**Contents**


## **About the Special Issue Editors**

**Antonio Colmenar Santos** has been a senior lecturer in the field of Electrical Engineering at the Department of Electrical, Electronic and Control Engineering at the National Distance Education University (UNED) since June 2014. Dr. Colmenar-Santos was an adjunct lecturer at both the Department of Electronic Technology at the University of Alcala and at the Department of ´ Electric, Electronic and Control Engineering at UNED. He has also worked as a consultant for the INTECNA project (Nicaragua). He has been part of the Spanish section of the International Solar Energy Society (ISES) and of the Association for the Advancement of Computing in Education (AACE), working in a number of projects related to renewable energies and multimedia systems applied to teaching. He was the coordinator of both the virtualisation and telematic Services at ETSII-UNED, and deputy head teacher and the head of the Department of Electrical, Electronics and Control Engineering at UNED. He is the author of more than 60 papers published in respected journals (http://goo.gl/YqvYLk) and has participated in more than 100 national and international conferences.

**David Borge Diez** has a Ph.D. in Industrial Engineering and an M.Sc. in Industrial Engineering, both from the School of Industrial Engineering at the National Distance Education University (UNED). He is currently a lecturer and researcher at the Department of Electrical, Systems and Control Engineering at the University of Leon, Spain. He has been involved in many national and ´ international research projects investigating energy efficiency and renewable energies. He has also worked in Spanish and international engineering companies in the field of energy efficiency and renewable energy for over eight years. He has authored more than 40 publications in international peer-reviewed research journals and participated in numerous international conferences.

**Enrique Rosales Asensio** is an industrial engineer with postgraduate degrees in electrical engineering, business administration, and quality, health, safety and environment managemen<sup>t</sup> systems. He has been a lecturer at the Department of Electrical, Systems and Control Engineering at the University of Leon, and a senior researcher at the University of La Laguna, where he has been ´ involved in water desalination project in which the resulting surplus electricity and water would be sold. He has also worked as a plant engineer for a company that focuses on the design, development and manufacture of waste-heat-recovery technology for large reciprocating engines; and as a project manager in a world-leading research centre. Currently he is an associate professor at the Department of Electrical Engineering at the University of Las Palmas de Gran Canaria.

## **Preface to "Locally Available Energy Sources and Sustainability"**

Renewable energy is electricity generated by fuel sources that restore themselves over a short period of time and do not diminish. Although some renewable energy technologies impact n the environment, renewables are considered environmentally preferable to conventional sources and, when replacing fossil fuels, have significant potential to reduce greenhouse gas emissions. This book focuses on the environmental and economic benefits of using renewable energy, which include: (i) generating energy that produces no greenhouse gas emissions from fossil fuels and reduces some types of air pollution, (ii) diversifying energy supply and reducing dependence on imported fuels, and (iii) creating economic development and jobs in manufacturing, installation, and more.

Local governments can dramatically reduce their carbon footprint by purchasing or directly generating electricity from clean renewable sources.

The most common renewable power technologies include: solar (photovoltaic (PV), solar thermal), wind, biogas (e.g., landfill gas, wastewater treatment digester gas), geothermal, biomass, low-impact hydroelectricity, and emerging technologies such as wave and tidal power.

Local governments can lead by example by generating energy on site, purchasing green power, or purchasing renewable energy. Using a combination of renewable energy options can help to meet local governmen<sup>t</sup> goals, especially in some regions where availability and quality of renewable resources vary. Options for using renewable energy include: generating renewable energy on site, using a system or device at the location where the power is used (e.g., PV panels on a state building, geothermal heat pumps, biomass-fueled combined heat and power), and purchasing renewable energy from an electric utility through a green pricing or green marketing program, where buyers pay a small premium in exchange for electricity generated locally from green power resources.

> **Antonio Colmenar Santos, David Borge Diez, Enrique Rosales Asensio**

> > *Special Issue Editors*

## *Article* **The Practice and Potential of Renewable Energy Localisation: Results from a UK Field Trial**

#### **Peter Boait 1,\*, J. Richard Snape 1, Robin Morris 2, Jo Hamilton 3 and Sarah Darby 4**


Received: 13 December 2018; Accepted: 1 January 2019; Published: 4 January 2019

**Abstract:** The adaptation of electricity demand to match the non-despatchable nature of renewable generation is one of the key challenges of the energy transition. We describe a UK field trial in 48 homes of an approach to this problem aimed at directly matching local supply and demand. This combined a community-based business model with social engagemen<sup>t</sup> and demand response technology employing both thermal and electrical energy storage. A proportion of these homes (14) were equipped with rooftop photovoltaics (PV) amounting to a total of 45 kWp; the business model enabled the remaining 34 homes to consume the electricity exported from the PV-equipped dwellings at a favourably low tariff in the context of a time-of-day tariff scheme. We report on the useful financial return achieved by all participants, their overall experience of the trial, and the proportion of local generation consumed locally. The energy storage devices were controlled, with user oversight, to respond automatically to signals indicating the availability of low cost electricity either from the photovoltaics or the time of day grid tariff. A substantial response was observed in the resulting demand profile from these controls, less so from demand scheduling methods which required regular user configuration. Finally results are reported from a follow-up fully commercial implementation of the concept showing the viability of the business model. We conclude that the sustainability of the transition to renewable energy can be strengthened with a community-oriented approach as demonstrated in the trial that supports users through technological change and improves return on investment by matching local generation and consumption.

**Keywords:** community energy; energy storage; time of use tariff; home battery; demand response; renewable energy; business model

## **1. Introduction**

Community energy initiatives are widely recognised as a valid and useful response to the sustainability challenges of climate change, energy security, and energy affordability. The UK governmen<sup>t</sup> published a Community Energy Strategy in 2014 [1], updated 2015 [2], aimed at encouraging both supply and demand side projects. Municipal and co-operative ownership models providing renewable generation capacity are playing a major role in Germany's "Energiewende" [3], while the USA's 900 rural electricity co-operatives [4], founded as a response to economic depression in the 1930s, are evolving to promote energy efficiency and adopt low carbon generation. Community energy schemes can have many organizational forms based on community of place or interest, but surveys and reviews such as [5–8] identify as typical benefits their ability to engage consumer participation in the systemic changes taking place and a contribution to societal cohesion through shared goals and a fair and transparent allocation of financial costs and returns. Another benefit found by [9] is that community members have a more favourable attitude overall to local renewable energy installations, mitigating the "not in my backyard" attitude that is otherwise common. However these studies also indicate that, at least in the UK, schemes are often financially fragile and depend on enthusiastic volunteers, so need policy and regulatory support to flourish.

An opportunity to strengthen the economic basis for community energy arises from the rising demand for electricity expected as electrification of transport and heating takes place. This prospect was reinforced in the UK by the government's publication of a policy to ban the sale of most petrol and diesel vehicles by 2040 [10], a goal also set by the governmen<sup>t</sup> of France [11]. The UK's electricity system operator, National Grid, predicted in their 2017 Future Energy Scenarios report [12] that with consumer engagemen<sup>t</sup> in end-use energy efficiency and demand response for system efficiency, peak demand by 2050 is limited to 74 GW (from a baseline of 62 GW), but rises to 85 GW without it. So there should be scope for rewarding communities for their contribution to a more efficient solution, which earlier studies such as [13] have valued as worth up to £30 Bn in deferred or avoided network reinforcement costs. Reflecting this potential, the 2018 Future Energy Scenarios report from National Grid [14] includes "Community Renewables" as one of the scenarios that can deliver UK commitments to the Paris Agreement.

The community contribution can be seen as the product of service expectations, activities and technologies within a given community at a given time: a 'demand response space' [15]. There may well be greater opportunities for developing demand response at community scale rather than focusing on individual customers. These opportunities could arise from norms and practices developed through social learning about new technologies and processes [16], from trust-building [17]; and from the diversity of activities, skills and technologies to be found within communities [18].

In this paper we report the results from trialing a combination of business model and "smart home" technology designed for communities of place whose common factor is that they reside on the same segmen<sup>t</sup> of the local electricity distribution network—e.g., they might share the same low voltage (LV) network. An assumption of the model is that there is some distributed low carbon electricity generation on the shared LV network. This can take any of the common forms such as solar photovoltaics (PV), wind generation, combined heat and power (CHP) or micro hydro. The technology and the incentives from the business model are then framed to work synergistically with community activities to empower participants to reduce their cost of electricity through three mechanisms in order of priority:


A benefit of the proposed business model is that it overcomes a legal constraint on financing local generation by forming a community co-operative. UK financial regulation requires that the investing members of a co-operative must either be workers in, or consumers of, the commercial product of the enterprise. This is to avoid the regulatory concessions available to co-operatives being exploited by purely speculative investment offers. Where the whole output of a generator is sold to an electricity supplier through a power purchase agreement, under a recent UK regulatory clarification [19] the investors in the generator cannot be considered consumers. But as this model allows consumers to purchase locally-generated electricity they can form a co-operative to fund the generator, which has advantages over other forms of legal entity such as a Community Interest Company (CIC) in allowing more flexibility in the use of profits.

There have been many trials of time-of-day tariffs and use of technology to influence patterns of residential electricity consumption, as for example summarised in [20,21]. More recently the potential of energy storage to contribute to demand flexibility has been recognized [22]. Drawing on lessons from that experience, the present trial incorporated a comprehensive combination of features and demand response measures not previously tested in the UK. These were:


The goals of the project (called CEGADS) undertaking this trial were to demonstrate the viability of the business model, test the acceptability of this level of innovation to consumers, and evaluate the amount of demand-side response to the measures deployed. The remainder of the paper is structured as follows. In Section 2 we describe the participant community, the business model, and the technology employed. Section 3 provides the results obtained, in respect of the use of local generation, financial outcome, demand side response, and the experience of participants. Section 4 describes briefly an agent-based modelling study of the scheme implemented, and a follow-up project implementing the business model now in fully commercial operation. The overall implications of the findings are discussed in Section 5 followed by conclusions.

#### **2. The CEGADS Trial**

#### *2.1. The Participants and Business Model*

The project name CEGADS stands for Community Electricity Generation, Aggregation, and Demand Shaping, indicating the key features involved. The project is also known by the acronym SWELL, referring to the Energy Local model in the cluster of Oxfordshire villages Shrivenham, Watchfield, and Longcot, in which the trial took place and the 48 participating households were recruited. Many of the residents had already subscribed to the local Westmill energy co-operatives [23,24] operating substantial wind and solar farms, but these generators were subject to wholesale power purchase agreements (entered into prior to the regulatory clarification at [19] summarised above), and being connected at 33 kV were not accessible to the present scheme. However the charitable trust associated with these co-operatives was thereby able to facilitate recruitment for this project from an informed community. Metering of electricity consumption and generation at one-minute intervals was installed in each household, along with the display and control technology. This equipment was installed at no cost to users, as were the batteries described later.

The generation for CEGADS was provided by roof-mounted PV panels already owned by 14 participants with a total capacity of 45 kWp. The export electricity from these panels (i.e., the generated electricity not consumed within the household) was metered and aggregated to form a resource which was considered available for supply at a favourable tariff (£0.065/kWh) to the remaining non-generating participants. The allocation of this export energy aggregate *A* in each half-hour was computed by finding iteratively a "fill level" *L* such that for each of *n* consumers with demand *ei* in the half hour greater than *L*, *L* kWh would be considered supplied from *A*, and for those

remaining *m* consumers with demand *ej* less than *L*, their demand would be fully met from *A*, with *L* also satisfying:

$$A = nL + \sum\_{j=1}^{j-m} c\_j \tag{1}$$

where *A* was large enough to more than supply all non-generating consumers then the residue was considered community export. This method of fair allocation of local generation is a key feature of the business model, which also allows the community export to be sold to an electricity supplier under a power purchase agreement. To enable participants in the trial to retain their existing electricity supplier and tariffs, the time-of-use tariff applied to electricity not generated locally was implemented as an incentive scheme where the difference between the actual cost of electricity to participants and the cost they would have incurred under the trial tariff is given to them in the form of credit vouchers exchangeable for goods at a supermarket chain associated with the electricity supplier supporting the project. The commercially-realistic (for 2016) time of use tariff rates offered are shown in Figure 1. Generators were credited with £0.065 for each export kWh matched with consumption, and £0.055 for each kWh not matched so considered as taken up by a power purchase agreemen<sup>t</sup> (PPA).

**Figure 1.** Time-of-use electricity tariff rates.

#### *2.2. The Metering and Demand Response System*

To execute the metering required for this scheme and enable the participants to make best use of the local generation and time-of-use tariffs a smart metering and control unit was installed in each household. Branded "Hestia", this unit provided from an internal web server a display of the tariff rates on any convenient IT device connected to the household broadband, but modified with a dip in the displayed rates during the middle of the day that reflected approximately the amount of local PV generation predicted to be available based on the overnight local weather forecast. It also provided displays of electricity consumption and generation over the last 24 h for the household, the participant community as a whole, the aggregate PV generation, and the PV generation for the household for those so equipped. To provide the data for these displays, metering data at one minute intervals was collected and processed in a central database using a commercial cloud service. A simplified view of the system is shown in Figure 2.

The Hestia control unit also performed automatic demand response for controllable appliances as illustrated in Figure 2. Six of the participating dwellings had space heating provided by electrically-heated thermal storage heaters and hot water from an immersion-heated tank. In aggregate these appliances provided about 60 kWh of thermal storage in each of the six homes. Charging of these useful thermal energy stores was controlled such that user comfort requirements as expressed on the Hestia user interface were prioritized, but was otherwise optimized against a tariff-dependent signal from the database server that ensured cost effective use of local generation and the time-of-day tariff while preventing peaks in aggregate demand at tariff boundaries by randomizing dispatch of loads. This signaling and optimization methodology is fully described in [25,26] and the peaking risk that is mitigated, which has been identified in many simulation studies, is identified for example in [27,28].

**Figure 2.** CEGADS system diagram.

An example page from the Hestia Hub display is shown in Figure 3a. All of the participants were given a smart plug as shown in Figure 3b for which the on/off status could be radio-controlled via a user interface provided by the Hestia unit. This allowed users to set a time window within which an appliance powered via the smart plug should operate, and the required operating duration. If some scheduling flexibility was available from the difference between the time window and the operating duration, the Hestia selected an optimized dispatch time using the demand response signal.

Nine of the participant households were equipped with a 2 kWh home battery unit (Figure 3c). The nine were deliberately chosen to exclude households with PV panels or thermal storage. These lithium-ion batteries were controlled to charge during low tariff periods and discharge during the early evening high tariff rate period, with the objective of improving the benefit these households obtained from the tariff scheme.

To summarise, the trial involved expanding the demand response potential in three villages by extending people's ideas of what community energy could do for them, by developing new activity in relation to the cooperative business model and tariff, and introducing technology in the shape of the Hestia control units, database server, smart plugs, batteries and display capability. In doing this, it was building on trust and knowledge that had been established through everyday social interactions and (for some) involvement in a local energy cooperative.

**Figure 3.** (**a**) Hestia Hub display. (**b**) Smart plug. (**c**) Home battery and heating control units.

#### **3. Results from the Trial**

#### *3.1. Utilisation of Local Generation*

Over the year of trial operation, out of the total PV generation of 43,406 kWh from the 14 generators, 18,307 kWh were used within the generating households, and 25,908 kWh were available to share with other participants. Of this available total, 22,154 kWh were matched with consumption using the algorithm described earlier, and the balance of 2944 kWh was allocated to the PPA. Figure 4 illustrates this outcome on a monthly basis. The generation tariffs gave an improved financial return simulated through the credit vouchers of about 80% to generators (£719 in addition to £868 from 50% deemed export feed-in tariff at £0.040/kWh making a total of £1587). The generation matched with consumption represented about 9.5% of the total electricity consumed (c. 233 MWh) by all the participants during the year.

**Figure 4.** Allocation of PV generation by month.

The savings with respect to their existing tariffs that accrued to all participants are shown in Figure 5. The "tariff" plot shows the savings from the baseline time-of-day tariff. "Local use" shows the savings to participants who consumed the matched generation shown in Figure 4. "PV shared" shows the additional return to the PV generators.

**Figure 5.** Financial savings to participants from the project.

#### *3.2. Demand Side Response*

The overall demand side response is illustrated in Figure 6a–d by comparing aggregate demand profiles at the start and end of the trial. The response derived from four sources:


The comparison between average consumption profiles in Figure 6 is influenced by the fact that December 2016 was much colder (295 degree-days) than December 2015 (154 degree-days). The six electrically-heated homes presented a special case in that they were already using a time-dependent tariff known as Economy 7. This comprises a low rate for 7 h overnight of about £0.07/kWh and a higher day rate of about £0.016 typically used, as in the present case, with thermal storage heating and domestic hot water tanks that can be charged at the low rate. So the automatic controls were given a signal which moved some of the heating demand into the middle of the day to take advantage of the local generation and lower mid-day tariff. The controls also ensured a more precise matching of stored thermal energy to the weather-dependent heating demand.

**Figure 6.** Comparison of average daily aggregate electricity use profile in December 2015 with December 2016, for different participant groups: (**a**) All participants (**b**) Participants with electric heating (**c**) Participants with batteries (**d**) All participants excluding those with batteries, electric heating and the pub.

The resulting shift in distribution of demand is illustrated in Figure 6b which shows the increased heating demand during the day and also a reduction of 16% in consumption during the peak tariff hours despite the colder weather in 2016. A comprehensive report focused on the performance of the heating controls is provided in [25].

The operation of eight batteries (one had to be removed before December 16) can be seen in Figure 6c with charging commencing at the start of the low tariff rate at 23:00 and continuing overnight. The reduction in demand by 20% during the evening high tariff period is also evident. It was found desirable to configure the batteries with a maximum discharge rate of 0.25 kW to ensure that the battery output always offset local consumption and was not exported, for which no reward could be offered. One of the participants with a battery also acquired an electric car during the trial year which contributed to the increased overnight demand seen in Figure 6c.

The relatively limited aggregate impact of the smart plugs and manual time-shifting can be seen in Figure 6d, which excludes the participants shown in Figure 6b,c and also the village pub (i.e., bar) which had a significant increase in power consumption during the year for commercial reasons unrelated to this trial. The participants in 6d had gas central heating so consumption was not greatly affected by the weather difference in the comparison. The increased overnight and mid-day consumption can be seen and also a slight reduction during the peak tariff time. The peak before 06:00 is believed to be wet appliance operation scheduled at the end of the low tariff period.

To examine the range of individual household responses, the average demand in each of the six tariff periods shown in Figure 1 was calculated for each household for October–December 2015 and for the same period in 2016. The correlation between changes in demand in each tariff period, and the tariff rate was then tested, with the hypothesis that demand would have changed over the year in inverse proportion to the tariff as consumers became accustomed to a time-of-day tariff and adjusted their demand accordingly. The results for different participant groups are shown in Table 1. A participant was counted as a "responder" in the table if a negative correlation was observed between change in demand and tariff rate with an R<sup>2</sup> value greater than 0.1. The much greater proportion of responders among participants with some additional technology that reinforces their engagemen<sup>t</sup> is evident. Note all the groups in Table 1 are independent i.e., there is no overlap of membership.


**Table 1.** Correlation of change in demand over a year with time-of-day tariff.

## *3.3. User Experience*

Most participants in the CEGADS project were interviewed in three rounds of surveys; (by telephone, and online using Survey Monkey) around the start in late 2015 early 2016, during summer 2016, and spring 2017. They began with largely positive attitudes, with responses to questions concerning demand response and time of use tariffs as shown in Figure 7a,b. In round 1 of the interviews, most participants were positive about being able to switch the time of use of some of their electricity although the extent to which they could was determined by patterns of household occupation (i.e., if they were in the home during the daytime). Most participants felt able to switch devices such as white goods such as washing machines, tumble driers, and dishwashers and battery charging (e.g., for computers) and where time and space flexibility permitted (e.g., either being able to use them, or program them to run, in off peak hours).

**Figure 7.** Initial perceptions of (**a**) demand response possibility and (**b**) a time-of-use tariff.

By the second round of interviews practices relating to the Hestia Hub technology provided were evident, with a range of usage levels as shown in Figure 8, and comments such as "*when I first started out I was looking at it daily or more frequently, but now* ... *I'm more familiar with it*". Overall, 13 of the 37 (35%) respondents reported looking at their Hestia Hub at least once a week. This is roughly consistent with findings relating to in-home display usage by smart-metered customers in Great Britain as a whole, where 44% of householders reported that they were consulting their display at least once a week, between seven and 29 months after installation [20].

However, on the broad question as to whether the project had influenced day to day habits of electricity use, 31 (out of 39) said "*yes*". Of those who responded "*yes*", many reported shifting of activities, such as using the washing machine and dish washer at different times, or changing cooking practices, for example: "*sometimes opting to microwave or grill instead of using oven*". Of the 6 who responded no, some were already producing electricity through solar PV, some considered they had already made changes in their consumption, and some had family routines which they didn't want to shift.

**Figure 8.** Engagement with smart metering and control device after six months.

The final survey revealed many practical changes in household practices that participants had taken, such as "*Weather—I was advised if I could use my washing machine when the sun was at its highest, that was the best time to do it*"; and "*Put the electric towel rail on a timer plug socket. It only operates sporadically in the peaks*". The motivation for demand response decisions was drawn from a wide range of factors as indicated in Figure 9, showing that the different channels used by the project to communicate with participants all had a role in the results obtained. The variety of integrated approaches enabled participants to learn and incorporate their learning into new routines in different ways, through different means, and at different times, for example: "*when you have the reports that came through and here you have the cheque for your money and your report on your energy use, you could practically see how it all fitted together*". Of 21 participants who said they had shifted usage during the trial, 17 said they would continue to do so, while 16 respondents who had reduced the amount of electricity usage during the trial were planning to maintain their reduction.
