Challenges and Prospects of Negawatt Trading in Light of Recent Technological Developments
Wayes Tushar, Tapan K. Saha, Chau Yuen, Peta Ashworth, H. Vincent Poor, Subarna Basnet
aa r X i v : . [ c s . C Y ] J u l Challenges and prospects for negawatt trading inlight of recent technological developments
Wayes Tushar a , ∗ , Tapan K. Saha a , Chau Yuen b , David Smith c , Peta Ashworth a , H. Vincent Poor d , and SubarnaBasnet ea The University of Queensland, Brisbane, QLD 4072, Australia b Singapore University of Technology and Design (SUTD), 8 Somapah Road, Singapore 487372 c CSIRO Data61, Eveleigh NSW 2015, Australia d Princeton University, Princeton, NJ 08544, USA e Massachusetts Institute of Technology, Cambridge, MA 02139, USA ∗ Corresponding author’s e-mail: [email protected]
Abstract —With the advancement of the smart grid, thecurrent energy system is moving towards a future where peoplecan buy what they need, sell when they have excess, andcan trade the right of buying to other proactive consumers(prosumers). While the first two schemes already exist in themarket, selling the right of buying - also known as negawatttrading - is something that is yet to be implemented. Here,we review the challenges and prospects of negawatt trading inlight of recent technological advancements. Through reviewinga number of emerging technologies, we show that the necessarymethodologies that are needed to establish negawatt trading asa feasible energy management scheme in the smart grid arealready available. Grid interactive buildings and distributedledger technologies for instance can ensure active participationand fair pricing. However, some additional challenges need toaddress for fully functional negawatt trading mechanisms intoday’s energy market.
I. I
NTRODUCTION
Approximately two-thirds of global greenhouse gas emis-sion stems from the energy sector [1]. This creates a signif-icant opportunity for prosumers – participants in the energygrid who both consume and produce energy – who can buyrenewable energy when needed; sell any excess to otherconsumers; and trade their right to buy energy to otherprosumers in the network. Such actions allow prosumers tocontribute to combating global climate change while savingmoney. The first two phenomena of managing energy arewell established and reported in many existing studies in theliterature. For example, see [2] and [3]. However, trading theright to buy energy – also known as negawatt trading [4] –is something that is yet to be implemented.The concept of negawatt was first introduced in the mid-eighties [5], [6] as a technique of energy management. It canbe defined as an energy customer’s right to buy energy, whichis produced due to a change in their energy consumptionbehaviour. For example, an energy customer may choose toeither reschedule their energy related activities to anothertime or decide to not use energy for some activities and selltheir right to buy the saved energy to other entities withinthe energy network. Note that different demand responseprograms such as direct load controls, demand biddingprograms, demand buyback programs, emergency demandresponse programs, capacity market programs, and interrupt-ible programs [7] also involve reduction in demand fromenergy consumers. However, the rules and decisions of these programs are set by the grid with very limited inputs fromcustomers. Such incapacity to contribute to pricing decisionshas long been a barrier to customers reaping their expectedreward from their provision of flexibility to the network.This is due to rules around minimum bid size and difficultiesfor aggregators to get market access. In contrast, negawatttrading provides prosumers with flexibility and independencein deciding when to reduce their demands, at which pricethey can sell their right to buy energy, and with whom theywant to trade their right of buying, i.e., whether with the gridin the national market or with other prosumers within thenetwork through a peer-to-peer platform. Thus, prosumerscan reap their expected rewards from their provision offlexibility to the network by participating in negawatt trading.The concept of negawatt trading is also fundamentallydifferent from peer-to-peer energy trading [8]. Peer-to-peertrading is a method of energy trading between two partieswithin an electricity network without the need for anyintermediate entity with possible detrimental impacts on thenetwork if strict constraints on the physical transfer of energyare not maintained [9], [10]. Negawatt trading, on the otherhand, is a specific trading mechanism for the trading ofthe right to buy energy instead of any physical exchangeof energy. Therefore, negawatt trading has no detrimentalimpacts on the electricity network. Further, it can be tradedeither with the grid with the involvement of intermediateentities or with other customers within a network through apeer-to-peer trading platform without any intermediate entity.Thus, negawatt trading offers an unprecedented opportu-nity to prosumers to participate in the energy market bytrading their right to buy energy as opposed to energydemand. However, while advances in the information andcommunication technology industry have made extraordinaryprogress in developing energy-efficient devices [11], pricingmechanisms [12], building energy management [13], localenergy trading [8], energy market structure [14], and energypolicy [15], [16], negawatt trading has remained elusive – aconcept without any serious development. Key reasons havebeen attributed to the lack of innovation in necessary tech-nologies and a limited understanding of the social, economicand environmental implication of trading the right to buyenergy.Nevertheless, in light of recent developments negawatttrading is starting to be seen as a reality. For example, in
Japan, a number of industries have participated in the Yoko-hama Smart City Project to demonstrate the effectiveness ofnegawatt hour trading in the future [17]. Similar initiativesfor demonstrating the feasibility and conceptual architectureof negawatt trading are also being taken in the USA [4]and Australia [18], [19]. These demonstration projects clearlyestablish that the technology era with the capacity to enablefeasible operation of negawatt trading is finally here.Considering these recent technological developments, thefocus of this study is to provide a multi-disciplinary perspec-tive on the feasibility of negawatt trading in energy markets,which will ultimately contribute to improving the energyefficiency of the smart grid. Note that energy efficiency isnot a part of wholesale markets and, at present, there is nomechanism to reward energy efficiency directly for providingcost savings to consumers. However, in some states in theUS, energy efficiency is taken into account when undertakingcost-benefit analyses and calculating the appropriate amountof energy efficiency investment by utilities [20]. In the restof the manuscript, first, we discuss the early challenges thathave hindered the large-scale adaptation of negawatt tradingin the energy market. Second, we provide a well-groundedevaluation of the recent developments in technologies, whichare critically contributing to the feasibility of negawatt trad-ing. Finally, we present a number of outstanding challengesfor future research that need to be addressed to ensure thewidespread deployment of negawatt trading, followed bysome concluding remarks.II. C
HALLENGES FOR W IDE -S CALE A DAPTATION
To enable trading negawatt, a typical energy system shouldhave a number of elements including active participants,appropriate communication facilities, secure information sys-tems, suitable market mechanisms, and fair pricing tech-niques. The main limitations in the capabilities of of eachof these network elements that have prevented negawatttrading from wide-scale implementation are summarised inthe following.
A. Active participants
In a network, negawatt needs to be exchanged or tradedbetween different prosumers, that is, negawatt sellers andnegawatt buyers. Negawatt sellers are the house or buildingowners who are willing to reduce their energy consumptionto trade their rights of buying energy with negawatt buyers.Thus, clearly, for a successful negawatt trading market,prosumers need to actively participate in the exchange. Un-fortunately, until very recently, the participation of prosumershas been insufficient due to a number of reasons as discussedbelow.Inadequate technical facilities - For trading negawatt, eachparticipating building or house must have enough flexibleload, the capacity to monitor energy demand and supply inreal-time, capability to interact with its appliances as wellas with other buildings, and the ability to make intelligentdecisions based on real-time and historic information withoutcompromising on convenience. Unfortunately, a vast majorityof buildings in the world today are equipped with inefficientappliances and have no, or very limited, building controls.Consequently, lights are turned on and off manually andair conditioning systems are controlled with local room-to-room switches and thermostats. Building management system solutions are also highly expensive and, thus, difficultto justify for use in small- and medium-size buildings [21].Limited prosumer-focused management system - To en-able prosumers to create enough negawatt relies on thecapabilities of energy management system of the building.Most existing energy management systems are technology-focused, rather than prosumer-focused. However, it is criticalthat the technology integrates with the users’ experienceand positively affects prosumers’ energy usage behaviourwithout creating any notable inconvenience [22]. Due todifferent type of prosumers and their diverse preferences, it isdifficult to develop one single energy management platformthat captures all prosumers’ preference within a network.Increased concern for privacy and security - Smart energymeters are necessary for most environmental and energy-saving initiatives, including trading negawatt. However,smart meters can also be used as surveillance tools [23]. Forexample, the data collected by a smart meter can easily revealhow many showers the occupants in a house have had, whenthey are cooking, and when they are in and out of the home.Thus, concerns for privacy and security arising from the useof such smart energy solutions has been a critical issue [24]that prevents prosumers from participating in negawatt orother energy exchange programs.Lack of education - Education can have a profoundimpact in changing people’s behaviour in combatting climatechange [25] and their energy use [26]. For example, by mo-tivating people to use renewable energy, adopting a lifestylethat helps reducing energy demand, and share renewableenergy with one another. Nevertheless, environmental edu-cation has not been emphasised in either secondary or post-secondary studies. Hence, most prosumers are not motivatedto change their energy usage behaviours, which has beena barrier for adopting environmental solutions like negawatttrading in the energy market. Similarly, wider communicationand education about the likely financial benefits of tradingenergy by household prosumers has been largely absent fromthe discussion.
B. Social engagement
Social engagement and relationships are important inshaping individual and group decisions and actions about en-ergy consumption. Key elements include social and personalnorms [27], social identity [28] and trust [29]. We know thatwhen individuals are faced with making decisions aroundenergy consumption they will be influenced by expectationsof their peers about what is acceptable, the likely conse-quences of their choices as well as their trust in the actors andinstitutions involved [30]. For example, when people gatherat social events, they may discuss their use of energy andwhy they consume in the ways they do and weigh up whetherthere is any motivation or peer pressure to change or adoptnew behaviours [26]. [31] also stressed the importance ofsocial relationships in shaping energy behaviours includingthose with friends and family, agencies and communitygroups and who they identify with. However, the impact ofsocial relations on energy use has often been neglected in thedesign of energy markets. Failing to cultivate relationships orunderstanding the social motivations that exist across energyconsumers can be detrimental for enabling negawatt sharingbetween prosumers. !" !" ."
Fig. 1:
Challenges of negawatt trading-
A number of prosumers,such as grid-efficient buildings, are located in a community thatwant to participate in negawatt trading. Now, to enable this buildingsto trade negawatt among themselves, a number of challenges atdifferent aspects of the negawatt trading framework need to address.Prosumers need to actively involved in the negawatt trading, forwhich educating them about the benefit of such participation isessential. Prosumers also need to be socially interactive to reap themaximum benefit of negawatt trading. There should be appropriatecommunication facilities for prosumers within the energy networkto securely communicate with one another without compromisingtheir privacies. Further, all information exchange and monetarytransactions should be done through a secured information system.All participants should have equal access to market mechanism andthe price per negawatt that each participants will receive shouldbe decided in a fair manner. Finally, the regulation of the regionshould environmentally friendly and support the establishment ofnegawatt trading to integrate negawatt trading into the existingmarket and supply system. (Source of clip art within the figure:https://pixabay.com/ )
C. Appropriate communication facilities
For trading negawatt, prosumers need to communicatewith one another as well as different energy stakeholderswithin the network. Therefore, a suitable communicationfacility is a major requirement for negawatt trading. Whilethere are many communication infrastructures that have beenreported in the literature including structured, unstructured,and hybrid architectures [32], the choice of a communicationarchitecture that fulfils the performance requirements oflatency, throughput, reliability, and security for negawatttrading is critical.
D. Secured information system
A critical element of an energy network, to facilitatenegawatt trading between different prosumers, is a well-functioning secure information system. The information sys-tem needs to enable all participants to be integrated withthe market mechanism, and provide participants with accessto the same accurate information about the price, energystatus (supply and demand), ancillary market information,and environmental conditions. Importantly, the informationsystem needs to be secured as well as conforming to theprivacy requirements of participants.
E. Suitable market mechanism
Market mechanisms consisting of market allocation, pay-ment rules, and clearly outlined pricing formats play acentral role in negawatt trading platforms. The main purposeof a market mechanism is to help participants to achieve their desired revenue (or cost reduction) by matching theselling and buying of orders in real-time. Each negawattseller can influence the maximum availability of negawattwithin the market through its demand reduction capacity andsubsequent pricing. Different market mechanisms may needto be created and made to co-exist to generate and tradeenough negawatt at each stage of market operation, which isyet to be implemented.
F. Fair pricing mechanism
Pricing mechanisms are designed to efficiently balance theenergy supply and demand within the network. However,pricing mechanisms for creating and selling negawatt couldbe different compared to the pricing of traditional energymarkets. This is because creating negawatt will not have anymarginal cost. Hence, participants may acquire more finan-cial benefits from selling their right to buy energy. However,creating negawatt does involve scheduling and regulatingenergy usage related activities that may cause additionalinconvenience to the negawatt sellers. This inconvenienceneeds to be taken into account when developing fair pricingmechanisms for negawatt trading. At the same time, theavailability of negawatt should influence the set price per unitof negawatt. For example, a higher availability of negawattwithin the network may lower the trading price and viceversa.
G. Environmental friendly regulation
The decision of the trading of negawatt in the future elec-tricity market will most likely be governed by regulation andenergy policy. Thus, the legislation in a country governs whatkind of market design will be allowed, whether there willbe any taxes or fees for such trading, and how the negawattmarket will be integrated into the existing energy market andsupply systems. Governments could provide new policies forsystem operators, network companies, and utilities to supportnegawatt trading. For example, distributed system operatorscould be motivated by new revenue streams and incentivesto engage in negawatt trading. Network companies couldconsider negawatt solutions alongside supply-side options(network upgrades, for example), and the utility could earna portion of the savings from negawatt trading within thenetwork [33].An overview of challenges in different aspects of a ne-gawatt trading framework is shown in Fig. 1.III. E
VALUATION OF E NABLING T ECHNOLOGICAL D EVELOPMENTS
To address these challenges, significant technological in-novations and developments have been made to establishnegawatt trading – an inter-disciplinary technology with rel-evance to energy engineering, IoT, optimisation, economics,and human behaviour – as a feasible practice in energymarkets. What follows is an overview of technologies thathave laid the pathway for negawatt trading among prosumersand between prosumers and the grid.
A. Grid-interactive efficient buildings
Grid interactive efficient buildings (GEBs) [34] are a newconcept, in which buildings have the capacity to monitor and control their real-time energy generation and dispatch as wellas optimise their energy usage for service, occupant needsand preferences, and cost reduction (or, revenue maximisa-tion) in an integrated fashion [35]. The main characteristicsof a GEB that would enable it to participate in negawatttrading include:Reliable and low latency communication facility - EachGEB is equipped with equipment that supports two-wayconnectivity and communication with devices, applianceswithin the buildings as well as with other GEBs within thenetwork and the grid. Equipment should have the capacityto monitor, report, and provide flexibility to shed, shift, ormodulate consumption in response to the control signal sentby the management system of the building.Intelligent management system - The management sys-tem of a GEB can monitor, incorporate, predict, and learnfrom occupant needs and preferences, outdoor conditions(weather), and from other GEBs’ needs and grid require-ments. Based on such prediction and learning, it can coor-dinate and execute complex control strategies that adapt tochanging conditions over multiple time scales. Further, themanagement system can quantitatively estimate and verifythe energy and demand savings from different strategies. Itcan optimise across a choice of multiple strategies to balanceefficiency with flexibility and occupancy comfort.Interoperable, secured, and trusted system - Finally, theoverall system responsible for negawatt trading of a GEBshould be interoperable and have the capacity to effectivelyexchange data and control signals among devices, appliances,management systems, and between different GEBs and thegrid in a secure fashion. Data security and protection shouldbe resilient against any cyber attack from unauthorisedsources, and as a trusted system, a GEB may need toenforce different specified security policies for performingapplications in different contexts.
B. Distributed ledger technology
Distributed ledger technology (DLT) is a digital, shared,and distributed database for recording transactions of assets,and has proven capabilities to improve the efficiency of cur-rent energy practices and processes [2], [36]. In particular, re-cent advancements in DLT including blockchain, smart con-tracts, consortium blockchain, Hyperledger, Ethereum, di-rected acyclic graph, Hashgraph, Holocahin, and Tempo [37]can contribute to negawatt trading due to a number of usefulcharacteristics.For example, DLT can realise automated billing for pro-sumers and help negawatt traders with micro-payments, pay-as-you-go solutions, and payment platforms for pre-paidmeters. DLT, in conjunction with artificial intelligence tech-niques, can identify prosumers’ energy usage behaviour pat-terns, and subsequently manage negawatt trading accordingto individual preferences, energy profiles, and environmentalconcerns. DLT-enabled distributed trading platforms alsohave the capacity to be part of market operations such aswholesale market management, commodity trading transac-tions, and risk management. It enforces trust within the sys-tem without the requirement for intermediaries through theuse of a distributed ledger, consensus algorithm, and token.Further, DLT has the capacity to improve the control of adecentralised system [38]. Thus, the adoption of blockchain !" $%&’(’)) )*’+,-&./,)0,-.’.12334-,)5$%&’(’)) )*’+,-&.6%)/%%-.12334-,),%7
Fig. 2:
Market for peer-to-peer negawatt sharing-
Prosumers canshare their negawatt with one another through peer-to-peer sharing.In a community, peers could be apartment buildings, individualhomes with distributed energy resources, electric vehicles, andcommunity storage devices. Negawatt can be shared within acommunity via a community manager such as a distribution systemoperator (DSO). However, it is also possible to share negawattamong different communities. In such cases, community managersneed to coordinate with one another in order to decide on thetrading parameters including the price and the amount of negawattto be traded. A third-party entity such as an independent systemoperator can facilitate this coordination among different communitymanagers to trade negawatts across communities. (Source of clipart within the figure: https://pixabay.com/ ) in the local negawatt trading market can increase behind-the-meter activities such as production of negawatt throughregulating self-consumption. Furthermore, DLT can be usedfor communication of smart devices, data transmission orstorage [38], which is very important to enable negawatt trad-ing. It can also assist network management of decentralisednetworks via flexible asset and service management. Finally,DLT has the capability to protect privacy, data confidentiality[38], and identity management [39]. The immutable recordsand transparency provided by DLT can significantly improveauditing and regulatory compliance of negawatt trading [39].
C. Peer-to-peer sharing
In Peer-to-peer sharing, the participants of an energynetwork can share some of their own resources with oneanother. These shared resources provide the service andcontent offered by the network and can be accessed byother peers directly, without the intervention of intermediaryentities [40]. Further, in a peer-to-peer network, any entitycan be removed or added, if necessary, without the networksuffering from any loss of network service.Over the last few years, a large number of findings havebeen reported on the advancement of peer-to-peer sharingin the electricity network. These results cover the aspects ofenergy cost reduction via peer-to-peer sharing [41], balancingof supply and demand of energy [42], engaging prosumersin peer-to-peer sharing [43], developing appropriate pricingmechanisms [44], identifying uncertainties in peer-to-peersharing [45], transaction security [46], and demonstration ofpilot projects [47]. Learning from these studies has the po-tential to significantly advance the research and deploymentof negawatt trading within energy systems. For example, ne-gawatt sharing could be thought of as an alternate version ofpeer-to-peer trading, in which prosumers’ willingness to buyenergy, i.e., negawatt, is shared among the participants withindecentralised markets [48], community-based markets [49], and composite markets [50], instead of watts. An examplediagram of such a trading paradigm is shown in Fig. 2.
D. High speed communication
For successful negawatt trading in energy markets, acritical requirement is a high-speed communication servicethat is capable of enabling immersive remote operationsand interactions with a physical world with low-latency.The 5th generation communication networks (5G) will befundamental to fulfilling this requirement with their capacityto support a wide range of highly demanding services andapplications, pushing the network capabilities to provideextreme performance benefits. This includes the support ofmassively interconnected devices, and providing necessaryservices to enable operations and manipulation of physicalobjects over distance with reliability and low latency, de-scribed as the tactile internet in [51].Essentially, the 5G network will provide several interde-pendent tools that will be able to offer the flexibility toenable negawatt trading. For example, 5G networks willbe highly programmable and built on network functionvirtualisation [51]. Thus, more of the network functionalitycould be implemented in software executed in virtual envi-ronments on general purpose hardware, and less specialisedhardware built and optimised specifically for certain networkfunctions. A key enabler for a flexible and programmable 5Gplatforms is the distributed cloud. As a result, the softwarecan be deployed and executed at an optimal place within thenetwork. The software can then become a virtual networkfunction (VNF) or an end-user application software, e.g.,BMS of a building, with bi-directional communication fornegawatt trading purposes.
E. Internet of Things
Internet of Things (IoT) is essentially the interconnectionbetween computing devices embedded in our everyday lifevia the Internet, which enables them to send and receivedata. It consists of multiple layers including the device layer,network layer, cloud management layer, and applicationlayer. While the device layer is responsible for sensing theenvironment, collecting data, and controlling flexible loadswithin a building, the purpose of the network layer is toconnect the devices to the application layer. The applica-tion layer provides services to the end-users by controllingflexible loads. Examples of such services include demandmanagement, dynamic pricing, energy management, andhome security services. The cloud management layer ensuresuser authentication, along with user and data management.More details of IoT can be found in [52].For participating in negawatt trading, it is important forbuildings to be aware – in real-time – of energy produc-tion, demand, and possibilities of scheduling or throttlingof flexible devices through sophisticated sensing and con-trol capabilities. Recent advancements in integrated dataacquisition and control systems based on open architecturesand cloud-enabled IoT allow building owners (or managers)to monitor and sense buildings’ environmental parameters,collect relevant human activity information, estimate energyusage, and then based on the available energy supply directBMSs to manipulate the activities of flexible loads accordingto expectations and specified rules [21]. An overview of application of IoT for negawatt trading is demonstrated inFig. 3.
F. Distributed energy resources
Distributed energy resources (DERs) include behind-the-meter generation, energy storage, inverters, electric vehicles,and controllable loads and their associated applications.DERs offer significant opportunities to reshape the energyfuture and negawatt trading is one of them. Recent advance-ment in DERs in terms of their manufacturing material, sizeof the resources, artificial intelligence, and computationalspeed has significantly increased their affordability via costreduction, reliability of operation, and security and privacy.For example, the price of storage has fallen from 2010to 2018 and is projected to reduce by an additional by2030 [53]. Further, with the integration of blockchain withDERs, security and privacy issues have been significantlyaddressed.Now, with artificial intelligence and novel computationalapproaches, participants can simply define rules in theirapplications, e.g., through their mobile phones, and the trans-active meters within their houses can automatically performtrading on their behalf [54]. Such advancements make iteasy for prosumers to trade their negawatt with one anotherand with the grid and contribute to improving environmentalsustainability.
G. Social media platforms
H. Behavioural economics
Negawatts are a result of reducing demand. While pro-sumers are gaining greater awareness of the value andneed for sustainable energy practices like negawatts, manyprosumers still fail to take noticeable steps towards en-ergy efficiency measures due to a significant discrepancybetween peoples’ self-reported knowledge, values, attitudesand intentions, and their observable behaviour [56]. Withthe emergence of the application of behavioural economicsin the area of smart grid and energy management [57],it is becoming possible to reduce such ‘knowledge-action’
Fig. 3:
IoT for prosumers decision-making-
Internet of Things (IoT) facilitates interconnection between the physical systems of theprosumers via the Internet and enables them to send and receive data over the virtual cloud-based system. Examples of such physicalsystems include household appliances and distributed energy resources. By converting the physical system to its digital-twin, IoT leveragesthe negawatt trading platform to sense the environment, collect data, monitor energy usage and generation, and predict and control the use offlexible loads within a building. Thus, IoT enables prosumers to opportunistically participate in both intra-community and inter-communitynegawatt sharing with appropriate market mechanism and pricing schemes. (Source of clip art within the figure: https://pixabay.com/ ) and ‘value-action’ gaps. For example, feedback interven-tions using digital technologies can be very effective atpromoting energy conservation behaviour. In particular, real-time feedback on specific and energy-intensive informationsuch as energy usage, negawatt demand in the network,energy price, and potential revenue for negawatt selling mayinduce considerable behavioural change and production ofnegawatts [58]. Further, energy education and behaviouralprograms on the potential impact of negawatt production onenergy savings and environmental sustainability could alsoimpact people’s energy savings behaviour in the long run.A trial conducted in [59] has shown that youth educationcan potentially influence environmental and/or knowledge,attitudes, and behaviour towards energy demand reduction,and that children’s energy behaviour also affects that ofparents. Further, prosumers’ perceptions of savings, e.g.,to generate negawatts, are affected by cognitive processessuch as the recall of previous bills [60], and, in the caseof negawatt trading, recall of previous economic benefitsobtained via trading negawatts with other participants in theenergy network.
I. Computational approaches
Substantial work has been done in terms of exploitingcomputation techniques for trading and sharing resourcesacross energy networks. Examples of such computationalapproaches include game theory, double auction, constrainedoptimisation, and artificial intelligence.Game theory - Game theory is a mathematical tool thatanalyses the strategic decision making process of a numberof players in a competitive situation, in which the actiontaken by one player depends on and affects the actions ofother players [61]. It has been used extensively in recentyears to balance energy supply and demand, develop pricingschemes, increase prosumers’ engagement, and provide net-work services. It has significant potential to be utilised forsharing energy resources including both watts and negawattwithin an energy network [54]. Double auction - Double auction involves a market of anumber of buyers and sellers seeking to interact with oneanother to trade or share their resources [62]. Both buyersand sellers need to truthfully report their bids for efficientand sustainable operation of the market. Double auction hasbeen used in the energy market for demand-supply balance,network services, and prosumers engagement [8].Constrained optimisation - Resource sharing in the energymarket has been heavily captured via various constrainedoptimisation techniques including linear programming, non-linear programming, mixed-integer linear programming, andalternating direction method of multipliers (ADMM). Forinstance, applications of these optimisation techniques can befound in storage management [63], energy management [64],and scheduling of flexible loads [65].Artificial intelligence - More recently, artificial intelligencetechniques have found many applications in learning theenergy usage behaviour of flexible loads and subsequentlycontrolling their energy consumption strategies to achievecertain objectives. Examples of such objectives may includedemand-response via reinforcement learning [66], buildingmanagement through deep learning [67], and energy costreduction through an artificial neural network-based ap-proaches [68].Indeed, the advancement of these computational tech-niques would also be valuable in capturing the decision-making processes of prosumers, e.g., see [69], in trading theirnegawatt, while simultaneously optimising their householdactivities without affecting their preferences. For instance,computational approaches can be utilised for market set-tlement, pricing mechanisms, forecasting of reduction indemand for individual participants, enabling strategic inter-action between different participants (for trading in P2P plat-forms) and between the retailer and participants (for tradingin the retail-based market), and for capturing and quantifyingparticipants’ convenience based on each participant’s uniquecircumstances and parameters. Further, by integrating designthinking [22] and motivational techniques [70] with the designed schemes, the computational approaches can bemade more prosumer-centric with potential to attract moreparticipants to actively involve in trading negawatt in theenergy market.IV. R
EMAINING C HALLENGES & F
UTURE R ESEARCH
Despite recent technological advancement and efforts, alarge number of challenges have yet to be addressed beforethe successful deployment of techniques that can facilitatewide-scale negawatt trading in the energy market can occur.Below is a summary of challenges that need to be addressedfor successful implementation of negawatt markets.
A. Appropriate pricing schemes for negawatt trading
Since the production of negawatt relies on managing de-mand of prosumers, it requires prosumers to either scheduletheir energy related activities or regulate the energy con-sumption of the loads. Therefore, there are possibilities thatprosumers may experience inconvenience to produce enoughnegawatt within the network. To compensate such inconve-nience and encourage prosumers to participate in negawatttrading markets, suitable pricing schemes are necessary.Further, inconvenience is not quantifiable like energy andcan vary extensively for different energy customers due todifferent circumstances of generation, demand, preferences,and views of environmental sustainability. As a consequence,the same price per unit of negawatt may not reflect thetrue reward for all negawatt producers to participate in suchtrading. Therefore, appropriate price discrimination will needto be introduced in order to ensure a fair and incentive-compatible revenue for all negawatt traders within the mar-ket.
B. Network security with low computational expenses
Two critical factors for negawatt trading are trust andsecurity. While identity check and verification might benecessary for prosumers to participate in trading, the securityof data injection needs to be ensured in an inexpensiveway. At present, blockchain is being considered as the mostappropriate trading platform due to its capability to ensuresecured and trusted transactions. However, as detailed in[71], providing secured trading transactions via blockchainis very computationally expensive. Thus, adopting such acomputationally expensive technique will require extensivepower to serve the participants. As a consequence, partici-pants will need to share the cost of this service, which willmake the trading of negawatts very expensive. Hence, forthe sustainability of the market in terms of cost, securityand trust, new and computationally less expensive platformswill be required for the wide-spread realisation of negawatttrading.
C. Comfort and convenience
Since the production of negawatt relies significantly uponthe scheduling of activities of flexible loads, there is ahigh chance that it would affect the comfort and day-to-day activities of building occupants. Hence, algorithms forregulating the schedule of different appliances needs to bedata-driven. Further, innovation is required in developing theartificial learning algorithms to take into account the diverse behavioural patterns of users, and subsequently scheduleflexible loads within houses such that prosumers’ comfortand regular day-to-day activities are not affected. An exampleof such a model for regulating HVAC’s temperature controlcan be found in [72].
D. Data accessibility with privacy
To improve prosumers’ decisions about negawatt trading,statistically relevant and accurate energy transaction andusage data need to be made available across communities.However, this accessible data also need to provide pri-vacy to each prosumer. Therefore, demonstrated private dataanonymisation needs to be facilitated for the necessary datasharing, while simultaneously providing sufficient accuracyfor interrogation of data.
E. A unified framework for both watt and negawatt trading
A suitable peer-to-peer network is necessary for negawatttrading between prosumers. With extensive use of DERs,it is reasonable to expect that a prosumer will be able toparticipate in both energy and negawatt trading within theenergy system. Hence, enough flexibility needs to be ensuredwithin the decision making framework of prosumers so thateach prosumer can switch its role as a participant betweenthe negawatt and energy market based on available marketinformation and its existing state of energy. Further, thenetwork should also possess sufficient interoperability tofacilitate such switching between markets for a very largenumber of participants.
F. Defining the roles of different stakeholders
Clearly, different stakeholders would be interested inexploiting negawatt trading to deliver different services totheir customers and maintain network security. For example,generators may want to use them for reducing productionvolatility, distribution network service providers for demandconstraint, and retailers may want to combat energy imbal-ances. Nevertheless, these goals could conflict with one an-other. Hence, negawatt trading schemes need to be designedsuch that they do not affect participants’ independence andbenefits. V. C
ONCLUSION
While the concept of negawatt trading is not new, thisperspective confirms that until recently, there have been anumber of barriers which have prevented its deployment.These range from social, technical and economic. However,amidst growing concerns of rising greenhouse gas emissions,energy prices and energy security it seems the time may nowbe ripe for widespread deployment of negawatt trading. Thishas been enabled through improvements in communicationand technological advances which were not present in theeighties when the concept was first introduced. In spite ofthese developments there remain a number of challengesthat will need to be addressed to ensure successful imple-mentation of such a scheme. These include establishing aframework that recognises the importance of appropriatepricing schemes, market mechanisms, the importance oftrust, security concerns and low barriers to entry. As moresuccessful case study examples are implemented negawatttrading should become a socially accepted norm as part ofany energy market platform. A CKNOWLEDGEMENTS
This work was supported in part by the Advance Queens-land Industry Research Fellowship AQRF11016-17RD2, inpart by the University of Queensland Solar (UQ Solar: solar-energy.uq.edu.au), in part by the SUTD-MIT InternationalDesign Centre (idc: idc.sutd.edu.sg), and in part by the U.S.National Science Foundation under Grants DMS-1736417and ECCS-1824710.C
OMPETING INTERESTS
The authors declare no competing interests.R
EFERENCES[1] Florian Egli, F., Steffen, B. & Schnidt, T. A dynamic analysis offinancing conditions for renewable energy technologies.
Nat. Energy , 1084–1092 (2018).[2] Thomas, L., Zhou, Y., Long, C., Wu, J. & Jenkins, N. A general formof smart contract for decentralized energy systems management. Nat.Energy , 140–149 (2019).[3] Morstyn, T., Farrell, N., Darby, S. J. & Mcculloch, M. D. Usingpeer-to-peer energy-trading platforms to incentivize prosumers to formfederated power plants. Nat. Energy , 94–101 (2018).[4] Jing, Z., Pipattanasomporn, M. & Rahman, S. Blockchain-basedNegawatt trading platform: Conceptual architecture and case studies.In Proc. of IEEE PES GTD Grand International Conference andExposition Asia (GTD Asia) , 68–73 (Bangkok, Thailand, 2019).[5] Lovins, A. B. Saving gigabucks with negawatts.
Public UtilitiesFortnightly , 19–26 (1985).
Pioneer paper on negawatta trading. [6] Lovins, A. B. The negawatt revolution – Solving the CO Optimal Design and Retrofit of Energy Efficient Buildings,Communities, and Urban Centers , chap. 4, 189–245 (Butterworth-Heinemann, Oxford, UK, 2018).[8] Tushar, W., Saha, T. K., Yuen, C., Smith, D. & Poor, H. V. Peer-to-peertrading in electricity networks: An overview.
IEEE Trans. Smart Grid , 3185–3200 (2020). A comprehensive overview of peer-to-peerenergy (Watt) sharing in electricity networks. [9] Guerrero, J., Chapman, A. C. & Verbiˇc, G. Decentralized P2P energytrading under network constraints in a low-voltage network.
IEEETrans. Smart Grid , 5163–5173 (2019).[10] Baroche, T., Pinson, P., Latimier, R. L. G. & Ahmed, H. B. Exogenouscost allocation in peer-to-peer electricity markets. IEEE Trans. PowerSyst. Renew. Sustain. Energy Rev. , 411–419 (2018).[13] Hannan, M. A. et al. A review of internet of energy based buildingenergy management systems: Issues and recommendations.
IEEEAccess , 38997–39014 (2018).[14] Sousa, T. et al. Peer-to-peer and community-based markets: Acomprehensive review.
Renew. Sustain. Energy Rev. , 367–378(2019).[15] Annunziata, E., Rizzi, F. & Frey, M. Enhancing energy efficiency inpublic buildings: The role of local energy audit programmes.
EnergyPolicy , 364–373 (2014).[16] Blanchet, T. Struggle over energy transition in Berlin: How dograssroots initiatives affect local energy policy-making? Energy Policy , 246–254 (2015).[17] Honda, K., Kusakiyo, K., Matsuzawa, S., Kosakada, M. & Miyazaki,Y. Experiences of demand response in yokohama demonstrationproject. CIRED - Open Access Proc. J. \ emergency-reserves-for-extreme-peaks.Accessed on 15 January, 2020.[19] Currie, G., Evans, R., Duffield, C. & Mareels, I. Policy options toregulate PV in low voltage grids-Australian case with internationalimplications. Technol. Econ. Smart Grids Sustain. Energy , 10 (2019).[20] Rosenow, J. & Thomas, S. Rewarding energy efficiency for energysystem services through markets: Opportunities and challenges inEurope (2020). URL https://doi.org/10.5281/zenodo.3634842. [21] Tushar, W. et al. Internet of things for green building management:Disruptive innovations through low-cost sensor technology and artifi-cial intelligence.
IEEE Signal Process. Mag. , 100–110 (2018).[22] Tushar, W. et al. Exploiting design thinking to improve energyefficiency of buildings.
Energy , 117141:1–117141:16 (2020).
A novel application of design innovation for improving energyefficiency. [23] Wang, Y., Chen, Q., Hong, T. & Kang, C. Review of smart meter dataanalytics: Applications, methodologies, and challenges.
IEEE Trans.Smart Grid , 3125–3148 (2019).[24] Zhang, K. et al. Security and privacy in smart city applications:Challenges and solutions.
IEEE Commun. Mag. , 122–129 (2017).[25] Anderson, A. Climate change education for mitigation and adaptation. J. Educ. Sustain. Dev. , 191–206 (2012).[26] Dowd, A., Ashworth, P., Carr-Cornish & Stenner, K. Energymark:Empowering individual Australians to reduce their energy consump-tion. Energy Policy , 264–276 (2012).[27] Schwartz, S. H. Words, deeds and the perception of consequences andresponsibility in action situations. J. Pers. Soc. Psychol. , 232–242(1968).[28] Tajfel, H. & Turner, J. C. The social identity theory of intergroupbehavior. In Worchel, S. & Austin, W. G. (eds.) Psychology ofIntergroup Relations , 7–24 (Nelson-Hall, Chicago, IL, 1986).[29] Siegrist, M. The influence of trust and perceptions of risks and benefitson the acceptance of gene technology.
Risk Anal. , 195–203 (2000).[30] Huijts, N. M. A., Molin, E. J. E. & Steg, L. Psychological factorsinfluencing sustainable energy technology acceptance: A review-basedcomprehensive framework. Renew. Sustain. Energy Rev. , 525–531(2012).[31] Hargreaves, T. & Middlemiss, L. The importance of social relationsin shaping energy demand. Nat. Energy , 195–201 (2020). Identifiesthree types of social relation that significantly influence energydemand. [32] Jogunola, O. et al.
Comparative analysis of P2P architecture for energytrading and sharing.
Energies IEEEInd. Electron. Mag. , 106–118 (2019).[37] Zia, M. F. et al. Microgrid transactive energy: Review, architectures,distributed ledger technologies, and market analysis.
IEEE Access Proc.of International Conference on Peer-to-Peer Computing , 101–102(Linkoping, Sweden, 2001).[41] Si, F. et al.
Cost-efficient multi-energy management with flexiblecomplementarity strategy for energy internet.
Appl. Energy , 803–815 (2018).[42] Li, Y., Yang, W., He, P., Chen, C. & Wang, X. Design and managementof a distributed hybrid energy system through smart contract andblockchain.
Appl. Energy , 390–405 (2019).[43] Kirchhoff, H. & Strunz, K. Key drivers for successful development ofpeer-to-peer microgrids for swarm electrification.
Appl. Energy ,46–62 (2019).[44] Morstyn, T. & McCulloch, M. Multi-class energy management forpeer-to-peer energy trading driven by prosumer preferences.
IEEETrans. Power Syst. , 4005–4014 (2019).[45] Moret, F., Baroche, T., Sorin, E. & Pinson, P. Negotiation algorithmsfor peer-to-peer electricity markets: Computational properties. In Proc.of Power Systems Computation Conference (PSCC) , 1–7 (Dublin,Ireland, 2018). [46] Noor, S., Yang, W., Guo, M., Dam, K. H. & Wang, X. Energy demandside management within micro-grid networks enhanced by blockchain.
Appl. Energy , 1385–1398 (2018).[47] Abrishambaf, O., Lezama, F., Faria, P. & Vale, Z. Towards transactiveenergy systems: An analysis on current trends.
Energy Strat. Rev. , 100418:1–100418:17 (2019). Very comprehensive review of thetransactive energy system. [48] Sorin, E., Bobo, L. & Pinson, P. Consensus-based approach to peer-to-peer electricity markets with product differentiation.
IEEE Trans.Power Syst. , 994–1004 (2019).[49] Tushar, W. et al. Energy storage sharing in smart grid: A modifiedauction-based approach.
IEEE Trans. Smart Grid , 1462–1475(2016). A new approach of storage sharing in an interactive energynetwork. [50] Liu, N. et al.
Online energy sharing for nanogrid clusters: A Lyapunovoptimization approach.
IEEE Trans. Smart Grid , 4624–4636 (2018).[51] Sachs, J. et al. Adaptive 5g low-latency communication for tactileinternet services.
Proc. IEEE , 325–349 (2019).[52] Viswanath, S. K. et al.
System design of the internet of things forresidential smart grid.
IEEE Wirel. Commun. , 90–98 (2016).[53] Goldie-Scot, L. A behind the scenes takeon lithium-ion battery prices (2019). URLhttps://about.bnef.com/blog/behind-scenes-take-lithium-ion-battery-prices/.[54] Tushar, W. et al. Transforming energy networks via peer-to-peerenergy trading: The potential of game-theoretic approaches.
IEEESignal Process. Mag. , 90–111 (2018).[55] Wolske, K. S., Gillingham, K. T. & Schultz, P. W. Peer influence onhousehold energy behaviours. Nat. Energy , 202–212 (2020).[56] Frederiks, E. R., Stenner, K. & Hobman, E. V. Household energy use:Applying behavioural economics to understand consumer decision-making and behaviour. Renew. Sustain. Energy Rev. , 1385–1394(2015).[57] Saad, W., Glass, A. L., Mandayam, N. B. & Poor, H. V. Towarda consumer-centric grid: A behavioral perspective. Proc. IEEE ,865–882 (2016).[58] Tiefenbeck, V., W orner, A., Sch ob, S., Fleisch, E. & Staake, T.Real-time feedback promotes energy conservation in the absence ofvolunteer selection bias and monetary incentives.
Nat. Energy , 35–41(2018). Discusses the importance of real-time feedback on energysavings. [59] Boudet, H. et al.
Effects of a behaviour change intervention for GirlScouts on child and parent energy-saving behaviours.
Nat. Energy ,1–10 (2016).[60] White, L. V. & Sintov, N. D. Inaccurate consumer perceptions of monetary savings in a demand-side response programme predictproramme acceptance. Nat. Energy , 1101–1108 (2018).[61] Bas¸ar, T. & Olsder, G. J. Dynamic Noncooperative Game Theory (Academic Press, New York, NY, 1995).[62] Saad, W., Zhu Han, Poor, H. V. & Bas¸ar, T. A noncooperativegame for double auction-based energy trading between phevs anddistribution grids. In
Proc. of IEEE International Conference onSmart Grid Communications (SmartGridComm) , 267–272 (Brussels,Belgium, 2011).[63] Long, C., Wu, J., Zhou, Y. & Jenkins, N. Peer-to-peer energysharing through a two-stage aggregated battery control in a communityMicrogrid.
Appl. Energy , 261–276 (2018).[64] Nguyen, S., Peng, W., Sokolowski, P., Alahakoon, D. & Yu, X.Optimizing rooftop photovoltaic distributed generation with batterystorage for peer-to-peer energy trading.
Appl. Energy , 2567–2580(2018).[65] L¨uth, A., Zepter, J. M., del Granado, P. C. & Egging, R. Localelectricity market designs for peer-to-peer trading: The role of batteryflexibility.
Appl. Energy , 1233–1243 (2018).[66] V´azquez-Canteli, J. R. & Nagy, Z. Reinforcement learning for demandresponse: A review of algorithms and modeling techniques.
Appl.Energy , 1072–1089 (2019).[67] Konstantakopoulos, I. C. et al.
A deep learning and gamification ap-proach to improving human-building interaction and energy efficiencyin smart infrastructure.
Appl. Energy , 810–821 (2019).[68] Reynolds, J., Rezgui, Y., Kwan, A. & Piriou, S. A zone-level, buildingenergy optimisation combining an artificial neural network, a geneticalgorithm, and model predictive control.
Energy , 729–739 (2018).[69] Okawa, Y. & Namerikawa, T. Distributed optimal power managementvia negawatt trading in real-time electricity market.
IEEE Trans. SmartGrid , 3009–3019 (2017). A demonstration of how negawatttrading can be performed for power management in real-timeenergy market. [70] Tushar, W. et al.
A motivational game-theoretic approach for peer-to-peer energy trading in the smart grid.
Appl. Energy , 10–20(2019).
Application of motivational psychology in peer-to-peerenergy sharing. [71] Fairley, P. Blockchain world - feeding the blockchain beast if bitcoinever does go mainstream, the electricity needed to sustain it will beenormous.
IEEE Spectr. , 36–59 (2017).[72] Li, W. et al. Data driven electricity management for residential airconditioning systems: An experimental approach.
IEEE Trans. Emerg.Top. Comput.7