different types of programs have specific objectives, and demand varying level of commitment from DR resources with different payoffs. Current market based DR programs are very successful in attracting participation in emergency or reliability based programs, which require relatively modest commitments from end users. If we expect DR resources to scale up to enable greater level of renewable integration, then we need to start evaluating DR technologies that are suitable for participating every day. We also need to evaluate if current programs are the right programs to provide incentives to customers for such role. For example, Economic DR programs will not provide sufficient revenues for customers to respond during off peak hours, whereas emergency programs are designed to provide response only during peak hours. On the other hand the wind energy is generally expected to have the largest share of the supply mix during night or early morning hours. We will need DR technologies such as EES that can respond in both
1560
R. Walawalkar et al. / Energy 35 (2010) 1553–1560
direction (absorb power as well as provide fill up power), and new DR programs that provide appropriate financial incentives for customers to participate. We are witnessing great advances in technology that can help DR to scale up substantially, but technology alone is not sufficient. The process of evolution of DR programs in electricity markets is not complete, and with changing needs of the market, stakeholders in NYISO and PJM need to continue efforts for increasing DR participation.
[14]
[15]
[16]
[17]
Acknowledgements
[18]
We thank Dr. Jay Apt, Dr. Lester Lave, Dr. Rahul Tongia, Dr. Granger Morgan (Carnegie Mellon University), Dr. Bruce Colburn (EPS Capital Corp.) and Dr. Seth Blumsack (Pennsylvania Stage University) for useful comments and conversations. We acknowledge timely inputs from our colleagues Rick Mancini and Rick Gilkey at Customized Energy Solutions. We also acknowledge contribution of the reviewers of this special issue, which helped us to identify areas that required further explanation.
[19]
References
[23]
[1] Electricity markets: consumers could benefit from demand programs, but challenges remain (US-GAO-04-844), A report to Senate Committee on Governmental Affairs. U.S. Government Accountability Office. Available from:, http://www.gao.gov/new.items/d04844.pdf; 2004. [2] Violette D, Freeman R, Neil C, DRR valuation and market analysis. Overview, a report for the International Energy Agency (IEA) DRR Task XIII, vol. I; 2006. Available from: http://www.demandresponsecommittee.org/id80.htm. [3] Violette D, Freeman R, Neil C, DRR valuation and market analysis, Assessing the DRR benefits and costs, a report for the International Energy Agency (IEA) DRR Task XIII, vol. II; 2006. Available from: http://www. demandresponsecommittee.org/id81.htm. [4] Rassenti S, Smith V, Wilson B. Demand-side bidding will reduce the level and volatility of electricity prices. Independent Review(3):441–5. Available from: http://www.independent.org/pdf/tir/tir_06_3_smith.pdf, 2002;6. [5] Goel L, Wu Q, Wang P. Reliability enhancement of a deregulated power system considering demand response. IEEE Power Engineering Conference Issue 2005;2:934–8. [6] Azami R, Fard A. Impact of demand response programs on system and nodal reliability of a deregulated power system. Sustainable Energy Technologies, ICSET 2008;2008:1262–6. [7] Shayesteh E, Yousefi A, Moghaddam P. ATC enhancement using emergency demand response program. IEEE Power Systems Conference and Exposition 2009;2009. PES ’09. [8] Green R. Electricity transmission pricing: how much does it cost to get it wrong?. MIT Center for Energy and Environmental Policy Research; 2004. MIT-CEEPR (Series); 04-020WP. [9] Boisvert R, Neenan B, Social welfare implications of demand-response programs in competitive electricity markets, Lawrence Berkeley National Laboratories working paper LBNL-52530; 2003. Available from: http://eetd.lbl. gov/ea/EMS/reports/LBNL-52530.pdf. [10] Lavy R, Herter K, Wilson J, Unlocking the potential for efficiency and demandresponse through advanced metering, Lawrence Berkeley National Laboratory working paper; 2004. Available from: http://repositories.cdlib.org/lbnl/LBNL55673. [11] Piette M, Watson D, Motegi N, Bourassa N. Findings from the 2004 fully automated demand response tests in large facilities. Demand Response Research Center; 2005. LBNL Report Number 58178. [12] FERC. A national assessment of demand response potential. staff report. Federal Energy Regulatory Commission; June 2009. [13] McKinsey & Company. The benefits of demand side management and dynamic pricing programs. Mckinsey and Company. Available from: http://republicans.
[20] [21]
[22]
[24]
[25] [26]
[27]
[28] [29]
[30]
[31]
[32] [33]
[34] [35]
[36]
energycommerce.house.gov/107/hearings/06222001Hearing265/swofford.pdf; May 2001. Lovins A, Datta E, Feiler T, Rabago K, Swisher J, Lehmann A, et al. Small is profitable: the hidden economic benefits of making electrical resources the right size. Rocky Mountain Institute; 2002. The Western States power crisis: imperatives and opportunities, EPRI white paper. EPRI. Available from: http://www.pserc.wisc.edu/ecow/get/publicatio/ specialepr/gridreliab/westernstatessynthesi1.pdf; 2001. ICF Consulting. Economic assessment of RTO policy. FERC. Available from: http://www.ferc.gov/legal/maj-ord-reg/land-docs/rtostudy_final_0226.pdf; 2002. PJM. A demand response roadmap in PJM symposium on demand response II. PJM Interconnection. Available from: www.pjm.com; 2008. PJM. Summary of PJM-initiated load management events. PJM Interconnection. Available from: http://www.pjm.com/planning/res-adequacy/ downloads/20070907-alm-history.pdf; Sept 2007. NYISO. Historic EDRP and SCR activation information. Nyiso. Available from: http://www.nyiso.com/public/webdocs/products/demand_response/general_ info/event_history.pdf; 2008. NYISO. State of the market report 2008. New York ISO. Available from: www. nyiso.com; 2007. NYISO, Annual compliance report on demand side management and new generation projects submitted to FERC; January 2009. Available from: http://www.nyiso.com/public/webdocs/documents/regulatory/filings/2009/01/ DemandSideMgmtCompFiling_1_15_09.pdf. PJM, Demand response program information. Available from: http://www.pjm. com/markets-and-operations/demand-response.aspx [Information accessed on 1 May 2009]. Walawalkar R, Blumsack S, Apt J, Fernands S. An economic welfare analysis of Demand Response in PJM electricity market. The Energy Policy October 2008;36(10). Brattle Group, 2007. Quantifying demand response benefits in PJM, A report for PJM Interconnection and Mid Atlantic Distributed Resource Initiative (MADRI). Available from: http://www.energetics.com/madri/pdfs/BrattleGroup Report.pdf. Docket EL08-12-000. Federal Energy Regulatory Commission (FERC); 2007. issued on 31 December 2007. O’Neil J. DSR activity report presented to PJM Demand Response Steering Committee. PJM 2009. Available from: http://www.pjm.com/committees-andgroups/committees/drsc.aspx; 2008. Gockley B. DSR activity report presented to PJM Demand Response Steering Committee. PJM 2009. Available from: http://www.pjm.com/committees-andgroups/committees/drsc.aspx; 2007. Stadler I. Power grid balancing of energy systems with high renewable energy penetration by demand response. Utilities Policy 2007;16(2):90–8. Hopper N, Goldman C, Bharvirkar R, Engel D. The summer of 2006: a milestone in the ongoing maturation of demand response. The Electricity Journal June 2007;20(5):62–75. Kujawski D. DSR activity report presented to PJM Demand Response Steering Committee. PJM 2009. Available from: http://www.pjm.com/committees-andgroups/committees/drsc.aspx; 2006. Walawalkar R, Tongia R, Colburn B. Web-enabled metering and controls for demand response. In: Capehart BL, Capehart LC, editors. Web based enterprise energy and building automation systems. Lilburn, Georgia: The Fairmont Press, ISBN 0849382351; 2007. p. 299–314. FERC. Electric power markets: national overview Available from: http://www. ferc.gov/market-oversight/mkt-electric/overview.asp. Covino S, PJM economic load response: summer 2006 potential and performance. A presentation to the Demand Side Response Working Group; November 2006. Available from: http://www.pjm.com/committees/workinggroups/dsrwg/downloads/20061130-item-09-summer-2006-performance. pdf. PJM locational marginal pricing data. Available from: http://www.pjm.com/ markets/jsp/lmp.jsp. PJM. PJM demand side response estimated revenue, Presented to PJM Demand Response Steering Committee; 2009. Available from: http://www. pjm.com/committees-and-groups/committees/w/media/committees-groups/ committees/drsc/postings/demand-side-response-estimated-revenue.ashx. Walawalkar R, Apt J, Market analysis of emerging electric energy storage systems, A report developed for Department of Energy (DOE) and National Energy Technology Labs (NETL), DOE/NETL-2008/1330; July 2008.