- Email: [email protected]
A review on price-driven residential demand response
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
Contents lists available at ScienceDirect
Renewable and Sustainable Energy Reviews journal homepage: www.elsevier.com/locate/rser
A review on price-driven residential demand response a,⁎
b
c
Xing Yan , Yusuf Ozturk , Zechun Hu , Yonghua Song
T
c,d
a
Energy Internet Research Institute, Tsinghua University, Beijing 100084, China Department of Electrical and Computer Engineering, San Diego State University, San Diego 92182, CA, USA Department of Electrical Engineering, Tsinghua University, Beijing 100084, China d Department of Electrical and Computer Engineering, University of Macau, Macau 519000, China b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Critical peak pricing Demand response Real-time pricing Time-of-use pricing
Smart grid enables the two-way communication between the suppliers and consumers. Price-driven demand response (PDDR) is one of the important demand response categories that uses price of the energy as control signals to affect consumers’ electricity consumption. The current PDDR programs include critical peak pricing (CPP), time-of-use (TOU) pricing, and real-time pricing. In this paper, we provide a review of the PDDR studies. Detailed evaluations on advantages and disadvantages of each PDDR are provided. Concerns and future research challenges on PDDR are also addressed. It is believed that with the installation of smart meter infrastructures at residential households, price signal can be an efficient market tool for peak demand shaving, risk and reliability management, carbon emission reduction, and energy cost reduction.
1. Introduction
response, it is still increasingly seen as a way of meeting emerging needs for better system management and concerns about energy security and environmental impact [2]. Demand response as a proactive measure can be implemented in both manual and automated ways. The manual demand response usually refers controlling use of certain appliances in different time periods of the same day. Semi-automated demand response involves a pre-programmed demand response strategy that is initiated by a person through a centralized control system. Fully automated demand response does not involve any human intervention but is initiated at a home, building, or facility through receipt of an external communications signal [3]. Although demand response is a very cost-effective way for peak demand shaving, risk and reliability management, carbon emission reduction, and energy cost reduction, industrial and residential sectors have different aspects on this topic. For industrial sector, the primary target is maximizing the profit. As electricity consumption is part of the production cost (in some industries such as chemical production companies electricity consumption is a huge part of the production cost) and demand response can reduce the electricity cost, industrial sector is able to adopt demand response very quickly. The residential sector, however, is not able to adopt demand response as quickly as the industrial sector because the primary focus of the residential sector is on increasing the comfort level. Different opinions of demand response are also rising important issues in applying such technology at the
Electricity is a very unique commodity which cannot be economically stored in large quantity. Nowadays, nearby electricity markets are interconnected with each other to improve system reliability, reduce energy cost, increase penetration of renewable resources, and reduce carbon emission. With a much larger interconnected grid, surplus electricity energy can be stored into different forms such as hydro power by pumping the water back to the dam for later use. According to the U.S. Department of Energy, smart grid generally refers to a class of technology used to bring utility electricity delivery systems into the 21st century, using computer-based remote control and automation [1]. These systems are made possible by two-way communication technology and computer processing that has been used for decades in other industries [1]. Smart grid enables the electricity energy exchange between electricity markets. It also enables integration of renewable energy resources to the energy exchange grid. Researchers are continuously working on new mechanism that can better serve the demand by either increasing supply or keeping demand bounded to the supply. While the proactive management of demand plays an important role, at times of high demand tapping into renewable energy resources or tapping into backup resources are critical components of smart grid solutions. All of these actions can also be referred as demand response. According to the current understanding of the impact of demand
⁎
Corresponding author. E-mail addresses: [email protected] (X. Yan), [email protected] (Y. Ozturk), [email protected] (Z. Hu), [email protected] (Y. Song). https://doi.org/10.1016/j.rser.2018.08.003 Received 13 April 2017; Received in revised form 29 June 2018; Accepted 3 August 2018 1364-0321/ © 2018 Elsevier Ltd. All rights reserved.
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
large quantities. Demand response can provide ancillary services to enhance the voltage stability in power systems operating under increased uncertainty, and replace some or even all of the spinning reserves typically supplied by conventional generators [7]. This will avoid using rolling blackouts when there is not enough energy supply and in turn increase the electricity grid reliability and reduce congestion. Consumers can also benefit from demand response. Under the current regulated energy transmission and distribution systems, electricity prices are generally averaged over the entire year and cost-of-service pricing is the norm. On the other hand, the real-time electricity price varies from time to time. Utility companies have to purchase highpriced wholesale energy whenever they cannot meet the demand. An extreme example occurred on November 28, 2005, eastern Denmark. The wholesale price reached DKK 13,469/MWh (equals to $2189.17/ MWh USD), which is 60 times more than the normal price level [8]. These high wholesale energy prices are eventually included in the calculation of the averaged energy price and are paid by the consumers. Demand response programs can lower the energy consumption during peak hours. This will reduce the wholesale energy prices and in turn reduce the energy price paid by the consumers. Renewable resources, such as solar, wind, biomass, and hydro, are continuously growing and playing a very important role in supplying electricity. According to the U.S. Energy Administration Renewable Electricity Standard (RES), by 2020, every state in the United States must generate one third of their electricity from renewable resources. German Advisory Council on the Environment states that Germany is aiming to transform its electricity supply to 100% renewable energy by 2050. Demand response programs can increase the penetration of renewable resources and benefit the environment in reducing carbon emission. It can contribute to the better integration of renewable energy resources such as wind power, solar, small hydro, biomass and combined heat and power (CHP) [9,10]. Excess energy from the renewable resources can be either stored into different form for later use or supply to the grid to reduce the electricity generated by the ordinary thermal generation plants. Surplus energy from the renewable resources can be stored in a different form of energy for later use in order to minimize the overall operating and environmental costs [11]. With renewable energy accessible only at certain times, the research on energy storage is also gaining momentum. We believe demand response programs will benefit from the developments in storage technologies as well as drive the developments in the storage technologies. Currently, demand response can be divided into main categories which are further subdivided into many forms. A complete list of demand response varieties is shown in Fig. 1. In this paper we will focused on residential PDDR programs. A detailed presentation of residential PDDR programs is given in Section 3. In the incentive or event-based demand programs, direct load control and emergency response programs are voluntary programs. Consumers will not get penalized if they did not adjust their energy consumption according to the suggestion by the local utility companies or the independent system/market operator. In the direct load program, utility companies or system operators can remotely control customers’ specific appliances (large energy consumption devices) in order to respond to the demand and reliability issues. Incentives are available for involved customers. Emergency response program provides incentives in exchange for voluntary load reduction during special events. Capacity market programs and interruptible/curtailable service are mandatory programs and enrolled consumers will get penalized based on the agreement if they did not adjust their energy consumption accordingly when needed. In the capacity market program, load reduction is pre-specified and mandatory as directed by the utility companies and system operators. Interruptible/curtailable program is very similar to the emergency response program except it is mandatory. Finally, demand bidding/buyback and ancillary service market programs are market mechanisms to balance energy supply and demand based on the
residential sector. According to a study proposed by Bartusch et al. [4] with 500 interviewed consumers under Swedish power system, only 50 are participated in the proposed study. People neither want their lives to be governed by tariff rates nor their demand to be remotely controlled, and they do most certainly not want to waste their time keep tracking on their electricity usage [4]. Although everyone wants to save money on their electricity bills, sacrificing the comfort level in exchange for financial incentives is not accepted by the majority of the residential sector, especially when the savings are very small. Another important reason is that most consumers are only charged by a fixed electricity rate based on the average electricity cost even though the real time electricity prices vary from time to time. In some cases, electricity costs may vary hourly or seasonally according to the seasonal and daily variations [5]. The U.S. Energy Policy Act of 2005 mentions that each electric power company should provide customers with timebased rates [5]. Currently, most utility companies offer consumers tiered residential electricity prices based on the climate zone, seasonal information, and time of use. When a consumer's total electricity usage exceeds a certain predetermined threshold, the consumer will be subjected to different electricity rates based on the tiered price structure. We believe that research on real time energy pricing and consumer behavior modification through pricing energy will gain acceleration as utility companies roll out pricing systems that charges consumers based on time of use and amount of energy consumed. In this paper, we reviewed the PDDR studies. Advantages and disadvantages of each PDDR program are evaluated in detail with experimental results. Based on the results, a discussion section is provided including our concerns and opinions on the current PDDR programs. The rest of the paper is organized as follows: Section 2 provides a general introduction on demand response. Detailed evaluation on recent PDDR programs are explained in Section 3. Section 4 discusses the issues with recent PDDR and conclusions are provided in Section 5. 2. Demand response According to the Federal Energy Regulatory Commission, demand response is defined as changes in electric usage by demand-side resources from their normal consumption patterns in response to changes in the price electricity over time, or to incentive payments designed to induce lower electricity usage at times of high wholesale market prices or when system reliability is jeopardized [6]. The studies on demand response started on early 1980s under the demand-side management programs. It is believed that utility companies, transmission and distribution system operators, and end-use consumers can all benefit from demand response. The benefits of demand response for utility companies presents themselves as reduced capital cost, operation cost, and reduced carbon emission. Compared to the huge capital cost and long construction time of period building new generation plants to match electricity consumption at peak hours, demand response is an economical, and environmental friendly program that can be used to balance the supply and demand within the jurisdiction of electricity grid. Demand response can also help the utility companies operate their power generation plants at optimized speeds. This in turn will result in reduced fuel consumption, increased productivity and increased profit margins. Carbon emission can also be reduced as the fuel consumption efficiency is increased. Moreover, utility companies can benefit from the demand response by avoid use of diesel powered backup/emergency power plants and in turn reduce carbon emission. Because of such advantages, demand response is among the top priorities of energy utility companies. For transmission and distribution system operators, demand response is also an essential asset. The transmission and distribution system operators are in charge of balancing the demand and supply within the jurisdiction of electricity grid at all time. Electricity is a special commodity which cannot, currently, be economically stored in 412
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Time-of-use Pricing Critical Peak Pricing Price-Driven Real-time Pricing Peak Time Pricing Demand Response Direct Load Control Emergency Demand Response Programs Capacity Market Programs Incentive or Event-Driven Interruptible/Curtailable Service Demand Bidding/Buyback Programs Ancillary Service Market Programs Fig. 1. Demand response categories.
Smart metering (advanced metering) is a measuring system that records customer's consumption (and possibly other parameters) on a timely basis (hourly or more frequently) and provides hourly or even minutely transmission of measurements over a communication network to a data consolidation point [5]. With the improved infrastructure in residential sectors with increased penetration of smart meters, real-time electrical energy consumption monitoring within the smart grid became widely available. Fig. 3 shows a detailed residential 24-h electricity consumption on an hourly basis. The U.S. Energy Policy Act of 2005 announced that it is the policy of the United States to encourage “timebased pricing and other forms of demand response” and encourage states to coordinate, on a regional basis [13]. State energy policies aims to provide reliable and affordable demand response services to the public [13]. New techniques are able to apply to the residential sectors and enhance the PDDR. The three different residential PDDR programs are evaluated in the following sections.
determination of market clearing price. Demand bidding/buyback program gives large energy consumption parties the ability to negotiate the price for the amount of load reduction. In ancillary service market program, negotiated amount of load reduction with the corresponding prices are used as the reserve energy for the electric grid. It is also referred as spinning reserve energy. Incentives based on the spot market prices and amount of gain from load reduction are paid to the participating customers when called by the system operators. 3. Residential price-driven demand response Price-driven demand response uses electricity prices as control signals to motivate consumers to change their energy consumption. The goals of applying these electricity price signals are to reduce overall energy consumption and shift some of the peak load into off-peak hours at the same time. PDDR contains time-of-use (TOU) pricing program, critical peak pricing (CPP) program, and real-time pricing program. The price patterns representing all 3 PDDR programs are shown in Fig. 2. PDDR is first practiced at the industrial sectors where energy consumed is part of the production cost. Higher energy costs results higher production costs. Therefore, company's large amounts of electrical energy such as chemical production companies are willing to participate in the PDDR programs in order to benefit from the cheaper electricity prices during off-peak hours. Although there were some early stage researches focusing on residential PDDR in the 1980s, due to lack of monitoring on the daily or even hourly electricity consumption, only limited techniques were able to applied to the PDDR in the residential sectors.
3.1. TOU pricing program The TOU pricing program shown in Fig. 2 uses a static pricing scheme based on predefined price values. The price pattern usually maintains the same prices during each season. New price portfolios are proposed yearly to cover the operation cost and long-term investment of the utility companies. The proposed price portfolios are approved by the state/provincial price regulation party whose chief objective was to protect the consumers from the inevitable consequences of a monopoly
Fig. 2. PDDR programs: a). TOU Pricing, b). CPP, and c). RTP [12]. 413
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
PDDR program, an approximate 10% peak shifting can be achieved [8,15–24,29] and consumers did save money on their electricity bills. However, according to Fig. 4, because participated consumers scheduled their large energy consumption appliances right after the price drop, new demand peak is created at the beginning of off-peak hours. These new demand peaks are usually worse than the ones under the static pricing. On the other hand, an overall energy consumption deduction is observed in all involved case studies which implies that TOU pricing did raise certain awareness on energy consumption. A similar observation is concluded in [14] that while households reduce peak demand significantly after the introduction of TOU tariffs and associated information, there is little incremental response to increasing differentials between peak and off-peak prices. 3.2. CPP program
Fig. 3. Daily energy consumption pattern for an individual household [12].
The CPP program is an event driven program that uses high electricity prices as control signals to affect electricity consumption during peak hours. CPP may be imposed if the system is severely constrained as in cold winter periods or warm summer periods for a limited number of hours [30]. Under these instances, consumers can receive incentives by either reducing peak consumption or shifting energy consumption to off-peak periods to protect system reliability. Since system under severe constrains does not happen on the daily basis, CPP is not a daily demand response program. Moreover, CPP values are higher than TOU pricing values. The advantages of CPP program includes: (1) easy to follow, (2) effective on shifting peak energy consumption, and (3) visualized incentives. Because CPP program is an event-driven tariff whose primary focus is ensuring system reliability, households are more willingly to participate in such event. A survey of 483 households in California by Herter in [31] concluded that statically a significant average participant response in each hour. Day-ahead CPP values are delivered to consumers through social media such as newspaper, text messages, and websites. This makes CPP program very easy to understand and follow by average consumers. The disadvantages of CPP program includes: (1) event driven and (2) not effective on reducing energy cost and carbon emission. Due to the fundamental event-driven principle of CPP program whose primary focus is ensuring system reliability, such program usually does not apply on a daily basis as system is not facing severe constrains on a daily basis. Therefore, CPP is not an effective program in improving energy consumption efficiency, reducing energy cost, and reducing carbon emission. Results of some previous case studies on CPP are shown in Table 2. CPP programs are usually applied in combination with other PDDR programs to optimize the results of shifting peak demand to off-peak periods.
industry. The successful results from the industrial sector containing a survey of 43 TOU pricing programs show that the cost savings can vary from − 72.0% to + 82.6%, depending on specific utility programs and switching strategies involved [14]. The advantages of TOU pricing program include: (1) easy to follow and (2) stable daily participating ratio. Because the TOU pricing portfolios maintain the same prices during the same season, it is very easy for consumers to understand, follow, and plan their daily electricity consumption portfolios. As a result, the daily percentage of people participating in the TOU pricing program is quite stable compared with other two PDDR programs. On the other hand, the advantage of the robust price portfolios of the TOU pricing program is its own fundamental disadvantage at the same time. Fig. 4 shows a daily electricity consumption portfolio comparison between static (fixed) and TOU pricing programs. It can be seen from Fig. 4 that there is a demand peak occurred right after the price dropped from peak hours to off-peak hours. Comparing with the demand curve under the static pricing program, TOU pricing program did reduce the overall electricity demand during peak hours, but created a new and much bigger demand peak during the off-peak hours. Although Fig. 4 only represents the electricity consumption of a single household, aggregated residential electricity consumption has a great chance of creating a new and much bigger demand peak from the residential sector, especially when increased plug-in electric vehicles are involved. On top of that, adjusting the time interval of peak and offpeak hours will not reduce the demand peak but shift it into different hours. Summarized results from previous studies are shown in Table 1. The best saving of 21.07% in financial incentives is obtain in the case study proposed by Giorgio et al. [29] on a low number of loads. However, the question is, will you sacrifice the promotional level of comfort in exchange with the financial incentives. Meanwhile, Gottwalt et al. [13] explored that an individual household can expect rather low benefits of an investment in smart appliances when TOU prices hardly exceed the static prices. According to the previous studies on TOU
3.3. RTP program Under the RTP program, electricity price varies continuously in response to the wholesale market prices. The dynamic electricity prices
Fig. 4. Daily consumption under different price programs [15]. 414
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Table 1 Case study results of TOU pricing. Studies
Place and year
Peak shifting
Comments
Bartusch et al. [4] Cosmo VD [14]
Sweden 2005–2008 Ireland 2007
0.1–2.5% N/A
Torriti [16]
Northern Italy 2010
N/A
Woo et al. [17]
Canada (British Columbia) 2007–2008
2.6–9.2%
Faruqui et at [18] CRA [19] Schlomann [20]
3–6% 13% 3–10%
Quantec [21]
U.S. 2010 U.S. (California) 2003–2004 Germany (Freiburg) 1970 and 1980s U.S. (Portland) 2004
Peak shifting is very limited during major consumption season (winter) 5000 Households involved. Peak demand reduced significantly, but little peak shifting observed. Consumption increased by 13.69% consumers’ electricity bills decreases by 2.21% 75.6% of substations experienced an increase in electricity demand during peak periods 2.6% at 2:1 peak-to-off-peak price ratio 9.2% at 12:1 peak-to-off-peak price ratio Participating customers are enthusiastic volunteers Summer afternoon peak
Hammerstrom [22] CER [23] Gyamfi [24] Ozturk [25]
U.S. (Washington) 2007 Ireland 2011 New Zealand (Christchurch) 2013, San Diego
Yan et al. [26–28]
2015, San Diego
• • •
1970 and 1980s
7–15% in winter 2% in summer 5–20% 8.8% 7–10% 2006 Demand forecasting and peak pricing Peak shifting
Owners set automatic response based on comfort and price preference
A residential energy management system has been developed. The home controller system controls appliances with user comfort in mind. A real time energy pricing system is designed and fully integrated to the smart grid via Green Button API. An android application deployed in cooperation with the local utility company. The energy residential energy management application “Energyelastics” is available on google app store.
applications with smart meters.
are available to the public an hour or sometime one day ahead. RTP programs can help utility companies better distribute the electricity price reflecting the demand-supply elastics. These dynamic price signals can motivate consumers to adjust their electricity consumption accordingly in order to achieve financial benefits. RTP signals combined with automation in end-use systems have the potential to deliver even more benefits to operators and consumers [36,37]. Some previous case studies on RTP program are shown in Table 3. Dynamic pricing remains as a new idea for residential customers [44]. Without fully implemented smart meter infrastructures, there is not enough hardware support to start with the RTP demand response program at the first place. Application of residential demand response (DR) programs is currently realized up to a limited extent due to customers’ difficulty in manually responding to the time-differentiated prices [45]. Even though at some places, the smart meter infrastructures are available in the residential sectors, the test of consumers’ acceptance of dynamic pricing remains very limited. Getting customers to voluntarily embrace dynamic pricing will take an ongoing, long term series of engagements reminders, pokes, and prods [44]. EnergyElastics [26–28] application is one of the first consumer advisory application that is fully integrated to the smart meter infrastructure enabling utility company to push real time prices to the user and monitor energy consumption in response to the prices pushed. It is integrated to the San Diego Gas and Electric smart meter infrastructure through Green Button API which is currently standardized for interfacing third party
3.4. Numerical example To better understand the effects of different pricing methods, a numerical example of a single household with a typical 24-h load profile from [27] is shown in Fig. 5. It is assumed that an upper level of 20% DR can be applied from hour 17–20 and shifted these demand starting from hour 21–23. The proposed static, TOU, CPP, and RT pricing demand response programs with experimental values are shown in Fig. 6. Numerical evaluations of applying different pricing methods are available in Table 4. According to Fig. 5, we can obtain that if a maximum of 20% DR is achieved, a new peak load is created at hour 21 as 365.7 W instead of 328.6 W at hour 19. This is an increase of 11.29% in peak load with a lower electricity cost. The new and bigger peak demand will create a new potential safety issue to the utility side as the original purpose of introducing the PDDR is to reduce the peak load rather than creating a new and a bigger one. Therefore, according to the results in Fig. 5, by applying PDDR, the utility side could face possible new critical peak load situation. According to Table 4, it can be obtained that with an upper level of 20% DR applied with different DR programs, some level of savings can be achieved by shifting some or all of the flexible demand into later time with lower electricity prices. The shifted demand may include taking a shower, plugging electric vehicles, doing laundries, and
Table 2 Case study results of CPP. Studies
Place and year
Peak shifting
Comments
Ericson [30]
Norway 2003
N/A
Herter et al. [31]
California 2004
1.9–10%
Aubin [32]
France 1996
15–45%
The study estimated that households with energy management system and electricity heating may favor the CPP program over flat rate program by 10%. 483 households involved. Results indicated that larger users respond more in both absolute and percentage terms. Households did not respond more to the higher CPP rate Experiment was conducted using newspaper to communicate with consumers about the day-ahead electricity prices. Although the experiment was successful, less than 20% of the customers have chosen this option.
Country Energy [33] Faruqui et al. [34] Renner et al. [35]
Australia 2005 North America 2010 Sweden 2011
2.4% 13–20% or 27–44% Approximate 50%
Lower value is without direct load control (DLC) while higher value is achieved with DLC allied. Customers given energy saving tips and day-ahead warnings of critical periods. Involved households are willing to support local utility.
415
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Table 3 Case study results of RTP. Studies
Place and year
Peak shifting
Comments
Avci et al. [38]
PJM 2011
4.3–6.7%
Yousefi et al. [39]
U.S. (New England) 2010
Approximate 5%
Yoon et al. [40]
U.S. (Austin, TX) 2012
12.8–24.7%
The proposed study only focused on the operation of heating, ventilation, and air conditioning (HVAC) system. The comfort level is also considered in the simulation. Simulated results is modeled based on the learning capability of an intelligent retail energy provider agent using day-ahead RTP program. Only focus on HVAC system. 12.8% sacrifice on temperature in exchange for 10.8% saving under RTP program.
Moghaddam et al. [41] Hammerstrom et al. [42]
Iran 2007 U.S. (Washington State) 2007 U.S. (Chicago) 2009
4.2–4.9% 5–20%
Owners set automatic demand response based on predefined comfort and price level.
5–14%
Day-ahead RTP.
Allcott [43]
programs in Table 5.
4. Discussions 4.1. Concerns of PDDR programs Although most studies presented above show that residential customers did respond to time-varying electricity prices and some level of peak demand shifting are achieved, a detailed analysis by Reiss et al. [46] indicates that significant proportion of households do not response to price. The level of response from a household is proportional related to the household's income and energy consumption. When household's income increases, the response decreases. Response of the lowest income households (less than $18,000 USD) were almost 50% higher than that of the highest income households (more than $60,000 USD). 44% of the involved households did not show any price responsiveness. It is worth to note that households with major appliances such as heating and air conditioners responded the most [38,40,46]. According to the previous PDDR programs [4–44], consumers’ education and awareness on their daily energy consumption has significant impact on the success of the program and customers’ feedback. Fewer than 50% of the US households have programmable thermostats, and even worse, the US Environmental Protection agency estimates that 30% or more of the US households with programmable thermostats are not using their thermostat's programing feature, instead they put them in hold mode and operate them manually [38,47,48]. Gyamfi et al. [24] explores that voluntary participation and behavior change in demand response are the critical issue for the developers of peak demand management programs. Another fact that effects the response to the price signals is that the residential consumers will not always be available to manage energy resources and decide, based on price signals and preference/needs, the best response actions to implement or the best usage of the electricity provide locally [49]. Maybe when commercialized storage units are available for residential consumers to store electricity energy from renewable resources, more people are willing to participate in the PDDR programs [50]. The current researches on PDDR programs also raised number of concerns. Torriti [16] concluded that issues with evening peaks were not resolved using TOU program and led to increment in electricity demand for substations at peak time. He also pointed out that on the positive side, the lesson from small combined heat and power systems suggest that two-way direct pricing approaches bring about better integration with the electricity supply. On the negative side, the attractiveness of the TOU pricing from a policy perspective is not matched by economic rigor [22]. Ericson [30] also pointed out his concern that time-differentiate tariffs may attract consumers who benefit without responding to the price, simply because they have a favorable consumption pattern. Moreover, higher prices discriminate against lower income households [51]. The lack of quantitative understanding of consumers’ behavior and end-use activity adaptive capacities are a significant barrier to design and deploy effective demand response
Fig. 5. 24-h electricity demand of a single household.
Fig. 6. Different pricing methods within a 24-h period. Table 4 Numerical example evaluation of different pricing methods. Type
Electricity cost ($ USD)
Electricity cost with DR ($ USD)
Percentage of savings (%)
Static TOU CPP RTP
2.6476 2.6361 2.9370 2.7276
2.6476 2.4522 2.8791 2.6145
0.00 7.50 2.01 4.33
turning on/off air conditioning. Depends on the level of comfort the users are willing to sacrifice such as shifting one and up to all of the above demand at a later time, a maximum percentage of savings of 7.5%, 2.01%, and 4.33% can be achieved accordingly using TOU, CPP, or RTP programs. It is worth to note that DR caused an increase of 11.29% in peak demand comparing with a maximum of only 7.5% in saving. By evaluating the numerical examples and reviewing the literatures, we summarized the characteristics of the three different PDDR 416
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Table 5 Key information of different PDDR programs. Type TOU
CPP
RTP
Type of DR can be applied ● ● ● ● ● ● ● ●
Manual operated DR Semi-automated DR Fully-automated DR Manual operated DR Semi-automated DR Fully-automated DR Semi-automated DR (partially) Fully-automated DR
Advantages
Disadvantages
● Easy to follow ● Stable daily participating ratio
● May shift the old peak load into different time
● ● ● ●
● Event driven ● Not effective in reducing energy cost and carbon emission
Easy to follow Effective on shifting peak energy consumption Visualized incentives Reflecting the demand and supply relationship
● Hard to participate at the moment
For instance, Darby et al. [2] points out the significance of thermal loads, supply mix, demand-side infrastructure, market regulation, and the framing of risks and opportunities associated with demand response. Siano [57] concluded further researchers are required to discuss regulatory and policy recommendations in order to ensure technical functioning and non-discriminatory physical access to all parties and precise rules, defining roles and responsibilities of all players and assuring fair sharing of costs and benefits among all stakeholders.
programs [49]. Socolow and his team showed that houses of similar sizes, occupied by demographically similar families, with a similar set of appliances, and under the same geographical condition, varied in energy consumption by as much as 200% [52]. Other concerns are people neither want their lives to be governed by tariff rates nor their demand to be remotely controlled, and they do most certainly not want to waste their time keep track on their electricity usage [4]. Although the above studies did demonstrate increased level of awareness on residential electricity consumption [3–44], this does not necessarily lead to peak demand shifting. Residential customers will certainly not postpone preparing their dinner until off-peak periods. Meanwhile, some other studies are trying to simulate the economic model of price/ incentive responsive loads which is based on the concept of flexible price elasticity of demand and customers’ benefit function to discover each of the demand response program's elasticity based on the electricity price before and after implementing demand response programs [17,39,41,53–55]. These researchers believe that the market regulator can use these flexible elastics to simulate the behavior of customers for different electricity prices, incentives, penalties, and participation level of customers. However, Bartusch et al. [4] has a different opinion on using price elastics as they believed estimating the price elasticity of demand would be questionable or even frivolous. On the other hand, positive effort such as Soares et al. [50] aims at characterizing and classifying in a detailed way the potentially controllable demand in the residential sector and assessing the impacts of implementing distinct automated demand response actions over some of the previously identified controllable end-use load is on the positive track. Moreover, demand response and dynamic pricing programs are expected to play increasing roles in the modern smart grid environment [56]. Positive demand response results are shown in major appliances such as HVAC which can respond to price signals while maintaining the comfort level within predefined tolerance [38,40,46]. Researchers are also aware of other factors that may be related to effecting peoples’ level of involvement in demand response programs. Gyamfi et al. [12] proposed the use of a hybrid engineering approach using social psychology and economical behavior models to overcome the challenges (price unresponsiveness, equity issues, and high cost of metering infrastructure) and realize the benefits of supply security and cost management. Allcott [43] argues that residential RTP should perhaps be thought of as a peak energy conservation program instead of a mechanism for peak demand shifting. Gottwalt et al. [13] pointed out that incentive mechanisms have to be designed to transfer parts of the benefits from more flexible and controllable residential loads from utilities to households. He also mentioned that it might be beneficial to evaluate the relationships between economic benefits and flexible load for residential homes [13]. Bartusch et el. [4] did conclude that in recent years, the distribution system operators have engaged in a range of activities for the purpose of promoting awareness and encouraging energy efficiency among their customers. In addition, the observed reduction in the electricity consumption may well be a sign that the efforts on building a demand response mechanism have paid off. Researchers are now starting to consider demand response as a complex system rather than a simple price-driven or event-driven mechanism.
4.2. P2P trading With the huge successes of the peer-to-peer (P2P) business model in the IT sector, the electricity system is considering the possibility of adopting such business model at the distribution level [58–60]. As an example, the primary goal in [61] is to enable the electricity trading between the citizens of a smart city in a continuous double auction market. With the increased number of aggregators entering the electricity market, it is possible for the P2P trading method to be applied in the PDDR program. An aggregator is a retail agent who represents a combined group of customers trading electricity at the wholesale market. The aggregator is able to react to PDDR signals by enabling a P2P trading among its customers. When a DR is needed, the spot market price reaches a peak value. The aggregator can use this information and design a protocol whenever a spot market release a PDDR price signal. Once a spot market price reached a threshold value, the PDDR P2P trading process will be initiated. Such trading process can help the aggregator purchase less electricity at the spot market during the DR period and reduce the spot price volatility. A completed trading cycle for an order of PDDR trading can be seen in Fig. 7. A timeslot of 15-min is used as the electricity consumption recording period. The P2P trading cycle contains 9 steps begins at minute 0 and finishes at minute 30. Minute 0 to minute 15 is the bidding cycle and minute 15 to minute 30 is the implementation cycle. Step 1: The aggregator initiates the PDDR P2P trading process. All participated consumers collect the DR information from the aggregator and the forecasted load information from their home smart infrastructure devices. Step 2: All participants place their first round of bidding orders in the aggregator operated the market. Step 3: The aggregator will match the trading orders. Step 4: The remaining quantity of electricity during the matching process and the unmatched orders will remain in the order book and enter the matching process for the next round with the new bidding orders. Step 2–4 will be repeated for a predetermined number of rounds. Step 5: At this moment, the bidding cycle is completed. All unmatched orders will be reset to 0 while all matched orders are recorded by the aggregator and related information such as matched electricity quantity, average trading price, total buying/selling quantity, and specific matching order with only quantity information will be released to all participants. 417
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Fig. 7. PDDR P2P trading mechanism.
Step 6: The aggregator will start to implement the matched orders taken place from the previous 15 min. Step 7: During the implementation process, due to the deviation of electricity, some of the matched orders cannot be fully delivered. The aggregator will settle the deviation quantity with other aggregators at the spot market. Step 8: Unmatched bidding orders will be charged with the current spot market price. Step 9: At minute 30, the aggregator will complete the implementation process of all matched orders.
temperate climates. Energy Policy 2012;49:759–69. [3] Kiliccote S, Piette MA. (Lawrence Berkeley National Laboratory). Automation development and testing of the capacity bidding program for San Diego gas and electric company. In: Proceedings of the PIER energy systems integration program. California energy commission. CEC-500-2008-059. [4] Bartusch C, Wallin F, Odlare M, Vassileva I, Wester L. Introducing a demand-based electricity distribution tariff in the residential sector: demand response and customer perception. Energy Policy 2011;19:5008–25. [5] Energy Policy Act of 2005. Public Law 109-58. August 8, 2005. [6] Federal Energy Regulatory Commission [Internet]. [cited 2015 May 18] Available: 〈http://www.ferc.gov/industries/electric/indus-act/demand-response/dem-resadv-metering.asp〉. [7] Wang D, Parkinson S, Miao W, et al. Online voltage security assessment considering comfort-constrained demand response control of distributed heat pump systems. Appl Energy 2012;96(3):104–14. [8] Aghaei J, Alizadeh MI. Demand response in smart electricity grids equipped with renewable energy sources: a review. Renew Sustain Energy Rev 2013;18:64–72. [9] Wang X, Palazoglu A, El-Farra NH. Operational optimization and demand response of hybrid renewable energy systems. Appl Energy 2015;143:324–35. [10] Yang H, Xiong T, Qiu J, et al. Optimal operation of DES/CCHP based regional multienergy prosumer with demand response. Appl Energy 2016;167:353–65. [11] Broeer T, Fuller J, Tuffner F, et al. Modeling framework and validation of a smart grid and demand response system for wind power integration. Appl Energy 2014;113(6):199–207. [12] Gyamfi S, Krumdieck S, Urmee T. Residential peak electricity demand responsehighlights of some behavioral issues. Renew Sustain Energy Rev 2013;25:71–7. [13] Gottwalt S, Ketter W, Block C, Collins J, Weinhardt C. Demand side management—a simulation of household behavior under variable prices. Energy Policy 2011;39:8163–74. [14] Cosmo VD, Lyons S, Nolan A, et al. Estimating the impact of time-of-use pricing on Irish electricity demand. Energy J -Camb Ma then Clevel Oh 2012;35(2):168–218. [15] Xu FY, Zhang T, Lai LL, Zhou H. Shifting boundary for price-based residential demand response and applications. Appl Energy 2015;146:353–70. [16] Torriti J. Price-based demand side management: assessing the impacts of time-ofuse tariffs on residential electricity demand and peak shifting in northern Italy. Energy 2012;44:576–83. [17] Woo CK, Li R, Shiu A, Horowitz I. Residential winter kWh responsiveness under optional time-varying pricing in British Columbia. Appl Energy 2013;108:288–97. [18] Faruqui A, Segici S. Household response to dynamic pricing of electricity: a survey of 15 experiments. J Regul Econ 2010;38:193–225. [19] Residential CRA. hourly load response to critical peak pricing in the state-wide pricing pilot. In: Presented by CRA international; 2006. [20] Schlomann B. Short- and long-term impact of a linear and time-dependent electricity tariff on the load curve of a utility. Berlin: ENER Joint Seminar; 1993. [21] Quantec LLC Analysis of the load impacts and economic benefits of the residential TOU rate option. Prepared for Portland general electric. [22] Hammerstrom DJ, Ambrosio R, Carlon TA, Desteese JG, Kajfasz R, Pratt RG Pacific Northwest GridWise ™ testbed demonstration projects Part I. Olympic Peninsula project. [23] CER, Electricity smart metering customer behaviour trials(CBT) rindings report; 2011. [24] Gyamfi S, Krumdieck S. Price, environment and security: exploring multi-modal motivation in voluntary residential peak demand response. Energy Policy 2011;39:2993–3004. [25] Ozturk Yusuf, Lee Gordon, Senthilkumar D, Kumar Sunil. An intelligent home energy management system to improve demand response. IEEE Trans smart Grid 2013;4(2):694–701. [26] Wright D, Yan X, Kumar S, Lee G, Ozturk Y. A Cloud-based system architecture for enabling consumer behavior modification in the electricity market. In: Proceedings of the international conference on soft computing and software engineering; 2015.
5. Conclusions The proposed paper provided a review among three different PDDR programs at the residential sector. The TOU program is easy to follow and can obtain a stable participating ratio by the end user. However, it only shift the peak into a different time period. The CPP program is also easy to follow and participants can visualize their incentives and result an effective peak shifting. The disadvantages of CPP are that it cannot be applied on a daily basis and it is not effective in reducing energy cost and carbon emission. Finally, the RTP program can reflect the dynamic relationship between the demand and supply, but it is not practical at the moment. Although the current PDDR programs achieved some level of success in peak demand shifting, other factors started to show a strong influence on the results of the current demand response mechanisms. A numerical example with an upper level of 20% DR is proposed and the results show that DR caused an increase of 11.29% in peak demand comparing with a maximum of only 7.5% in saving. With the implementation of smart metering infrastructure, demand response will play a major role in the future smart grid. It is glad to see that a group of researchers is focusing on social contexts such as policy, psychology, individual behavior, education, and income level. We believe that social level analysis will provide a better understanding of residential electricity usage that can really benefit the design of future residential demand response mechanisms. Acknowledgements This work was supported by the China Postdoctoral Science Foundation Funded Project (2016M591174). References [1] U.S. Department of Energy [Internet]. [cited 2015 May 18] Available: 〈http:// energy.gov/oe/services/technology-development/smart-grid〉. [2] Darby SJ, McKenna E. Social implications of residential demand response in cool
418
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
[27] Yan X, Wright D, Kumar S, Lee G, Ozturk Y. Real-time residential time-of-use pricing: a closed-loop consumers feedback approach. In: Proceedings of the IEEE GreenTech; 2015. [28] Yan X, Wright D, Kumar S, Lee G, Ozturk Y. Enabling consumer behavior modification through real timeenergy pricing. In: Proceedings of the first IEEE workshop on pervasive energy services; 2015. [29] Giorgio AD, Pimpinella L. An event driven smart home controller enabling consumer economic saving and automated demand side management. Appl Energy 2012;96:92–103. [30] Ericson T. Households' self-selection of dynamic electricity tariffs. Appl Energy 2011;88:2541–7. [31] Herter K, Wayland S. Residential response to critical-peak pricing of electricity: california evidence. Energy 2010;35(4):1561–7. [32] Aubin CD, Fougere E, Ivaldi M. Real-time pricing for residential customers: econometric analysis of an experiment. J Appl Econ 1995;10(4):171–91. [33] Country Energy. The country energy home energy efficiency trial; 2005. [34] Faruqui A, Sergici S. Household response to dynamic pricing of electricity: a survey of 15 experiments. J Regul Econ 2010;38:193–225. [35] Renner S, Albu M, Elburg H, Heinemann C, Lazicki A, Penttinen L, Puente F, Saele H. European smart metering landscape report 2011. [36] Widergren S, Marinovici C, Berliner T. et al. Real-time pricing demand response in operations. In: Proceedings of the IEEE power and energy society general meeting; 2012. p. 1–5. [37] Widergren S, Chan SC. Demand response optimization for smart home scheduling under real-time pricing. IEEE Trans Smart Grid 2012;3(4):1812–21. [38] Avci M, Erkoc M, Asfour SS. Residential HVAC load control strategy in real-time electricity pricing environment. In: Proceedings of the IEEE Energytech; 2012. p. 1–6. [39] Yousefi S, Moghaddam MP, Majd VJ. Optimal real time pricing in and agent-based retail market suing a comprehensive demand response model. Energy 2011;36:57165727. [40] Yoon JH, Bladick R, Novoselac A. Demand response for residential buildings based on dynamic price of electricity. Energy Build 2014;80:531–41. [41] Moghaddam MP, Abdollahi A, Rashidinejad M. Flexible demand response programs modeling in competitive electricity markets. Appl Energy 2011;88:3257–69. [42] Hammerstrom DJ, Ambrosio R, Carlon T, Desteese JG, Kajfasz R, Pratt RG. Pacific Northwest GridWise Testbed demonstration projects. Part I. Olympic Peninsula Project 2007. [43] Allcott H. Rethinking real-time electricity pricing. Resour Energy Econ Resour Energy Econ 2011;33:820–42. [44] Star A, Isaacson M, Haeg D, Kotewa L. The dynamic pricing mousetrap: why isn’t the world beating down our door? n: Proceedings of the ACEEE summer study; 2010. p. 2257–68.
[45] Rastegar M, Fotuhifiruzabad M, Choi J. Investigating the impacts of different pricebased demand response programs on home load management. J Electr Eng Technol 2014;9(3):1125–31. [46] Reiss PC, White MW. Household electricity demand, revisited. Stanford University Publication; 2002. [47] Fuller JC, Schneider KP, Chassin D. Analysis of residential demand response and double-auction markets. In: Proceedings of the IEEE power and energy society general meeting; 2011. p. 1–7. [48] EPA. Summary of research findings from the programmable thermostat market. 2005. [Online]. Available: 〈http://www.energystar.gov/ia/partners/prod_ development/revisions/dowasnloads/thermostats/Summary.pdf〉. [49] DRRC, Understanding customer behavior to improve demand response delivery in California. PIER demand response research centre research opportunity notice DRRC RON -3; 2007. [50] Soares A, Gomes A, Antunes GH. Categorization of residential electricity consumption as a basis for the assessment of the impacts of demand response actions. Renew Sustain Energy Rev 2014;30:490–503. [51] Simons SI, Rowlands IH. TOU rates and vulnerable households: electricity consumption behavior in a Canadian case study. Dep Environ Resour Stud 2007. [52] Socolow RH. Saving energy in the home: princeton's experiments at twin rivers. Cambridge, MA, USA: Ballinger Publishing Company; 1978. [53] Giorgio AD, Liberati F. Near real time load shifting control for residential electricity prosumers under designed and market indexed pricing models. Appl Energy 2014;128:119–32. [54] Corradi O, Ochsenfeld H, Madsen H, Pinson P. Controlling electricity consumption by forecasting its response to varying prices. IEEE Trans Power Syst 2013;28(1):421–9. [55] Faria F, Vale Z. Demand response in electrical energy supply: an optimal real time pricing approach. Energy 2011;36:5374–84. [56] Venkatesan N, Solanki J, Solanki SK. Residential demand response model and impact on voltage profile and losses of an electric distribution network. Appl Energy 2012;96:84–91. [57] Siano P. Demand response and smart grids-a survey. Renew Sustain Energy Rev 2014;30:461–78. [58] The NOBEL market. [Online] [cited at Sep. 2016]. Available at: 〈http://www.ictnobel.eu〉. [59] Piclo P2P energy trading, [Online] [cited at Sep. 2016]. Available at: 〈https:// www.openutility.com/piclo/〉. [60] Rahimi F, Ipakchi A, Fletcher F. The changing electrical landscape end-to-end power system operation under the transactive energy paradigm. IEEE Power Energy Mag 2016;14(3):52–62. [61] Hug G, Kar S, Wu C. Consensus + innovations approach for distributed multiagent coordination in a microgrid. IEEE Trans Smart Grid 2015;6(4):1893–903.
419
Contents lists available at ScienceDirect
Renewable and Sustainable Energy Reviews journal homepage: www.elsevier.com/locate/rser
A review on price-driven residential demand response a,⁎
b
c
Xing Yan , Yusuf Ozturk , Zechun Hu , Yonghua Song
T
c,d
a
Energy Internet Research Institute, Tsinghua University, Beijing 100084, China Department of Electrical and Computer Engineering, San Diego State University, San Diego 92182, CA, USA Department of Electrical Engineering, Tsinghua University, Beijing 100084, China d Department of Electrical and Computer Engineering, University of Macau, Macau 519000, China b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Critical peak pricing Demand response Real-time pricing Time-of-use pricing
Smart grid enables the two-way communication between the suppliers and consumers. Price-driven demand response (PDDR) is one of the important demand response categories that uses price of the energy as control signals to affect consumers’ electricity consumption. The current PDDR programs include critical peak pricing (CPP), time-of-use (TOU) pricing, and real-time pricing. In this paper, we provide a review of the PDDR studies. Detailed evaluations on advantages and disadvantages of each PDDR are provided. Concerns and future research challenges on PDDR are also addressed. It is believed that with the installation of smart meter infrastructures at residential households, price signal can be an efficient market tool for peak demand shaving, risk and reliability management, carbon emission reduction, and energy cost reduction.
1. Introduction
response, it is still increasingly seen as a way of meeting emerging needs for better system management and concerns about energy security and environmental impact [2]. Demand response as a proactive measure can be implemented in both manual and automated ways. The manual demand response usually refers controlling use of certain appliances in different time periods of the same day. Semi-automated demand response involves a pre-programmed demand response strategy that is initiated by a person through a centralized control system. Fully automated demand response does not involve any human intervention but is initiated at a home, building, or facility through receipt of an external communications signal [3]. Although demand response is a very cost-effective way for peak demand shaving, risk and reliability management, carbon emission reduction, and energy cost reduction, industrial and residential sectors have different aspects on this topic. For industrial sector, the primary target is maximizing the profit. As electricity consumption is part of the production cost (in some industries such as chemical production companies electricity consumption is a huge part of the production cost) and demand response can reduce the electricity cost, industrial sector is able to adopt demand response very quickly. The residential sector, however, is not able to adopt demand response as quickly as the industrial sector because the primary focus of the residential sector is on increasing the comfort level. Different opinions of demand response are also rising important issues in applying such technology at the
Electricity is a very unique commodity which cannot be economically stored in large quantity. Nowadays, nearby electricity markets are interconnected with each other to improve system reliability, reduce energy cost, increase penetration of renewable resources, and reduce carbon emission. With a much larger interconnected grid, surplus electricity energy can be stored into different forms such as hydro power by pumping the water back to the dam for later use. According to the U.S. Department of Energy, smart grid generally refers to a class of technology used to bring utility electricity delivery systems into the 21st century, using computer-based remote control and automation [1]. These systems are made possible by two-way communication technology and computer processing that has been used for decades in other industries [1]. Smart grid enables the electricity energy exchange between electricity markets. It also enables integration of renewable energy resources to the energy exchange grid. Researchers are continuously working on new mechanism that can better serve the demand by either increasing supply or keeping demand bounded to the supply. While the proactive management of demand plays an important role, at times of high demand tapping into renewable energy resources or tapping into backup resources are critical components of smart grid solutions. All of these actions can also be referred as demand response. According to the current understanding of the impact of demand
⁎
Corresponding author. E-mail addresses: [email protected] (X. Yan), [email protected] (Y. Ozturk), [email protected] (Z. Hu), [email protected] (Y. Song). https://doi.org/10.1016/j.rser.2018.08.003 Received 13 April 2017; Received in revised form 29 June 2018; Accepted 3 August 2018 1364-0321/ © 2018 Elsevier Ltd. All rights reserved.
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
large quantities. Demand response can provide ancillary services to enhance the voltage stability in power systems operating under increased uncertainty, and replace some or even all of the spinning reserves typically supplied by conventional generators [7]. This will avoid using rolling blackouts when there is not enough energy supply and in turn increase the electricity grid reliability and reduce congestion. Consumers can also benefit from demand response. Under the current regulated energy transmission and distribution systems, electricity prices are generally averaged over the entire year and cost-of-service pricing is the norm. On the other hand, the real-time electricity price varies from time to time. Utility companies have to purchase highpriced wholesale energy whenever they cannot meet the demand. An extreme example occurred on November 28, 2005, eastern Denmark. The wholesale price reached DKK 13,469/MWh (equals to $2189.17/ MWh USD), which is 60 times more than the normal price level [8]. These high wholesale energy prices are eventually included in the calculation of the averaged energy price and are paid by the consumers. Demand response programs can lower the energy consumption during peak hours. This will reduce the wholesale energy prices and in turn reduce the energy price paid by the consumers. Renewable resources, such as solar, wind, biomass, and hydro, are continuously growing and playing a very important role in supplying electricity. According to the U.S. Energy Administration Renewable Electricity Standard (RES), by 2020, every state in the United States must generate one third of their electricity from renewable resources. German Advisory Council on the Environment states that Germany is aiming to transform its electricity supply to 100% renewable energy by 2050. Demand response programs can increase the penetration of renewable resources and benefit the environment in reducing carbon emission. It can contribute to the better integration of renewable energy resources such as wind power, solar, small hydro, biomass and combined heat and power (CHP) [9,10]. Excess energy from the renewable resources can be either stored into different form for later use or supply to the grid to reduce the electricity generated by the ordinary thermal generation plants. Surplus energy from the renewable resources can be stored in a different form of energy for later use in order to minimize the overall operating and environmental costs [11]. With renewable energy accessible only at certain times, the research on energy storage is also gaining momentum. We believe demand response programs will benefit from the developments in storage technologies as well as drive the developments in the storage technologies. Currently, demand response can be divided into main categories which are further subdivided into many forms. A complete list of demand response varieties is shown in Fig. 1. In this paper we will focused on residential PDDR programs. A detailed presentation of residential PDDR programs is given in Section 3. In the incentive or event-based demand programs, direct load control and emergency response programs are voluntary programs. Consumers will not get penalized if they did not adjust their energy consumption according to the suggestion by the local utility companies or the independent system/market operator. In the direct load program, utility companies or system operators can remotely control customers’ specific appliances (large energy consumption devices) in order to respond to the demand and reliability issues. Incentives are available for involved customers. Emergency response program provides incentives in exchange for voluntary load reduction during special events. Capacity market programs and interruptible/curtailable service are mandatory programs and enrolled consumers will get penalized based on the agreement if they did not adjust their energy consumption accordingly when needed. In the capacity market program, load reduction is pre-specified and mandatory as directed by the utility companies and system operators. Interruptible/curtailable program is very similar to the emergency response program except it is mandatory. Finally, demand bidding/buyback and ancillary service market programs are market mechanisms to balance energy supply and demand based on the
residential sector. According to a study proposed by Bartusch et al. [4] with 500 interviewed consumers under Swedish power system, only 50 are participated in the proposed study. People neither want their lives to be governed by tariff rates nor their demand to be remotely controlled, and they do most certainly not want to waste their time keep tracking on their electricity usage [4]. Although everyone wants to save money on their electricity bills, sacrificing the comfort level in exchange for financial incentives is not accepted by the majority of the residential sector, especially when the savings are very small. Another important reason is that most consumers are only charged by a fixed electricity rate based on the average electricity cost even though the real time electricity prices vary from time to time. In some cases, electricity costs may vary hourly or seasonally according to the seasonal and daily variations [5]. The U.S. Energy Policy Act of 2005 mentions that each electric power company should provide customers with timebased rates [5]. Currently, most utility companies offer consumers tiered residential electricity prices based on the climate zone, seasonal information, and time of use. When a consumer's total electricity usage exceeds a certain predetermined threshold, the consumer will be subjected to different electricity rates based on the tiered price structure. We believe that research on real time energy pricing and consumer behavior modification through pricing energy will gain acceleration as utility companies roll out pricing systems that charges consumers based on time of use and amount of energy consumed. In this paper, we reviewed the PDDR studies. Advantages and disadvantages of each PDDR program are evaluated in detail with experimental results. Based on the results, a discussion section is provided including our concerns and opinions on the current PDDR programs. The rest of the paper is organized as follows: Section 2 provides a general introduction on demand response. Detailed evaluation on recent PDDR programs are explained in Section 3. Section 4 discusses the issues with recent PDDR and conclusions are provided in Section 5. 2. Demand response According to the Federal Energy Regulatory Commission, demand response is defined as changes in electric usage by demand-side resources from their normal consumption patterns in response to changes in the price electricity over time, or to incentive payments designed to induce lower electricity usage at times of high wholesale market prices or when system reliability is jeopardized [6]. The studies on demand response started on early 1980s under the demand-side management programs. It is believed that utility companies, transmission and distribution system operators, and end-use consumers can all benefit from demand response. The benefits of demand response for utility companies presents themselves as reduced capital cost, operation cost, and reduced carbon emission. Compared to the huge capital cost and long construction time of period building new generation plants to match electricity consumption at peak hours, demand response is an economical, and environmental friendly program that can be used to balance the supply and demand within the jurisdiction of electricity grid. Demand response can also help the utility companies operate their power generation plants at optimized speeds. This in turn will result in reduced fuel consumption, increased productivity and increased profit margins. Carbon emission can also be reduced as the fuel consumption efficiency is increased. Moreover, utility companies can benefit from the demand response by avoid use of diesel powered backup/emergency power plants and in turn reduce carbon emission. Because of such advantages, demand response is among the top priorities of energy utility companies. For transmission and distribution system operators, demand response is also an essential asset. The transmission and distribution system operators are in charge of balancing the demand and supply within the jurisdiction of electricity grid at all time. Electricity is a special commodity which cannot, currently, be economically stored in 412
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Time-of-use Pricing Critical Peak Pricing Price-Driven Real-time Pricing Peak Time Pricing Demand Response Direct Load Control Emergency Demand Response Programs Capacity Market Programs Incentive or Event-Driven Interruptible/Curtailable Service Demand Bidding/Buyback Programs Ancillary Service Market Programs Fig. 1. Demand response categories.
Smart metering (advanced metering) is a measuring system that records customer's consumption (and possibly other parameters) on a timely basis (hourly or more frequently) and provides hourly or even minutely transmission of measurements over a communication network to a data consolidation point [5]. With the improved infrastructure in residential sectors with increased penetration of smart meters, real-time electrical energy consumption monitoring within the smart grid became widely available. Fig. 3 shows a detailed residential 24-h electricity consumption on an hourly basis. The U.S. Energy Policy Act of 2005 announced that it is the policy of the United States to encourage “timebased pricing and other forms of demand response” and encourage states to coordinate, on a regional basis [13]. State energy policies aims to provide reliable and affordable demand response services to the public [13]. New techniques are able to apply to the residential sectors and enhance the PDDR. The three different residential PDDR programs are evaluated in the following sections.
determination of market clearing price. Demand bidding/buyback program gives large energy consumption parties the ability to negotiate the price for the amount of load reduction. In ancillary service market program, negotiated amount of load reduction with the corresponding prices are used as the reserve energy for the electric grid. It is also referred as spinning reserve energy. Incentives based on the spot market prices and amount of gain from load reduction are paid to the participating customers when called by the system operators. 3. Residential price-driven demand response Price-driven demand response uses electricity prices as control signals to motivate consumers to change their energy consumption. The goals of applying these electricity price signals are to reduce overall energy consumption and shift some of the peak load into off-peak hours at the same time. PDDR contains time-of-use (TOU) pricing program, critical peak pricing (CPP) program, and real-time pricing program. The price patterns representing all 3 PDDR programs are shown in Fig. 2. PDDR is first practiced at the industrial sectors where energy consumed is part of the production cost. Higher energy costs results higher production costs. Therefore, company's large amounts of electrical energy such as chemical production companies are willing to participate in the PDDR programs in order to benefit from the cheaper electricity prices during off-peak hours. Although there were some early stage researches focusing on residential PDDR in the 1980s, due to lack of monitoring on the daily or even hourly electricity consumption, only limited techniques were able to applied to the PDDR in the residential sectors.
3.1. TOU pricing program The TOU pricing program shown in Fig. 2 uses a static pricing scheme based on predefined price values. The price pattern usually maintains the same prices during each season. New price portfolios are proposed yearly to cover the operation cost and long-term investment of the utility companies. The proposed price portfolios are approved by the state/provincial price regulation party whose chief objective was to protect the consumers from the inevitable consequences of a monopoly
Fig. 2. PDDR programs: a). TOU Pricing, b). CPP, and c). RTP [12]. 413
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
PDDR program, an approximate 10% peak shifting can be achieved [8,15–24,29] and consumers did save money on their electricity bills. However, according to Fig. 4, because participated consumers scheduled their large energy consumption appliances right after the price drop, new demand peak is created at the beginning of off-peak hours. These new demand peaks are usually worse than the ones under the static pricing. On the other hand, an overall energy consumption deduction is observed in all involved case studies which implies that TOU pricing did raise certain awareness on energy consumption. A similar observation is concluded in [14] that while households reduce peak demand significantly after the introduction of TOU tariffs and associated information, there is little incremental response to increasing differentials between peak and off-peak prices. 3.2. CPP program
Fig. 3. Daily energy consumption pattern for an individual household [12].
The CPP program is an event driven program that uses high electricity prices as control signals to affect electricity consumption during peak hours. CPP may be imposed if the system is severely constrained as in cold winter periods or warm summer periods for a limited number of hours [30]. Under these instances, consumers can receive incentives by either reducing peak consumption or shifting energy consumption to off-peak periods to protect system reliability. Since system under severe constrains does not happen on the daily basis, CPP is not a daily demand response program. Moreover, CPP values are higher than TOU pricing values. The advantages of CPP program includes: (1) easy to follow, (2) effective on shifting peak energy consumption, and (3) visualized incentives. Because CPP program is an event-driven tariff whose primary focus is ensuring system reliability, households are more willingly to participate in such event. A survey of 483 households in California by Herter in [31] concluded that statically a significant average participant response in each hour. Day-ahead CPP values are delivered to consumers through social media such as newspaper, text messages, and websites. This makes CPP program very easy to understand and follow by average consumers. The disadvantages of CPP program includes: (1) event driven and (2) not effective on reducing energy cost and carbon emission. Due to the fundamental event-driven principle of CPP program whose primary focus is ensuring system reliability, such program usually does not apply on a daily basis as system is not facing severe constrains on a daily basis. Therefore, CPP is not an effective program in improving energy consumption efficiency, reducing energy cost, and reducing carbon emission. Results of some previous case studies on CPP are shown in Table 2. CPP programs are usually applied in combination with other PDDR programs to optimize the results of shifting peak demand to off-peak periods.
industry. The successful results from the industrial sector containing a survey of 43 TOU pricing programs show that the cost savings can vary from − 72.0% to + 82.6%, depending on specific utility programs and switching strategies involved [14]. The advantages of TOU pricing program include: (1) easy to follow and (2) stable daily participating ratio. Because the TOU pricing portfolios maintain the same prices during the same season, it is very easy for consumers to understand, follow, and plan their daily electricity consumption portfolios. As a result, the daily percentage of people participating in the TOU pricing program is quite stable compared with other two PDDR programs. On the other hand, the advantage of the robust price portfolios of the TOU pricing program is its own fundamental disadvantage at the same time. Fig. 4 shows a daily electricity consumption portfolio comparison between static (fixed) and TOU pricing programs. It can be seen from Fig. 4 that there is a demand peak occurred right after the price dropped from peak hours to off-peak hours. Comparing with the demand curve under the static pricing program, TOU pricing program did reduce the overall electricity demand during peak hours, but created a new and much bigger demand peak during the off-peak hours. Although Fig. 4 only represents the electricity consumption of a single household, aggregated residential electricity consumption has a great chance of creating a new and much bigger demand peak from the residential sector, especially when increased plug-in electric vehicles are involved. On top of that, adjusting the time interval of peak and offpeak hours will not reduce the demand peak but shift it into different hours. Summarized results from previous studies are shown in Table 1. The best saving of 21.07% in financial incentives is obtain in the case study proposed by Giorgio et al. [29] on a low number of loads. However, the question is, will you sacrifice the promotional level of comfort in exchange with the financial incentives. Meanwhile, Gottwalt et al. [13] explored that an individual household can expect rather low benefits of an investment in smart appliances when TOU prices hardly exceed the static prices. According to the previous studies on TOU
3.3. RTP program Under the RTP program, electricity price varies continuously in response to the wholesale market prices. The dynamic electricity prices
Fig. 4. Daily consumption under different price programs [15]. 414
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Table 1 Case study results of TOU pricing. Studies
Place and year
Peak shifting
Comments
Bartusch et al. [4] Cosmo VD [14]
Sweden 2005–2008 Ireland 2007
0.1–2.5% N/A
Torriti [16]
Northern Italy 2010
N/A
Woo et al. [17]
Canada (British Columbia) 2007–2008
2.6–9.2%
Faruqui et at [18] CRA [19] Schlomann [20]
3–6% 13% 3–10%
Quantec [21]
U.S. 2010 U.S. (California) 2003–2004 Germany (Freiburg) 1970 and 1980s U.S. (Portland) 2004
Peak shifting is very limited during major consumption season (winter) 5000 Households involved. Peak demand reduced significantly, but little peak shifting observed. Consumption increased by 13.69% consumers’ electricity bills decreases by 2.21% 75.6% of substations experienced an increase in electricity demand during peak periods 2.6% at 2:1 peak-to-off-peak price ratio 9.2% at 12:1 peak-to-off-peak price ratio Participating customers are enthusiastic volunteers Summer afternoon peak
Hammerstrom [22] CER [23] Gyamfi [24] Ozturk [25]
U.S. (Washington) 2007 Ireland 2011 New Zealand (Christchurch) 2013, San Diego
Yan et al. [26–28]
2015, San Diego
• • •
1970 and 1980s
7–15% in winter 2% in summer 5–20% 8.8% 7–10% 2006 Demand forecasting and peak pricing Peak shifting
Owners set automatic response based on comfort and price preference
A residential energy management system has been developed. The home controller system controls appliances with user comfort in mind. A real time energy pricing system is designed and fully integrated to the smart grid via Green Button API. An android application deployed in cooperation with the local utility company. The energy residential energy management application “Energyelastics” is available on google app store.
applications with smart meters.
are available to the public an hour or sometime one day ahead. RTP programs can help utility companies better distribute the electricity price reflecting the demand-supply elastics. These dynamic price signals can motivate consumers to adjust their electricity consumption accordingly in order to achieve financial benefits. RTP signals combined with automation in end-use systems have the potential to deliver even more benefits to operators and consumers [36,37]. Some previous case studies on RTP program are shown in Table 3. Dynamic pricing remains as a new idea for residential customers [44]. Without fully implemented smart meter infrastructures, there is not enough hardware support to start with the RTP demand response program at the first place. Application of residential demand response (DR) programs is currently realized up to a limited extent due to customers’ difficulty in manually responding to the time-differentiated prices [45]. Even though at some places, the smart meter infrastructures are available in the residential sectors, the test of consumers’ acceptance of dynamic pricing remains very limited. Getting customers to voluntarily embrace dynamic pricing will take an ongoing, long term series of engagements reminders, pokes, and prods [44]. EnergyElastics [26–28] application is one of the first consumer advisory application that is fully integrated to the smart meter infrastructure enabling utility company to push real time prices to the user and monitor energy consumption in response to the prices pushed. It is integrated to the San Diego Gas and Electric smart meter infrastructure through Green Button API which is currently standardized for interfacing third party
3.4. Numerical example To better understand the effects of different pricing methods, a numerical example of a single household with a typical 24-h load profile from [27] is shown in Fig. 5. It is assumed that an upper level of 20% DR can be applied from hour 17–20 and shifted these demand starting from hour 21–23. The proposed static, TOU, CPP, and RT pricing demand response programs with experimental values are shown in Fig. 6. Numerical evaluations of applying different pricing methods are available in Table 4. According to Fig. 5, we can obtain that if a maximum of 20% DR is achieved, a new peak load is created at hour 21 as 365.7 W instead of 328.6 W at hour 19. This is an increase of 11.29% in peak load with a lower electricity cost. The new and bigger peak demand will create a new potential safety issue to the utility side as the original purpose of introducing the PDDR is to reduce the peak load rather than creating a new and a bigger one. Therefore, according to the results in Fig. 5, by applying PDDR, the utility side could face possible new critical peak load situation. According to Table 4, it can be obtained that with an upper level of 20% DR applied with different DR programs, some level of savings can be achieved by shifting some or all of the flexible demand into later time with lower electricity prices. The shifted demand may include taking a shower, plugging electric vehicles, doing laundries, and
Table 2 Case study results of CPP. Studies
Place and year
Peak shifting
Comments
Ericson [30]
Norway 2003
N/A
Herter et al. [31]
California 2004
1.9–10%
Aubin [32]
France 1996
15–45%
The study estimated that households with energy management system and electricity heating may favor the CPP program over flat rate program by 10%. 483 households involved. Results indicated that larger users respond more in both absolute and percentage terms. Households did not respond more to the higher CPP rate Experiment was conducted using newspaper to communicate with consumers about the day-ahead electricity prices. Although the experiment was successful, less than 20% of the customers have chosen this option.
Country Energy [33] Faruqui et al. [34] Renner et al. [35]
Australia 2005 North America 2010 Sweden 2011
2.4% 13–20% or 27–44% Approximate 50%
Lower value is without direct load control (DLC) while higher value is achieved with DLC allied. Customers given energy saving tips and day-ahead warnings of critical periods. Involved households are willing to support local utility.
415
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Table 3 Case study results of RTP. Studies
Place and year
Peak shifting
Comments
Avci et al. [38]
PJM 2011
4.3–6.7%
Yousefi et al. [39]
U.S. (New England) 2010
Approximate 5%
Yoon et al. [40]
U.S. (Austin, TX) 2012
12.8–24.7%
The proposed study only focused on the operation of heating, ventilation, and air conditioning (HVAC) system. The comfort level is also considered in the simulation. Simulated results is modeled based on the learning capability of an intelligent retail energy provider agent using day-ahead RTP program. Only focus on HVAC system. 12.8% sacrifice on temperature in exchange for 10.8% saving under RTP program.
Moghaddam et al. [41] Hammerstrom et al. [42]
Iran 2007 U.S. (Washington State) 2007 U.S. (Chicago) 2009
4.2–4.9% 5–20%
Owners set automatic demand response based on predefined comfort and price level.
5–14%
Day-ahead RTP.
Allcott [43]
programs in Table 5.
4. Discussions 4.1. Concerns of PDDR programs Although most studies presented above show that residential customers did respond to time-varying electricity prices and some level of peak demand shifting are achieved, a detailed analysis by Reiss et al. [46] indicates that significant proportion of households do not response to price. The level of response from a household is proportional related to the household's income and energy consumption. When household's income increases, the response decreases. Response of the lowest income households (less than $18,000 USD) were almost 50% higher than that of the highest income households (more than $60,000 USD). 44% of the involved households did not show any price responsiveness. It is worth to note that households with major appliances such as heating and air conditioners responded the most [38,40,46]. According to the previous PDDR programs [4–44], consumers’ education and awareness on their daily energy consumption has significant impact on the success of the program and customers’ feedback. Fewer than 50% of the US households have programmable thermostats, and even worse, the US Environmental Protection agency estimates that 30% or more of the US households with programmable thermostats are not using their thermostat's programing feature, instead they put them in hold mode and operate them manually [38,47,48]. Gyamfi et al. [24] explores that voluntary participation and behavior change in demand response are the critical issue for the developers of peak demand management programs. Another fact that effects the response to the price signals is that the residential consumers will not always be available to manage energy resources and decide, based on price signals and preference/needs, the best response actions to implement or the best usage of the electricity provide locally [49]. Maybe when commercialized storage units are available for residential consumers to store electricity energy from renewable resources, more people are willing to participate in the PDDR programs [50]. The current researches on PDDR programs also raised number of concerns. Torriti [16] concluded that issues with evening peaks were not resolved using TOU program and led to increment in electricity demand for substations at peak time. He also pointed out that on the positive side, the lesson from small combined heat and power systems suggest that two-way direct pricing approaches bring about better integration with the electricity supply. On the negative side, the attractiveness of the TOU pricing from a policy perspective is not matched by economic rigor [22]. Ericson [30] also pointed out his concern that time-differentiate tariffs may attract consumers who benefit without responding to the price, simply because they have a favorable consumption pattern. Moreover, higher prices discriminate against lower income households [51]. The lack of quantitative understanding of consumers’ behavior and end-use activity adaptive capacities are a significant barrier to design and deploy effective demand response
Fig. 5. 24-h electricity demand of a single household.
Fig. 6. Different pricing methods within a 24-h period. Table 4 Numerical example evaluation of different pricing methods. Type
Electricity cost ($ USD)
Electricity cost with DR ($ USD)
Percentage of savings (%)
Static TOU CPP RTP
2.6476 2.6361 2.9370 2.7276
2.6476 2.4522 2.8791 2.6145
0.00 7.50 2.01 4.33
turning on/off air conditioning. Depends on the level of comfort the users are willing to sacrifice such as shifting one and up to all of the above demand at a later time, a maximum percentage of savings of 7.5%, 2.01%, and 4.33% can be achieved accordingly using TOU, CPP, or RTP programs. It is worth to note that DR caused an increase of 11.29% in peak demand comparing with a maximum of only 7.5% in saving. By evaluating the numerical examples and reviewing the literatures, we summarized the characteristics of the three different PDDR 416
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Table 5 Key information of different PDDR programs. Type TOU
CPP
RTP
Type of DR can be applied ● ● ● ● ● ● ● ●
Manual operated DR Semi-automated DR Fully-automated DR Manual operated DR Semi-automated DR Fully-automated DR Semi-automated DR (partially) Fully-automated DR
Advantages
Disadvantages
● Easy to follow ● Stable daily participating ratio
● May shift the old peak load into different time
● ● ● ●
● Event driven ● Not effective in reducing energy cost and carbon emission
Easy to follow Effective on shifting peak energy consumption Visualized incentives Reflecting the demand and supply relationship
● Hard to participate at the moment
For instance, Darby et al. [2] points out the significance of thermal loads, supply mix, demand-side infrastructure, market regulation, and the framing of risks and opportunities associated with demand response. Siano [57] concluded further researchers are required to discuss regulatory and policy recommendations in order to ensure technical functioning and non-discriminatory physical access to all parties and precise rules, defining roles and responsibilities of all players and assuring fair sharing of costs and benefits among all stakeholders.
programs [49]. Socolow and his team showed that houses of similar sizes, occupied by demographically similar families, with a similar set of appliances, and under the same geographical condition, varied in energy consumption by as much as 200% [52]. Other concerns are people neither want their lives to be governed by tariff rates nor their demand to be remotely controlled, and they do most certainly not want to waste their time keep track on their electricity usage [4]. Although the above studies did demonstrate increased level of awareness on residential electricity consumption [3–44], this does not necessarily lead to peak demand shifting. Residential customers will certainly not postpone preparing their dinner until off-peak periods. Meanwhile, some other studies are trying to simulate the economic model of price/ incentive responsive loads which is based on the concept of flexible price elasticity of demand and customers’ benefit function to discover each of the demand response program's elasticity based on the electricity price before and after implementing demand response programs [17,39,41,53–55]. These researchers believe that the market regulator can use these flexible elastics to simulate the behavior of customers for different electricity prices, incentives, penalties, and participation level of customers. However, Bartusch et al. [4] has a different opinion on using price elastics as they believed estimating the price elasticity of demand would be questionable or even frivolous. On the other hand, positive effort such as Soares et al. [50] aims at characterizing and classifying in a detailed way the potentially controllable demand in the residential sector and assessing the impacts of implementing distinct automated demand response actions over some of the previously identified controllable end-use load is on the positive track. Moreover, demand response and dynamic pricing programs are expected to play increasing roles in the modern smart grid environment [56]. Positive demand response results are shown in major appliances such as HVAC which can respond to price signals while maintaining the comfort level within predefined tolerance [38,40,46]. Researchers are also aware of other factors that may be related to effecting peoples’ level of involvement in demand response programs. Gyamfi et al. [12] proposed the use of a hybrid engineering approach using social psychology and economical behavior models to overcome the challenges (price unresponsiveness, equity issues, and high cost of metering infrastructure) and realize the benefits of supply security and cost management. Allcott [43] argues that residential RTP should perhaps be thought of as a peak energy conservation program instead of a mechanism for peak demand shifting. Gottwalt et al. [13] pointed out that incentive mechanisms have to be designed to transfer parts of the benefits from more flexible and controllable residential loads from utilities to households. He also mentioned that it might be beneficial to evaluate the relationships between economic benefits and flexible load for residential homes [13]. Bartusch et el. [4] did conclude that in recent years, the distribution system operators have engaged in a range of activities for the purpose of promoting awareness and encouraging energy efficiency among their customers. In addition, the observed reduction in the electricity consumption may well be a sign that the efforts on building a demand response mechanism have paid off. Researchers are now starting to consider demand response as a complex system rather than a simple price-driven or event-driven mechanism.
4.2. P2P trading With the huge successes of the peer-to-peer (P2P) business model in the IT sector, the electricity system is considering the possibility of adopting such business model at the distribution level [58–60]. As an example, the primary goal in [61] is to enable the electricity trading between the citizens of a smart city in a continuous double auction market. With the increased number of aggregators entering the electricity market, it is possible for the P2P trading method to be applied in the PDDR program. An aggregator is a retail agent who represents a combined group of customers trading electricity at the wholesale market. The aggregator is able to react to PDDR signals by enabling a P2P trading among its customers. When a DR is needed, the spot market price reaches a peak value. The aggregator can use this information and design a protocol whenever a spot market release a PDDR price signal. Once a spot market price reached a threshold value, the PDDR P2P trading process will be initiated. Such trading process can help the aggregator purchase less electricity at the spot market during the DR period and reduce the spot price volatility. A completed trading cycle for an order of PDDR trading can be seen in Fig. 7. A timeslot of 15-min is used as the electricity consumption recording period. The P2P trading cycle contains 9 steps begins at minute 0 and finishes at minute 30. Minute 0 to minute 15 is the bidding cycle and minute 15 to minute 30 is the implementation cycle. Step 1: The aggregator initiates the PDDR P2P trading process. All participated consumers collect the DR information from the aggregator and the forecasted load information from their home smart infrastructure devices. Step 2: All participants place their first round of bidding orders in the aggregator operated the market. Step 3: The aggregator will match the trading orders. Step 4: The remaining quantity of electricity during the matching process and the unmatched orders will remain in the order book and enter the matching process for the next round with the new bidding orders. Step 2–4 will be repeated for a predetermined number of rounds. Step 5: At this moment, the bidding cycle is completed. All unmatched orders will be reset to 0 while all matched orders are recorded by the aggregator and related information such as matched electricity quantity, average trading price, total buying/selling quantity, and specific matching order with only quantity information will be released to all participants. 417
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
Fig. 7. PDDR P2P trading mechanism.
Step 6: The aggregator will start to implement the matched orders taken place from the previous 15 min. Step 7: During the implementation process, due to the deviation of electricity, some of the matched orders cannot be fully delivered. The aggregator will settle the deviation quantity with other aggregators at the spot market. Step 8: Unmatched bidding orders will be charged with the current spot market price. Step 9: At minute 30, the aggregator will complete the implementation process of all matched orders.
temperate climates. Energy Policy 2012;49:759–69. [3] Kiliccote S, Piette MA. (Lawrence Berkeley National Laboratory). Automation development and testing of the capacity bidding program for San Diego gas and electric company. In: Proceedings of the PIER energy systems integration program. California energy commission. CEC-500-2008-059. [4] Bartusch C, Wallin F, Odlare M, Vassileva I, Wester L. Introducing a demand-based electricity distribution tariff in the residential sector: demand response and customer perception. Energy Policy 2011;19:5008–25. [5] Energy Policy Act of 2005. Public Law 109-58. August 8, 2005. [6] Federal Energy Regulatory Commission [Internet]. [cited 2015 May 18] Available: 〈http://www.ferc.gov/industries/electric/indus-act/demand-response/dem-resadv-metering.asp〉. [7] Wang D, Parkinson S, Miao W, et al. Online voltage security assessment considering comfort-constrained demand response control of distributed heat pump systems. Appl Energy 2012;96(3):104–14. [8] Aghaei J, Alizadeh MI. Demand response in smart electricity grids equipped with renewable energy sources: a review. Renew Sustain Energy Rev 2013;18:64–72. [9] Wang X, Palazoglu A, El-Farra NH. Operational optimization and demand response of hybrid renewable energy systems. Appl Energy 2015;143:324–35. [10] Yang H, Xiong T, Qiu J, et al. Optimal operation of DES/CCHP based regional multienergy prosumer with demand response. Appl Energy 2016;167:353–65. [11] Broeer T, Fuller J, Tuffner F, et al. Modeling framework and validation of a smart grid and demand response system for wind power integration. Appl Energy 2014;113(6):199–207. [12] Gyamfi S, Krumdieck S, Urmee T. Residential peak electricity demand responsehighlights of some behavioral issues. Renew Sustain Energy Rev 2013;25:71–7. [13] Gottwalt S, Ketter W, Block C, Collins J, Weinhardt C. Demand side management—a simulation of household behavior under variable prices. Energy Policy 2011;39:8163–74. [14] Cosmo VD, Lyons S, Nolan A, et al. Estimating the impact of time-of-use pricing on Irish electricity demand. Energy J -Camb Ma then Clevel Oh 2012;35(2):168–218. [15] Xu FY, Zhang T, Lai LL, Zhou H. Shifting boundary for price-based residential demand response and applications. Appl Energy 2015;146:353–70. [16] Torriti J. Price-based demand side management: assessing the impacts of time-ofuse tariffs on residential electricity demand and peak shifting in northern Italy. Energy 2012;44:576–83. [17] Woo CK, Li R, Shiu A, Horowitz I. Residential winter kWh responsiveness under optional time-varying pricing in British Columbia. Appl Energy 2013;108:288–97. [18] Faruqui A, Segici S. Household response to dynamic pricing of electricity: a survey of 15 experiments. J Regul Econ 2010;38:193–225. [19] Residential CRA. hourly load response to critical peak pricing in the state-wide pricing pilot. In: Presented by CRA international; 2006. [20] Schlomann B. Short- and long-term impact of a linear and time-dependent electricity tariff on the load curve of a utility. Berlin: ENER Joint Seminar; 1993. [21] Quantec LLC Analysis of the load impacts and economic benefits of the residential TOU rate option. Prepared for Portland general electric. [22] Hammerstrom DJ, Ambrosio R, Carlon TA, Desteese JG, Kajfasz R, Pratt RG Pacific Northwest GridWise ™ testbed demonstration projects Part I. Olympic Peninsula project. [23] CER, Electricity smart metering customer behaviour trials(CBT) rindings report; 2011. [24] Gyamfi S, Krumdieck S. Price, environment and security: exploring multi-modal motivation in voluntary residential peak demand response. Energy Policy 2011;39:2993–3004. [25] Ozturk Yusuf, Lee Gordon, Senthilkumar D, Kumar Sunil. An intelligent home energy management system to improve demand response. IEEE Trans smart Grid 2013;4(2):694–701. [26] Wright D, Yan X, Kumar S, Lee G, Ozturk Y. A Cloud-based system architecture for enabling consumer behavior modification in the electricity market. In: Proceedings of the international conference on soft computing and software engineering; 2015.
5. Conclusions The proposed paper provided a review among three different PDDR programs at the residential sector. The TOU program is easy to follow and can obtain a stable participating ratio by the end user. However, it only shift the peak into a different time period. The CPP program is also easy to follow and participants can visualize their incentives and result an effective peak shifting. The disadvantages of CPP are that it cannot be applied on a daily basis and it is not effective in reducing energy cost and carbon emission. Finally, the RTP program can reflect the dynamic relationship between the demand and supply, but it is not practical at the moment. Although the current PDDR programs achieved some level of success in peak demand shifting, other factors started to show a strong influence on the results of the current demand response mechanisms. A numerical example with an upper level of 20% DR is proposed and the results show that DR caused an increase of 11.29% in peak demand comparing with a maximum of only 7.5% in saving. With the implementation of smart metering infrastructure, demand response will play a major role in the future smart grid. It is glad to see that a group of researchers is focusing on social contexts such as policy, psychology, individual behavior, education, and income level. We believe that social level analysis will provide a better understanding of residential electricity usage that can really benefit the design of future residential demand response mechanisms. Acknowledgements This work was supported by the China Postdoctoral Science Foundation Funded Project (2016M591174). References [1] U.S. Department of Energy [Internet]. [cited 2015 May 18] Available: 〈http:// energy.gov/oe/services/technology-development/smart-grid〉. [2] Darby SJ, McKenna E. Social implications of residential demand response in cool
418
Renewable and Sustainable Energy Reviews 96 (2018) 411–419
X. Yan et al.
[27] Yan X, Wright D, Kumar S, Lee G, Ozturk Y. Real-time residential time-of-use pricing: a closed-loop consumers feedback approach. In: Proceedings of the IEEE GreenTech; 2015. [28] Yan X, Wright D, Kumar S, Lee G, Ozturk Y. Enabling consumer behavior modification through real timeenergy pricing. In: Proceedings of the first IEEE workshop on pervasive energy services; 2015. [29] Giorgio AD, Pimpinella L. An event driven smart home controller enabling consumer economic saving and automated demand side management. Appl Energy 2012;96:92–103. [30] Ericson T. Households' self-selection of dynamic electricity tariffs. Appl Energy 2011;88:2541–7. [31] Herter K, Wayland S. Residential response to critical-peak pricing of electricity: california evidence. Energy 2010;35(4):1561–7. [32] Aubin CD, Fougere E, Ivaldi M. Real-time pricing for residential customers: econometric analysis of an experiment. J Appl Econ 1995;10(4):171–91. [33] Country Energy. The country energy home energy efficiency trial; 2005. [34] Faruqui A, Sergici S. Household response to dynamic pricing of electricity: a survey of 15 experiments. J Regul Econ 2010;38:193–225. [35] Renner S, Albu M, Elburg H, Heinemann C, Lazicki A, Penttinen L, Puente F, Saele H. European smart metering landscape report 2011. [36] Widergren S, Marinovici C, Berliner T. et al. Real-time pricing demand response in operations. In: Proceedings of the IEEE power and energy society general meeting; 2012. p. 1–5. [37] Widergren S, Chan SC. Demand response optimization for smart home scheduling under real-time pricing. IEEE Trans Smart Grid 2012;3(4):1812–21. [38] Avci M, Erkoc M, Asfour SS. Residential HVAC load control strategy in real-time electricity pricing environment. In: Proceedings of the IEEE Energytech; 2012. p. 1–6. [39] Yousefi S, Moghaddam MP, Majd VJ. Optimal real time pricing in and agent-based retail market suing a comprehensive demand response model. Energy 2011;36:57165727. [40] Yoon JH, Bladick R, Novoselac A. Demand response for residential buildings based on dynamic price of electricity. Energy Build 2014;80:531–41. [41] Moghaddam MP, Abdollahi A, Rashidinejad M. Flexible demand response programs modeling in competitive electricity markets. Appl Energy 2011;88:3257–69. [42] Hammerstrom DJ, Ambrosio R, Carlon T, Desteese JG, Kajfasz R, Pratt RG. Pacific Northwest GridWise Testbed demonstration projects. Part I. Olympic Peninsula Project 2007. [43] Allcott H. Rethinking real-time electricity pricing. Resour Energy Econ Resour Energy Econ 2011;33:820–42. [44] Star A, Isaacson M, Haeg D, Kotewa L. The dynamic pricing mousetrap: why isn’t the world beating down our door? n: Proceedings of the ACEEE summer study; 2010. p. 2257–68.
[45] Rastegar M, Fotuhifiruzabad M, Choi J. Investigating the impacts of different pricebased demand response programs on home load management. J Electr Eng Technol 2014;9(3):1125–31. [46] Reiss PC, White MW. Household electricity demand, revisited. Stanford University Publication; 2002. [47] Fuller JC, Schneider KP, Chassin D. Analysis of residential demand response and double-auction markets. In: Proceedings of the IEEE power and energy society general meeting; 2011. p. 1–7. [48] EPA. Summary of research findings from the programmable thermostat market. 2005. [Online]. Available: 〈http://www.energystar.gov/ia/partners/prod_ development/revisions/dowasnloads/thermostats/Summary.pdf〉. [49] DRRC, Understanding customer behavior to improve demand response delivery in California. PIER demand response research centre research opportunity notice DRRC RON -3; 2007. [50] Soares A, Gomes A, Antunes GH. Categorization of residential electricity consumption as a basis for the assessment of the impacts of demand response actions. Renew Sustain Energy Rev 2014;30:490–503. [51] Simons SI, Rowlands IH. TOU rates and vulnerable households: electricity consumption behavior in a Canadian case study. Dep Environ Resour Stud 2007. [52] Socolow RH. Saving energy in the home: princeton's experiments at twin rivers. Cambridge, MA, USA: Ballinger Publishing Company; 1978. [53] Giorgio AD, Liberati F. Near real time load shifting control for residential electricity prosumers under designed and market indexed pricing models. Appl Energy 2014;128:119–32. [54] Corradi O, Ochsenfeld H, Madsen H, Pinson P. Controlling electricity consumption by forecasting its response to varying prices. IEEE Trans Power Syst 2013;28(1):421–9. [55] Faria F, Vale Z. Demand response in electrical energy supply: an optimal real time pricing approach. Energy 2011;36:5374–84. [56] Venkatesan N, Solanki J, Solanki SK. Residential demand response model and impact on voltage profile and losses of an electric distribution network. Appl Energy 2012;96:84–91. [57] Siano P. Demand response and smart grids-a survey. Renew Sustain Energy Rev 2014;30:461–78. [58] The NOBEL market. [Online] [cited at Sep. 2016]. Available at: 〈http://www.ictnobel.eu〉. [59] Piclo P2P energy trading, [Online] [cited at Sep. 2016]. Available at: 〈https:// www.openutility.com/piclo/〉. [60] Rahimi F, Ipakchi A, Fletcher F. The changing electrical landscape end-to-end power system operation under the transactive energy paradigm. IEEE Power Energy Mag 2016;14(3):52–62. [61] Hug G, Kar S, Wu C. Consensus + innovations approach for distributed multiagent coordination in a microgrid. IEEE Trans Smart Grid 2015;6(4):1893–903.
419