The Case of the Greek Tilos Island

Home , Energy in Greece, Gyali, Kalymnos, Tilos

Article Supporting the Clean Electriﬁcation for Remote Islands: The Case of the Greek Tilos Island

John K. Kaldellis

Lab of Soft Energy Applications & Environmental Protection, University of West Attica, 12241 Athens, Greece; [email protected]

Abstract: Many islands around the world present considerable energy supply problems, while their energy mixture is controlled by oil products. Meanwhile, several of these isolated islands enjoy excellent RES potential that support actions to maximize RES integration. In order to ameliorate energy supply security and energy autonomy of the Aegean Archipelagos Greek islands, an integrated solution is deployed based on the exploitation of the existing RES potential in conjunction with the application of an appropriate energy storage scheme, as well as complementary smart-grid elements. The proposed solution has been applied for the Greek Tilos island in the framework of the Tilos- Horizon 2020 program. In this context, the implementation of the integrated Tilos energy solution under the current local legislative frame is a great success story introducing several important innovative characteristics in the European market, like the combined operation of a wind turbine and a PV installation, the application of new technology battery energy storage, the installation of a DSM network/platform and the development of a large number of reliable forecasting algorithms. The innovative integrated solution is a real-world working operating example offering knowledge and proving that the solution deployed could be equally well applied in various other remote islands throughout the European territory with very promising results.

Citation: Kaldellis, J.K. Supporting Keywords: hybrid power station; green island; energy storage; remote community the Clean Electrification for Remote Islands: The Case of the Greek Tilos Island. Energies 2021, 14, 1336. https://doi.org/10.3390/en14051336 1. Introduction 1.1. The Current Situation of Greek Remote Islands Electricity System Academic Editor: George Many islands around the world present considerable energy supply problems, while S. Stavrakakis their energy mixture is controlled by oil products [1,2]. Actually, energy supply security is related to both the dependence on oil imports as well as to the fact that the majority of Received: 2 November 2020 these islands are subject to even greater challenges in the occasion of critical damage to Accepted: 18 February 2021 Published: 1 March 2021 local thermal power generation stations or a power network failure [3]. In this context, improving energy security would dictate to not only reduce the energy dependence on imported fuels [4], but also to establish a diversified energy mix [5], considering the Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in increasing contribution of high percentages of Renewable Energy Sources (RES) power published maps and institutional affil- generation. Unfortunately, introducing high RES percentages normally entails a price, iations. since the risk that has to be addressed by the local operator of isolated weak electrical grids, either through the employment of adequate reserve capacity [6] or the implementation of RES curtailments [7,8], is proportional to the relevant RES contribution in such grids. To face this challenge, local grid operators impose certain constraints that are related to the strongly variable penetration of RES and the compliance with the technical limitations Copyright: © 2021 by the author. (technical minima) that characterize the currently operating oil-based power generation Licensee MDPI, Basel, Switzerland. This article is an open access article units [9,10]. The result of all these restrictions normally limits the maximum contribution distributed under the terms and of RES concerning the local load fulfillment, in the range of 15–20% on an annual basis conditions of the Creative Commons (Figure1). More specifically, according to the data of Figure1, the RES production during Attribution (CC BY) license (https:// the last years is presented on a monthly basis (taking values from 60 GWhe up to 150 GWhe creativecommons.org/licenses/by/ per month), while the RES percentage (penetration) on monthly electricity demand varies 4.0/). between 12% and 22%.

Energies 2021, 14, 1336. https://doi.org/10.3390/en14051336 https://www.mdpi.com/journal/energies Energies 2021, 14, x FOR PEER REVIEW 2 of 24

Energies 2021, 14, 1336 2 of 22 per month), while the RES percentage (penetration) on monthly electricity demand varies between 12% and 22%.

FigureFigure 1. The 1. The Renewable Renewable Energy Energy Sources Sources (RES) (RES contribution) contribution on the on Non-Interconnected the Non-Interconnected Islands’ Islands’ (NIIs) (NIIs) electricity electricity consumption. consumption. Actually, during the last fifty years, the Non-Interconnected Islands’ (NIIs) electricity requirementsActually, have during been the almost last fifty exclusively years, the Non covered-Interconnected by autonomous Islands’ power (NIIs) generation electricity sta- tionsrequirements (APS) comprised have been of almost internal exclusively combustion covered engines by autonomous and gas turbines. power generation More specifically, sta- thetions electricity (APS) comprised demand of of internal the aforementioned combustion engines islands and is gas mainly turbines. covered More by specifically, almost thirty (30)the APSs electricity [4,11 ],demand based onof permanentthe aforementioned as well asislands portable is mainly units ofcovered a broad by nominal almost thirty capacity range.(30) APSs The [4,11] total, capacitybased on ofpermanent all these as APSs well isas approximately portable units of 700 a broad MW nominal (excluding capacity the elec- range. The total capacity of all these APSs is approximately 700 MW (excluding the electrical systems of Crete and Rhodes islands, including, however, the ones of the Cyclades’ trical systems of Crete and Rhodes islands, including, however, the ones of the Cyclades’ complex just recently interconnected to the Greek mainland electricity grid). Moreover, the complex just recently interconnected to the Greek mainland electricity grid). Moreover, significant increase encountered in the local electricity demand of NIIs during the tourism the significant increase encountered in the local electricity demand of NIIs during the season results in a limited utilization of thermal power units during the rest of the year, tourism season results in a limited utilization of thermal power units during the rest of particularlythe year, particularly if the latter if the are latter oversized are oversized with regards with regards to the average to the average autonomous autonomous network electricitynetwork requirements.electricity requirements. This is exactly This is the exactly case for the the case very for small the very to medium small to scale medium Aegean Archipelagosscale Aegean islands.Archipelagos islands. TheThe direct direct result result of of this this situation situation is is the the low low capacity capacity factors factors for for all all these these APS, APS, combined com- alsobined with also excessive with excessive specific specific fuel consumption fuel consumption (SFC) values,(SFC) values, in the rangein the ofrange200–300 of 20 gr/kWh0–300 e, which,gr/kWh ine, turn, which results, in turn in, quiteresults elevated in quite elevated electricity electricity generation generation cost for cost most for of most the localof the elec- tricallocal systems. electrical Insyste thisms. context, In this thecontext, combined the combined effect of effect high of SFC high values SFC values in conjunction in conjunc- with extensivetion with maintenance extensive maintenance requirements requirements and APS assets’and APS amortization assets’ amortization results in results a continuous in a increasecontinuous of the increase total electricity of the total generation electricity cost.generation More specifically,cost. More specifically, during the during last years, the the annuallast ye electricityars, the annual generation electricity cost generation of all the cost NIIs of has allexceeded the NIIs has 300 exceed M€ (electricaled 300 M systems€ (elec- of Cretetrical and systems Rhodes of Crete not included) and Rhodes and not is included) strongly affectedand is strongly by the importedaffected by oil the quantities imported and theiroil quantities price volatility. and their price volatility.

1.2. RES Power Generation in NIIs of the Aegean Archipelagos According to long-term measurements, most of these isolated islands possess excellent RES potential that supports actions to maximize RES integration. Albeit the very high- quality wind potential that characterizes most of the NIIs as well as the excellent solar potential of the entire Aegean Archipelagos (Figure2), the progress noted in RES power

applications has not met the expectations. Based on available ofﬁcial data [12,13], wind power has presented a stagnation at ~75 MW, distributed on approximately 100 wind parks, dispersed mainly across the biggest and medium-scale Aegean Archipelagos islands, Energies 2021, 14, x FOR PEER REVIEW 3 of 24 Energies 2021, 14, x FOR PEER REVIEW 3 of 24

1.2. RES Power Generation in NIIs of the Aegean Archipelagos 1.2. RES Power Generation in NIIs of the Aegean Archipelagos According to long-term measurements, most of these isolated islands possess excel- According to long-term measurements, most of these isolated islands possess excellent RES potential that supports actions to maximize RES integration. Albeit the very high- lent RES potential that supports actions to maximize RES integration. Albeit the very high- quality wind potential that characterizes most of the NIIs as well as the excellent solar quality wind potential that characterizes most of the NIIs as well as the excellent solar Energies 2021, 14, 1336 potential of the entire Aegean Archipelagos (Figure 2), the progress noted in RES power3 of 22 potential of the entire Aegean Archipelagos (Figure 2), the progress noted in RES power applications has not met the expectations. Based on available official data [12,13], wind applications has not met the expectations. Based on available official data [12,13], wind power has presented a stagnation at ~75 MW, distributed on approximately 100 wind power has presented a stagnation at ~75 MW, distributed on approximately 100 wind parks, dispersed mainly across the biggest and medium-scale Aegean Archipelagos is- parks, dispersed mainly across the biggest and medium-scale Aegean Archipelagos is- whilelands, while PV capacity PV capacity slightly slightly exceeds exceeds 40 MW40 MpW, alsop, also gathered gathered in in thethe biggestbiggest islands islands of of the lands, while PV capacity slightly exceeds 40 MWp, also gathered in the biggest islands of aforementionedthe aforementioned area area (see (see also also Figure Figure3). 3). the aforementioned area (see also Figure 3).

WIND & SOLAR ENERGY IN GREECE WIND & SOLAR ENERGY IN GREECE

V = 3.2 m/s V = 3.2 m/s F G F G V = 5.3 m/s V = 5.3 m/s E Annual mean wind E Annualspeed mean at wind30m F speed at 30m F D D E E C C D D B B F C F C V = 7.5 m/s D V = 7.5 m/s C E D C E

Solar potential Solar potential 2 A ZONE kWh/m2 A ZONEA kWh/m>1650 A B 1600>1650 – 1649 V = 5.3 m/s V = 5.3 m/s B C 16001550 – 1649 – 1599 B B C D 15501500 – 1599 – 1549 D E 15001450 – 1549 – 1499 C E 1450 – 1499 C F 1400 – 1449 A F G 1400 –<1399 1449 A G <1399

Figure 2. Wind & Solar Potential in Aegean Sea. FigureFigure 2. Wind 2. Wind & Solar & Solar Potential Potential in Aegean in Aegean Sea. Sea.

(a) (a) Figure 3. Cont.

Energies 2021, 14, 1336 4 of 22 Energies 2021, 14, x FOR PEER REVIEW 4 of 24

(b)

FigureFigure 3. 3.(a()a Installed) Installed wind wind power power capacity; capacity; (b) ( bPV) PV power power capacity capacity in NIIs in NIIs of the of the Aegean Aegean Archipel- Archipelagos. agos. Besides, stagnation encountered in new RES power plants implementation is further reflectedBesides, in stagnation Figure1, where encountered one may in new find RES the RESpower contribution plants implementation in the local is load further demand reflectedcoverage in forFigure all NIIs,1, where on aone monthly may find basis. the RES From contribution the data presented, in the local it load can demand be concluded coveragethat RES for share all NIIs, presents on a monthly an almost basis. constant From the pattern, data presented despite the, it smallcan be increaseconcluded starting thatfrom RES 15% share in 2011presents to attain an almost 18% inconstant 2017. Aspattern, already despite mentioned, the small this increase specific starting penetration fromupper 15% limit in 2011 is the to attain outcome 18%of in local2017. electricityAs already grids’mentioned, restrictions this specific (thermal penetration power units’ upper limit is the outcome of local electricity grids’ restrictions (thermal power units’ tech- technical minima and local grid dynamic penetration margins) that discourage new RES nical minima and local grid dynamic penetration margins) that discourage new RES power stations’ implementation, due to the expected high values of curtailments. Such power stations’ implementation, due to the expected high values of curtailments. Such curtailments for already operating wind parks are already apparent in saturated insular curtailments for already operating wind parks are already apparent in saturated insular electrical networks, such as the relatively big one configured by Kos and Kalymnos islands, electrical networks, such as the relatively big one configured by Kos and Kalymnos is- to where Tilos island’s electricity grid is connected. In the abovementioned electrical lands, to where Tilos island’s electricity grid is connected. In the abovementioned electri- calnetwork, network, the the curtailments curtailments that that local local wind wind parks parks(with (with anan installedinstalled capacitycapacity approximatelyapproxi- matelyequal equal to 15 MW)to 15 M faceW) face could could even even approach approach the the value value of of 30% 30% of of the the pertinent pertinent annual wind windenergy energy yield yield [13]. [13]. Consequently, Consequently, the the imposed imposed grid grid and and thermal thermal power power units’ units’ constraints con- straintschallenge challenge the effective the effective capacity capacity factor factor of the of the existing existing wind wind parks, parks, strongly strongly hinderinghinder- the ingmaximum the maximum exploitation exploitation of the of available the available wind wind potential potential [7– [710–].10]. ToTo improve improve the the limited limited share share of RES of RES power power generation generation in all inthese all theseremote remote islands, islands, pilotpilot projects projects being being characterized characterized by state by state-of-the-art,-of-the-art, integrated integrated solutions solutions which which combine combine intelligentintelligent and and sophisticated sophisticated management management features features (forecasting (forecasting and DSM) and DSM) as also as energy also energy storage,storage, could could pave pave the the way. way. To Tothis this end, end, the pioneering the pioneering TILOS TILOS Horizon Horizon 2020 project 2020 project [14], [14], deployeddeployed on on the the Greek Greek Tilos Tilos island island that that belongs belongs as already as already mentioned mentioned to the tosaturated the saturated—— withwith regards regards to toRES RES applications applications—Kos—Kos and and Kalymnos Kalymnos electrical electrical network, network, indicates indicates the the wayway community community-scale-scale battery battery storage storage can canoptimally optimally cooperate cooperate with withlocally locally developed developed RES RES powerpower generation generation and and advanced advanced techniques techniques such such as DSM. as DSM.

2.2. Proposed Proposed Integrated Integrated Electricity Electricity Solution Solution for Remote for Remote Islands Islands ToTo improve improve energy energy autonomy autonomy and andenergy energy supply supply security security of the several of the existing several is- existing landsislands,, an integrated an integrated solution solution has been has been developed developed (Figure (Figure 4) based4) based on the on available the available wind wind andand solar solar potential potential exploitation exploitation along along with with an appropriate an appropriate energy energy storage storage infrastructure, infrastructure, introducingintroducing also also some some extra extra smart smart-grid-grid elements. elements. This solution This solution is quite promising is quite promising for var- for iousvarious Greek Greek islands islands belonging belonging to the Aegean to the AegeanArchipelagos Archipelagos area, and area,particularly and particularly the most the distantmost distantand small and-scale small-scale ones. ones.

Energies 2021, 14, 1336 5 of 22 Energies 2021, 14, x FOR PEER REVIEW 5 of 24

Figure 4. 4. TheThe RES RES-based-based energy energy solution solution for for NIIs. NIIs. The The case case of T ofilos Tilos island. island.

MoreMore specifically, specifically, for for every every island island,, one one should should estimate estimate the theavailable available RES RES (i.e., (i.e., wind/solar/biomass and and geothermal) geothermal) potential. potential In this. In this context, context, real- real-worldworld operational operational data are are necessary necessary (see (see,, for for example example,, Figures Figures 5 and5 and 6),6 originated), originated—if—if possi possible—fromble—from long long-- term in in situ situ measurements. measurements. Actually, Actually, for for a period a period of three of three years years (201 (2015–2017),5–2017), detailed detailed measurements were were carried carried out out concerning concerning the the wind wind speed speed and and the the solar solar irradiance irradiance data data as wellas well as theas the main main meteorological meteorological parameters parameters by by our our research research team team (Soft (Soft Energy Energy Applications Appli- andcations Environmental and Environmental Protection Protection Lab ofLab UNIWA) of UNIWA) [14]. [14]. To thisTo this end, end, the the expected expected RES-based RES- based annual electricity generation “Etot” can be estimated by defining the maximum rated annual electricity generation “Etot” can be estimated by deﬁning the maximum rated power “PRES” for several configurations consisted of wind “Pw” and photovoltaic parks power “PRES” for several conﬁgurations consisted of wind “Pw” and photovoltaic parks “PPV”, i.e.: “PPV”, i.e.,: (1) EtotE =tot 8760.(CF= 8760.(CFw.Pw w+ .PCFwPV+.P CFPV)PV .PPV) (1) Energies 2021, 14, x FOR PEER REVIEWwhere “CFw” and “CFPV” are the capacity factor values of the selected wind turbines6 ofand 24 where “CFw” and “CFPV” are the capacity factor values of the selected wind turbines and PV panels according according to to the the available available wind wind and and solar solar potential potential,, correspondingly. correspondingly. On the basis of existing licensing procedure in Greece, the maximum RES power to be installed on an autonomous island grid cannot exceed the local consumption peak load demand (e.g., 960 kW), hence:

Pres = Pw + PPV ≤ Max Load Demand (2) Considering that the rated power of the smallest contemporary commercial wind turbine in the market is 900 kW, a similar machine has been selected and has been down- rated at 800 kW, in order to install also a small PV power station of 160 kWp. In the near future, additional PV panels may be added if the peak power of the island increases also, since the PV installations have the option of modularity.

Figure 5. 5. WindWind s speedpeed time time series series for for Tilos Tilos island. island.

Figure 6. Solar irradiance time series at horizontal plane for Tilos island.

Subsequently, using accurate measurements (Figure 7), even on a sec-time-interval basis, stemming from official recordings, the pertinent load demand time-evolution is in- vestigated to ensure the local electricity grid stability as well as the uninterruptible power supply of high priority loads. Furthermore, both deferrable loads as well as loads of secondary priority have been pinpointed (Figure 8). Actually, in Figure 8, one may find the electricity consumption needs for water pumping (including also water transferring between water tanks), which is maximum during the summer due to the increased tourist activities. In the same Figure 8, one may find the monthly public lighting electricity consumption for the four villages of the island (Megalo Chorio, Livadia, Agios Antonios, Er- istos). As can be revealed from the existing data, the relevant electricity consumption is also increased during the summer period since all the local tourist facilities are in operation. Finally, the technical specifications of the “in-operation” thermal power units and the actual interconnections deployed until now to satisfy the local load demand have been scrutinized as well.

Energies 2021, 14, x FOR PEER REVIEW 6 of 24

Energies 2021, 14, 1336 6 of 22

Figure 5. Wind speed time series for Tilos island.

Figure 6. 6. SolarSolar irradiance irradiance time time series series at athorizontal horizontal plane plane for forTilos Tilos island. island.

Subsequently,On the basis of using existing accurate licensing measurements procedure (Figure in Greece, 7), even the maximum on a sec-time RES-interval power to be installedbasis, stemming on an autonomousfrom official recordings, island grid the cannot pertinent exceed load thedemand local time consumption-evolution peakis in- load demandvestigated (e.g., to ensure 960 kW), the local hence: electricity grid stability as well as the uninterruptible power supply of high priority loads. Furthermore, both deferrable loads as well as loads of secondary priority have been pinpointedPres = Pw + ( PFigurePV ≤ Max8). Actually, Load Demand in Figure 8, one may find the (2) electricity consumption needs for water pumping (including also water transferring be- tweenConsidering water tanks), that which the is rated maximum power during of the the smallest summer contemporary due to the increased commercial tourist wind turbineactivities. in In the the market same Figure is 900 8, kW, one a may similar find machine the monthly has beenpublic selected lighting and electricity has been con- down- ratedsumption at 800 for kW,the four in order villages to installof the island also a (Megalo small PV Chorio, power Livadia, station Agios of 160 Antonios, kWp. In theEr- near future,istos). As additional can be revealed PV panels from may the beexisti addedng data, if the the peak relevant power electricity of the island consumption increases is also, sincealso increased the PV installations during the summer have the period option since of modularity. all the local tourist facilities are in operation. Subsequently,Finally, the technical using specifications accurate measurements of the “in-operation (Figure7”), thermal even on power a sec-time-interval units and basis,the actual stemming interconnection from officials deployed recordings, until now the to satisfy pertinent the local load load demand demand time-evolution have been is investigatedscrutinized as to well. ensure the local electricity grid stability as well as the uninterruptible power supply of high priority loads. Furthermore, both deferrable loads as well as loads of secondary priority have been pinpointed (Figure8). Actually, in Figure8, one may find the electricity consumption needs for water pumping (including also water transferring between water tanks), which is maximum during the summer due to the increased tourist activities. In the same Figure8, one may find the monthly public lighting electricity consumption for the four villages of the island (Megalo Chorio, Livadia, Agios Antonios, Eristos). As can be revealed from the existing data, the relevant electricity consumption is also increased during the summer period since all the local tourist facilities are in operation. Finally, the technical specifications of the “in-operation” thermal power units and the actual interconnections deployed until now to satisfy the local load demand have been scrutinized as well. Energies 2021, 14, 1336 7 of 22 Energies 2021, 14, x FOR PEER REVIEW 7 of 24 Energies 2021, 14, x FOR PEER REVIEW 7 of 24

Figure 7. Load demand data for Tilos island. The 4.5 year average 10 min load. Figure 7. 7. LoadLoad d demandemand data data for for Tilos Tilos island. island. The The 4.5 4.5year year average average 10 min 10 minload. load.

Figure 8. Public lighting-water pumping electricity consumption breakdown. Figure 8. 8. PublicPublic lighting lighting-water-water pumping pumping electricity electricity consumption consumption breakdown. breakdown. Consequently, a thoroughly designed and sized energy storage system has been in- Consequently, a athoroughly thoroughly designed designed and andsized sized energy energy storage storage system systemhas been has in- been stalled [15–18], to improve the electricity system power (active and reactive) balance and stalled [15–18], to improve the electricity system power (active and reactive) balance and installedreduce the [ 15dependence–18], to improve on fossil the fuels. electricity The corresponding system power energy (active storage and technology reactive) and balance reduce the dependence on fossil fuels. The corresponding energy storage technology and andits main reduce operational the dependence parameters onfossil (i.e., fuels.input– Theoutput corresponding power, charging/discharging energy storage technology rate, its main operational parameters (i.e., input–output power, charging/discharging rate, andround its-trip main efficiency, operational depth parameters of discharge, (i.e., energy input–output capacity, power,etc.) are charging/discharging determined by the is- rate, round-trip efficiency, depth of discharge, energy capacity, etc.) are determined by the is- round-tripland’s peak efﬁciency,load demand, depth its inherent of discharge, characteristics energy (e.g. capacity,, land etc.)availability, are determined topography, by the land’s peak load demand, its inherent characteristics (e.g., land availability, topography, island’setc.) as well peak as loadthe degree demand, of energy its inherent autonomy characteristics to be attained (e.g., (or land the availability,intended maximum topography, etc.) as well as the degree of energy autonomy to be attained (or the intended maximum etc.)oil contribution as well as thein the degree island of energy energy mix). autonomy to be attained (or the intended maximum oil contribution in the island energy mix). oil contributionThe introduction in the of island DSM techniques energy mix). in conjunction with smart meters for the major The introduction of DSM techniques in conjunction with smart meters for the major deferrableThe introduction loads stands out of DSM as another techniques crucial in parameter conjunction for obtaining with smart enhanced meters electrical for the major deferrable loads stands out as another crucial parameter for obtaining enhanced electrical deferrable loads stands out as another crucial parameter for obtaining enhanced electrical grid power balance [13,19]. Moreover, load management may also be applied in the residential and hotel sectors as well as for other principal consumers of the island (Figure9) .

Energies 2021, 14, x FOR PEER REVIEW 8 of 24

Energies 2021, 14, 1336 8 of 22

grid power balance [13,19]. Moreover, load management may also be applied in the residential and hotel sectors as well as for other principal consumers of the island (Figure 9). One of the most promising cases is the application of load management strategies at the One of the most promising cases is the application of load management strategies at the existingexisting eight eight (8) (8) water water pumpingpumping installationsinstallations with maximum maximum power power of of 7 700 k kW,W, represent- representing aing considerable a considerable part part (e.g., (e.g. 15%), 15%) of the of island’sthe island’s average average load. load. In In this this context, context the, the local local load demandload demand can be can adjusted be adjusted to better to better match match the available the available RES productionRES production without without jeopardizing jeop- theardizing local population the local population living standards. living standards. Additionally, Additionally, penetration penetration of electric of electric vehicles vehi- in the transportationcles in the transportation sector of the sector island of the (Figure island 10 (Figure) could 10) also could assist also in assist this in direction, this direction, with the local,with relativelythe local, relatively restricted restricted (maximum (maximum distance distance 15 km), 1 road5 km), networks road networks suggesting suggesting the ideal groundthe ideal for ground electro-mobility for electro- applications.mobility applications.

(a) (b)

FigureFigure 9. 9.DSM DSM panels—overall panels—overall andand atat villagevillage centre @ @ Livadia Livadia (a (,ab,b) and) and @ @M. M. Chorio Chorio (c,d (c)., d).

Furthermore, forecasting methods and associated systems can also be taken into account under the scope of upgrading the operation and ameliorating the flexibility of the complete electricity production system (EPS), while in parallel maximizing RES contribution. Such an example is the deployment of an innovative Forecasting Platform (FP) that is capable of providing reliable predictions of load demand, solar power and wind power generation several hours ahead [20,21]. The aforementioned features comprise the core of a High-Level Energy Management System (EMS) and Centre, in charge of coordinating the programming and optimum operation of the various subsystems of such an integrated solution. In this context, based on forecasting signals, the operation and dispatch strategy of energy storage can be improved, while also keeping end users informed on the potential for demand response, hence favoring RES contribution maximization and cost-effectiveness.

Figure 10. Photos of the charging area and EV charger at TILOS info-kiosk in Livadia. Figure 10. Photos of the charging area and EV charger at TILOS info-kiosk in Livadia. Furthermore, forecasting methods and associated systems can also be taken into ac- countAs undera final point, the scope it is important of upgrading also theto mention operation that and all the ameliorating aforementioned the ﬂexibility steps im- of the posecomplete continuous electricity education production and interaction system (EPS), with the while local in community parallel maximizing [22]. To that RES end, contribu- close collaboration with the local authorities should be established, respecting their priorities, while at the same time the potential project developer should disseminate to the local people the information for the new infrastructure introduction and the benefits anticipated.

3. The Application Paradigm of Tilos Island Tilos is a unique, small-scale “S”-shaped island found at the SE side of the Aegean Sea in Greece, a region of great symbolism and major geopolitical interest for Europe as a whole. Part of the Dodecanese group of islands, Tilos lies mid-sea between Kos and Rhodes (Figure 11). Northwest to south-east, the island is ~14.5 km long, with a maximum stretch of 8 km and an overall area of approximately 64 km2. Standing at a distance of 240 n. miles from the Greek mainland (Piraeus, Attica), Tilos belongs to the very special group of remote and small-scale European islands. The island counts four main (populated) settlements/villages, namely, Megalo Cho- rio, which is the capital; Livadia, which is the biggest village and where the island’s port and main center of economic activities are found; Agios Antonios and Eristos. The climate of Tilos is typical of the Mediterranean region, with mild winters and hot, sunny summers. Concerning the local RES potential, Tilos appreciates an excellent- quality solar potential, determined by ~1800 kWh/(m2.a) at the horizontal plane (Figure 6). On the other hand, the local wind potential is of medium quality, with the average wind speed ranging between 6 and 8 m/sec in the largest part of the island (see also Figure 5).

Energies 2021, 14, 1336 9 of 22

tion. Such an example is the deployment of an innovative Forecasting Platform (FP) that is capable of providing reliable predictions of load demand, solar power and wind power generation several hours ahead [20,21]. The aforementioned features comprise the core of a High-Level Energy Management System (EMS) and Centre, in charge of coordinating the programming and optimum operation of the various subsystems of such an integrated solution. In this context, based on forecasting signals, the operation and dispatch strategy of energy storage can be improved, while also keeping end users informed on the potential for demand response, hence favoring RES contribution maximization and cost-effectiveness. As a ﬁnal point, it is important also to mention that all the aforementioned steps impose continuous education and interaction with the local community [22]. To that end, close collaboration with the local authorities should be established, respecting their priorities, while at the same time the potential project developer should disseminate to the local people the information for the new infrastructure introduction and the beneﬁts anticipated.

Figure 11. 11. TilosTilos island island location. location.

AccordingThe island to counts the most four recent main (populated)census, Tilos settlements/villages,has a total of 829 registered namely, inh Megaloabitants. Chorio, whichThe actual is the population capital; Livadia, of the island which during is the the biggest winter village months and, however where the, is island’sestimated port to and narrowmain center down of to economic ~400–500, activities as most of are the found; residents Agios move Antonios to the more and Eristos.populated, vicinal islandsThe of climateRhodes ofand Tilos Crete. is typical The rest of of the the Mediterranean permanent residents region, graduall with mildy return winters to the and hot, islandsunny (between summers. April Concerning and May the each local year), RES on potential, time for Tilosthe tourist appreciates season, an with excellent-quality tourism suggestingsolar potential, the main determined economic by activity ~1800 on kWh/(m Tilos. In2.a) this at context, the horizontal and as already plane (Figure implied,6). On the local other population hand, the localincreases wind dramatically potential is ofduring medium the tourist quality, season with theof the average year, windwhich speed normallyranging betweenlasts for a 6 full and 6 8 months m/sec in(mid the-April largest to mid part-October). of the island The (seeisland also can Figure host ~12005). visitorsAccording, with the tonumber the most of tourists recent census,normally Tilos exceeding has a total the hosting of 829 registered capacity of inhabitants. hotels and The apartments by far, especially during the peak summer period of August. actual population of the island during the winter months, however, is estimated to narrow down to ~400–500, as most of the residents move to the more populated, vicinal islands 3.1. Tilos Electricity Grid Tilos belongs to the electrical complex of the Kos and Kalymnos islands, forming an electricity system of nine (9) islands in total, with Tilos located at the very southern tip of the complex. In more detail, Tilos is connected to the island of Kos via the R-44 feeder line (see relevant configuration in Figure 12). The R-44 feeder (in light blue) departs from the central bus bar at the Kos thermal power station, has three underwater sections between four islands and serves—via overhead lines—a few MV consumers on the Kos and Gyali island, as well as the entire electricity demand of the Nisyros and Tilos islands. Moreover, and for redundancy reasons, there are two undersea cables, each corresponding to a different cable link on Kos, with only one being typically operated, which means that the switch from one side of the second link remains open. The feed-in point from Kos stands also as a point of common coupling between the broader electricity system of Kos and Kalymnos and the electrical grid of Tilos. This means that Tilos can be isolated from the R-44 line and thus be operated as a completely independent, geographical island microgrid, relying exclusively on its own power generation sources.

Energies 2021, 14, 1336 10 of 22

of Rhodes and Crete. The rest of the permanent residents gradually return to the island (between April and May each year), on time for the tourist season, with tourism suggesting the main economic activity on Tilos. In this context, and as already implied, the local population increases dramatically during the tourist season of the year, which normally lasts for a full 6 months (mid-April to mid-October). The island can host ~1200 visitors, with the number of tourists normally exceeding the hosting capacity of hotels and apartments by far, especially during the peak summer period of August.

3.1. Tilos Electricity Grid Tilos belongs to the electrical complex of the Kos and Kalymnos islands, forming an electricity system of nine (9) islands in total, with Tilos located at the very southern tip of the complex. In more detail, Tilos is connected to the island of Kos via the R-44 feeder line (see relevant conﬁguration in Figure 12). The R-44 feeder (in light blue) departs from the central bus bar at the Kos thermal power station, has three underwater sections between four islands and serves—via overhead lines—a few MV consumers on the Kos and Gyali island, as well as the entire electricity demand of the Nisyros and Tilos islands. Moreover, and for redundancy reasons, there are two undersea cables, each corresponding Energies 2021, 14, x FOR PEER REVIEW 11 of 24 to a different cable link on Kos, with only one being typically operated, which means that the switch from one side of the second link remains open.

Figure 12. 12. TheThe R R-44-44 Interconnector Interconnector from from Kos. Kos.

TheThe Kos feed-in and Kalymnos point from island Kos complex, stands also to whi as ach point Tilos ofbelongs, common comprises, coupling as already between the mentioned,broader electricity an electrical system system of of Kos nine and (9) Kalymnosislands, with and an installed the electrical capacity grid of ~14 of Tilos.5 MW This (mainlymeans thatthermal, Tilos oil can-fired be isolatedunits), an from annual the peak R-44 load line demand and thus of be more operated than 9 as0 M aW completely and anindependent, annual electricity geographical demand island in the microgrid,order of 36 relying5 GWhe. exclusively The system on relies its ownmainly power on the genera- thermaltion sources. power station of Kos island (capacity of ~102 MW), employing heavy-oil fired enginesThe of Kos 87.3 and MW Kalymnos and diesel island-fired complex,units of 14. to6 which MW, and Tilos secondarily belongs, comprises, on the thermal as already powermentioned, station an of electrical Kalymnos system (capacity of nine of 18.1 (9)5 islands, MW), employing with an installed heavy-oil capacity units alone. of ~145 At MW the(mainly same thermal,time, RES oil-fired units across units), the ancomplex annual (excluding peak load Tilos) demand include of four more wind than parks 90 MW of and total installed capacity equal to 15.2 MW and 92 PV parks of ~8.8 MWp in total. The annual an annual electricity demand in the order of 365 GWhe. The system relies mainly on the energy yield of the latter is estimated in the order of 16.6 GWh (capacity factor of ~19%), thermal power station of Kos island (capacity of ~102 MW), employing heavy-oil fired while for the existing wind parks, the estimated annual energy yield reaches almost 46 engines of 87.3 MW and diesel-fired units of 14.6 MW, and secondarily on the thermal GWh (capacity factor of 34.5%) (see also Figure 6), with 1/3 of this number, however, being power station of Kalymnos (capacity of 18.15 MW), employing heavy-oil units alone. curtailed. This reflects the saturation levels of the Kos and Kalymnos electricity system in At the same time, RES units across the complex (excluding Tilos) include four wind terms of RES capacity. More specifically, due to the dynamic penetration limit for RES on theparks one of hand total (e.g. installed, RES contribution capacity equal should to 15.2 be no MW more and than 92 PV30% parks of instantaneous of ~8.8 MW loadp in total. dTheemand) annual and energythe technical yield minima of the latterof the isin- estimatedoperation thermal in the orderunits on of the 16.6 other, GWh the (capacity re- sidualfactor load of ~19%), available while for the for absorption the existing of windconsiderable parks, RES the estimatedpower generation annual is energy signifi- yield cantlyreaches reduced. almost The 46 GWhspecific (capacity shortcoming factor is critical of 34.5%) for the (see further also Figureincrease6), of with renewables 1/3 of this in the system, which, if not supported by energy storage and/or new interconnections, does not seem to suggest an economically viable option. The average load demand profile of Tilos currently adopted (Figure 7) is configured on the basis of a long-term measurement campaign carried out within the context of TILOS project, covering a period of almost 4.5 full years, i.e., from mid-2015 to autumn 2019 on the basis of 10 min measurements. To this end, the total electricity consumption of the island is found (Figure 13) to marginally exceed 3 GWhe (3.035 GWhe), while the average peak demand of the year exceeds 800 kW, although, if looking at individual years, summer peak demand is actually in the order of 900 kW.

Energies 2021, 14, 1336 11 of 22

number, however, being curtailed. This reflects the saturation levels of the Kos and Kalymnos electricity system in terms of RES capacity. More specifically, due to the dynamic penetration limit for RES on the one hand (e.g., RES contribution should be no more than 30% of instantaneous load demand) and the technical minima of the in-operation thermal units on the other, the residual load available for the absorption of considerable RES power generation is significantly reduced. The specific shortcoming is critical for the further increase of renewables in the system, which, if not supported by energy storage and/or new interconnections, does not seem to suggest an economically viable option. The average load demand profile of Tilos currently adopted (Figure7) is configured on the basis of a long-term measurement campaign carried out within the context of TILOS project, covering a period of almost 4.5 full years, i.e., from mid-2015 to autumn 2019 on the basis of 10 min measurements. To this end, the total electricity consumption of the island is found (Figure 13) to marginally exceed 3 GWhe (3.035 GWhe), while the average Energies 2021, 14, x FOR PEER REVIEW 12 of 24 peak demand of the year exceeds 800 kW, although, if looking at individual years, summer peak demand is actually in the order of 900 kW.

Figure 13. 13. TheThe 4.5 4.5 year year average average monthly monthly consumption consumption profiles proﬁles of Tilos of Tilos island. island.

3.2. Tilos Tilos Hybrid Hybrid Power Power Station Station InIn view view of of the the successful successful implementation implementation of the of theTILOS TILOS project, project, the first the-ever first-ever battery battery-- based, wind wind and and PV PV Hybrid Hybrid Power Power Station Station (HPS) (HPS) in Greece in Greece has hasbeen been developed. developed. At the At the same time, time, an an integrated integrated microgrid microgrid has has also also been been developed developed on the on island the island and several and several innovative elements elements have have been been introduced, introduced, altogether altogether transforming transforming Tilos Tilos into into an an exem- exemplary plaryisland island case incase terms in terms of local-scale of local-scale clean clean energy energy production production and and management. management. ThisThis comprised comprised a a major major breakthrough breakthrough,, not not only only for for Tilos, Tilos, but but also also for for the the Greek Greek en- energy ergymarket market as a whole,as a whole, disrupting disrupting the norms the norms of the of past the andpast presentingand presenting a new a energynew energy paradigm paradigmand solution and forsolution the electrification for the electrification of island of regions.island regions. Acknowledging Acknowledging the above, the above, the main theassets main and assets elements and elements of the power of the generationpower generation side are side first are presented. first presented. ConcerningConcerning the the electricity electricity generation generation sector, sector, the themain main power power generation generation assets assets that that are currently in in operation operation on on Tilos Tilos include include::  • thethe Tilos Tilos HPS, HPS, comprising: comprising: o an 800 kW, medium scale wind turbine an 800 kW, medium scale wind turbine o an 160 kWp PV station o# an an800 160 kW/2.8 kW8p MPVWh station integrated Battery Energy Storage System (BESS)  the# backan-up 800 diesel kW/2.88 genset MWhof 1.45 integrated MW, located Battery in the Energy village of Storage M. Chorio System (BESS) • distributed,the# back-up small diesel-scale genset PV installations of 1.45 MW, supporting located in theearly village prosumer of M. schemes Chorio and in- • cluding:distributed, small-scale PV installations supporting early prosumer schemes and in- ocluding: two monitored PV installations in M. Chorio & Livadia (capacity of 3.36 kWp at local residence and 4.93 kWp at the TILOS info-kiosk/EV charging station respectively) o four non-monitored PV installations (operational capacity of ~10 kWp in total). The above assets, together with the rest of the innovative elements that will also be analysed in the following paragraphs, comprise the integrated Tilos energy solution. As also reflected in Figure 14, the integrated Tilos energy solution suggests an energy microgrid that is normally found to interact with the host electricity system of Kos and Kalymnos (mainly the HPS), or, in rare cases, operated in isolation (such as in cases of emergency or microgrid testing).

Energies 2021, 14, 1336 12 of 22

two monitored PV installations in M. Chorio & Livadia (capacity of 3.36 kWp # at local residence and 4.93 kWp at the TILOS info-kiosk/EV charging station respectively) four non-monitored PV installations (operational capacity of ~10 kWp in total). #The above assets, together with the rest of the innovative elements that will also be analysed in the following paragraphs, comprise the integrated Tilos energy solution. As also reﬂected in Figure 14, the integrated Tilos energy solution suggests an energy microgrid that is normally found to interact with the host electricity system of Kos and Energies 2021, 14, x FOR PEER REVIEW 13 of 24 Energies 2021, 14, x FOR PEER REVIEWKalymnos (mainly the HPS), or, in rare cases, operated in isolation (such as13 in of cases 24 of emergency or microgrid testing).

Figure 14. Description of Tilos island microgrid and geographical location of main assets. FigureFigure 14. DescriptionDescription of of Ti Tiloslos island island microgrid microgrid and and geographical geographical location location of main of main assets. assets. 3.2.1. HPS-Wind Turbine 3.2.1.3.2.1. HPSHPS--WindWind TurbineTurbine The Tilos HPS wind turbine was installed in June 2017 by the Greek energy company, TheThe TilosTilos HPSHPS windwind turbineturbine waswas installedinstalled inin JuneJune 20172017 byby thethe GreekGreek energyenergy company,company, Eunice,Eunice, whichwhich which actedacted acted bothboth both asas as aa private aprivate private investorinvestor investor andand and aa majormajor a major partnerpartner partner ofof thethe of TILOSTILOS the TILOS projecprojec project.t.t. AsAs can can alsoalso bebe be seenseen seen fromfrom from FigureFigure Figure 14,14, 14 thethe, the windwind wind turbineturbine turbine isis coco is--locatedlocated co-located withwith with thethe HPSHPS the HPSbatterybattery battery storagestorage system,system, system, atat at thethe the northwesternnorthwestern northwestern tiptip tip ofof ofTilosTilos Tilos (area(area (area ofof Pachy),Pachy), of Pachy), nextnext next toto thethe to underseaundersea the undersea cablecable cable junctionjunction (( (FigureFigureFigure 1515 15))..). ToTo To thisthis this end,end, end, thethe the mainmain main technicaltechnical technical featuresfeatures features ofof thethe of En theEnerconercon Enercon EE--5353 E-53windwind wind turbineturbine areare are providedprovided provided inin in TableTable Table 1.1.1 .

Figure 15. Photo of the Tilos HPS wind turbine, installed in the northwestern part of the island. FigureFigure 15. PhotoPhoto of of the the Tilos Tilos HPS HPS wind wind turbine, turbine, installed installed in the in thenorthwestern northwestern part partof the of island. the island. Table 1. Main characteristics of installed wind turbine. Table 1. Main characteristics of installed wind turbine.

WindWind TurbineTurbine TypeType:: ENERCONENERCON EE--5353 NominalNominal capacity:capacity: 808000 kkWW RotorRotor diameter:diameter: 52.52.99 mm HubHub height:height: 6600 mm TowerTower type:type: TubularTubular steelsteel towertower WindWind Class/IECClass/IEC 6140061400--1:1: SS ElectricalElectrical configuration:configuration: GridGrid PerformancePerformance FTFT Location:Location: PachyPachy

Energies 2021, 14, 1336 13 of 22

Table 1. Main characteristics of installed wind turbine.

Wind Turbine Type: ENERCON E-53 Nominal capacity: 800 kW Rotor diameter: 52.9 m Hub height: 60 m Tower type: Tubular steel tower Wind Class/IEC 61400-1: S Energies 2021, 14, x FOR PEER REVIEW Electrical conﬁguration: Grid Performance FT 14 of 24 Location: Pachy

3.2.2.3.2.2. HPS HPS-PV-PV Plant Plant The Tilos Tilos HPS HPS PV PV plant plant was was also also installed installed in inJune June 2017 2017 by Eunice, by Eunice, and andis located is located at at thethe central part part of of the the island island (Figure (Figuress 14 14 and and 16), 16 sharing), sharing the thedistance distance between between the villages the villages ofof M. Chorio and and Livadia. Livadia. The The PV PV plant plant is iscomprise comprisedd of 592 of 592 polycrystalline polycrystalline PV panels PV panels of of 270270 Wp maximummaximum capacity capacity each, each, manufactured manufactured by byJA JASolar. Solar. Every Every 74 panels 74 panels form form an array an array withwith a maximum capacity capacity of of 19.9 19.988 kW kWp andp and are are connected connected to the to therespective respective array array inverter inverter ofof 2020 k kWW nominal nominal capacity, capacity, manufactured manufactured by by FRONIUS. FRONIUS. The The produced produced energy energy is collec is collectedted fromfrom eacheach stringstring inverterinverter andand directeddirected toto thethe MVMV transformertransformer andand consequentlyconsequently toto thethe local gridlocal overheadgrid overhead lines. lines. To this To end, this theend, main the main technical technical features features of the of PV the panels PV panels employed em- are providedployed are in provided Table2 on in theTable upper 2 on right,the upper together right, withtogether additional with additional information, information, such as the exactsuch as location the exact of location the installation. of the installation.

FigureFigure 16. PhotoPhoto of of the the Tilos Tilos HPS HPS PV PV plant, plant, installed installed at the at thecentral central part part of the of island. the island.

Table 2. Main characteristics of installed PV park. Table 2. Main characteristics of installed PV park. Type: JAP6(K)-60-270/4BB Type: JAP6(K)-60-270/4BB Maximum Power: 270 Wp Max PowerMaximum Current Power: (lmp): 8.67 270A Wp Max Power Current (lmp): 8.67 A Max Power Voltage (Vmp): 31.13 V Max Power Voltage (Vmp): 31.13 V ShortShort Circuit Circuit Current Current (lsc): (lsc): 9.18 A 9.18 A OpenOpen Circuit Circuit Voltage Voltage (Voc): (Voc): 38.17 38.17V V TemperatureTemperature Coefficient Coefﬁcient of of Voc: Voc: −0.330%/°C−0.330%/ ◦C OperatingOperating Temperature: Temperature: −40 °C−40 ~ +85◦C~ °C +85 ◦C MaximumMaximum System System Voltage: Voltage: 1000 1000V DC V DC × × Dimensions:Dimensions: 1650 1650× 991 ×991 35 mm35 mm Total surface/tilt angle: 1635.2 m2/30◦ Total surface/tilt angle: 1635.2 m2/30° Location/orientation Agios Konstantinos/South Location/orientation Agios Konstantinos/South

3.2.3.3.2.3. HPS HPS-BESS-BESS The integrated integrated battery battery system system of ofTilos Tilos was was fully fully installed installed on the on theisland island in January in January 2018,2018, following following the the earlier earlier installation installation of two of twobattery battery containers containers and the and introduction the introduction of individual battery modules right after. It comprises a fully containerized solution based on the high-temperature NaNiCl2 technology, standing as a fully recyclable storage configuration. The battery modules, which are manufactured by the Italian FZSonick, i.e., one of the lead partners of TILOS, under the type FZSonick ST523, are containerized into two 20 ft containers (Figure 17). Each of the containers comprises an energy storage system of 64 FZSonick ST523 modules that are connected in parallel. The technical specifications of the battery module and of the standard energy storage application are provided in Table

Energies 2021, 14, 1336 14 of 22

of individual battery modules right after. It comprises a fully containerized solution based on the high-temperature NaNiCl2 technology, standing as a fully recyclable storage configuration. The battery modules, which are manufactured by the Italian FZSonick, i.e., one of the lead partners of TILOS, under the type FZSonick ST523, are containerized into Energies 2021, 14, x FOR PEER REVIEWtwo 20 ft containers (Figure 17). Each of the containers comprises an energy storage15 of system24 of 64 FZSonick ST523 modules that are connected in parallel. The technical specifications of the battery module and of the standard energy storage application are provided in Table3. Con3cerning. Concerning the AC the side, AC each side, of eachthe battery of the containers battery containers is connected is connected to a dedicated to a dedicated bat- batterytery inverter inverter (PCS) (PCS) manufactured manufactured by the by Swiss the SwissIndrivetec Indrivetec (IDT). The (IDT). two The employed two employed in- invertersverters are are of of500 500 kVA kVA each each under under the thetype type SOLO SOLO S and S andare combined are combined with withtwo 63 two0 kVA, 630 kVA, 20/0.320/0.3 kV kV MV MV transformers, transformers, altogether altogether forming forming the thebattery battery AC ACside. side. The Theinverters’ inverters’ main main specificationsspecifications are are given given in in Table Table 4.4 .

Figure 17. PhotoPhoto of of the the Tilos Tilos HPS HPS BESS, BESS, installed installed at the at the northwestern northwestern part partof the of island. the island.

Table 3. Main characteristics of installed BESS. Table 3. Main characteristics of installed BESS. 1.4 MWh-100% NominalNominal Energy Energy Capacity Capacity 1.4 MWh-100% DOD @C/10 DOD @C/10 NominalNominal energy energy (nom. (nom. voltage-capacity) voltage-capacity) 1.44 MWhMWh NominalNominal current current capacity capacity 243 24322 A Ahh (100% DOD)DOD) Constant power discharge 400 kW for 3 h Constant power discharge 400 kW for 3 h Standard charge/discharge 8 h/3 h StandardMax heating charge/discharge power 8 64 h/ kW3 h Heater av.Max consumption@ heating power ﬂoating 6 104 kWkW HeaterMax av. charge consumption@ current floating 1 9600 kW A MaxMax discharge charge current current 1920960 A A Max chargeMax power discharge for controlled current charge 2161920 kW A No. of FIAMM gateway 1 Max charge power for controlled charge 216 kW No. of FIAMM gateway 1 Table 4. Main technical characteristics of the Indrivetec SOLO S inverter. Table 4. Main technical characteristics of the Indrivetec SOLO S inverter. Nominal AC Power (PAC) 450 kW At Power Factor cosϕ = 1 Nominal AC Power (PAC) 450 kW At Power Factor cosφ = 1 MaximumMaximum apparent apparent power power 50 5000 k kVAVA Power factor cosϕ ±0.9 Power factor cosφ ±0.9 AC Nominal operating voltage (UAC) 300 V UACmax = 330 V AC NominalAC Nominal operating current voltage (IAC) (UAC) 875300 AV UACmax 970 APEAK = 330 V RMS AC NominalGrid Frequency current (IAC) 875 50 Hz A 970 APEAK±10% RMS Harmonic DistortionGrid Frequency (% THD IAC)50 <3% Hz ±10% Harmonic DistortionEfﬁciency (% THD IAC) 98.2%<3% Charging and discharging ControlEfficiency Modes P/Q—U/f98.2% Charging and discharging Control Modes P/Q—U/f In contract with the two RES units, the integrated BESS was commissioned in April– MayIn 2018. contract This with enabled the two the energizationRES units, the and integrated warm-up BESS of thewas battery commissioned modules in and April allowed– forMay the 2018. execution This enabled of a series the energization of Site Acceptance and warm Tests-up (SAT).of the Thebattery SAT modules lasted forand a al- total of 2lowed weeks for and the included execution communication of a series of Site tests, Acceptance software Tests integration, (SAT). The emergency SAT lasted tests, for battery a total of 2 weeks and included communication tests, software integration, emergency tests, battery warm-up, power exchange between containers, cycling, grid perturbation and internal black start, as well as a set of set-point profile tests.

Energies 2021, 14, 1336 15 of 22

warm-up, power exchange between containers, cycling, grid perturbation and internal black start, as well as a set of set-point proﬁle tests.

3.3. Tilos Island Demand Side Management Following the detailed presentation of the generation side and energy management aspects, the demand side of the island is next presented, with the focus given on the Smart Metering and DSM platform of Tilos, bringing together a pool of almost 100 end consumers and community loads. Note that the remote electricity grid of Tilos comprises a vulnerable system often faced with power cuts, which affects living conditions on the island in an adverse manner, hindering also economic activities, especially during the summer period. Concerning consumers, there are no MV customers on the island, neither large-scale consumers, which is attributed to the lack of local industrial activities. Essentially, it is the local residential and tertiary sector that carry the biggest share of electricity consumption, with local hotels and summer apartments increasing the local demand during the summer period, altogether belonging to the broader category of the building sector analysed accordingly. Apart from the local building sector, other important sources of electricity consumption correspond to community-scale loads, such as water pumping (a total of 8 water pumping systems of ~62 kW and several smaller scale borehole pumps for irrigation) and public lighting (see also the respective monthly consumption profile in Figure8), while grid-connected is also the telecommunication infrastructure of the island. As already seen, the Smart Metering and DSM platform, developed by Eurosol (also partner of TILOS) and currently owned by local end users and the Municipality of Tilos, comprises one of the main components of the integrated energy solution of Tilos, bringing together a pool of ~100 Smart Meter and DSM panels (Figure9). The prototype DSM panels can be used both directly by local end users (mainly for electricity consumption monitoring) and offer a pool of controllable loads. In this context, the DSM platform can in aggregate serve different purposes in order to improve the operation of the Tilos microgrid (e.g., load shifting in favour of increased RES shares or control of loads during recovery from blackout events / stand-alone island operation of Tilos in support of the local HPS operation, etc.), comprising at the same time the end-nodes of the integrated Smart Metering and DSM platform. The latter is also responsible for the monitoring, data collection, classification and remote, centralized control of all active DSM loads, acting at the same time as an integrated element of the developed High-Level Energy Management Centre (HL-EMC). DSM panels were gradually installed during the implementation of TILOS, distributed across local residences, hotels, commercial stores and the public sector, as well as in eight (8) water pumping stations of the island standing as community-level loads. Their distribution over the two main villages of Livadia and M. Chorio is given in Figure9, followed by the distribution of controlled community loads across the entire island. Apart from the monitoring and control of loads at the end-user level, the entire pool of loads can also be centrally monitored at the DSM server, co-located with the HL-EMS server at the old power station of the island. Following the commissioning of the HL- EMC/EMS, the Smart Metering and DSM solution acquired a twofold role, acting as an independent system branch (aggregate pool of DSM loads that can support different types of grid services) on the one hand, and supporting interoperability with the rest of the system components as part of the HL-EMC on the other. In this context, load demand monitoring is also active in aggregate fashion, supported by an advanced UI and relevant dashboards that group together DSM loads per type of end user. In the meantime, further classification, based on the type and elasticity/flexibility of loads has been carried out. What should be stressed at this point is that the total DSM loads’ installed capacity rises to more than 700 kW, which is expected to ensure a minimum DSM pool of 15–20% in comparison to the appearing load demand of the island at all times. Energies 2021, 14, 1336 16 of 22 Energies 2021, 14, x FOR PEER REVIEW 17 of 24

4. Evaluation of Tilos Island Solution 4. Evaluation of Tilos Island Solution After the the completion completion of of the the commissioning commissioning phase phase for forthe theRES RES units, units, the HPS the HPSof Tilos of Tilos became fully fully functional functional and and,, as as such such,, it entered it entered the the “prosumer “prosumer”” stage stage of operation. of operation. To this To this end, consecutive, consecutive, 20 20 day day duration duration time time windows windows of trial of trial operation operation were were exploited exploited in order in order to test the the HPS HPS performance, performance, first first under under a certain a certain daily daily profile/schedule profile/schedule of max ofmax-permittedpermitted power output output per per hour hour of of the the day. day. The The specific specific profile profile (provided (provided by the by local the local network network manager-Kosmanager-Kos APS) APS) is is depicted depicted in inFigure Figure 18 18(with (with the themax max-permitted-permitted pow powerer output output of the of the entire HPS HPS set set to to vary vary between between the the fixed fixed values values of 0 of kW, 0 kW, 400 400kW kWand and800 k 800W) kW)and it and is the it is the one governing the the operation operation of ofthe the HPS HPS from from mid mid-September-September until until early early in January, in January, i.e., i.e., when a a second second round round of of trial trial operation operation started started for forthe thetestin testingg of different of different set-point set-point profiles profiles for the HPS. HPS.

Standard, Max Set-Point Profile of the Consecutive 20-day Time Windows (14 September 2018 to 8 January 2019) 900

800

700

600

500

400

300

Max HPS Power (kW) Max HPS Power 200

100

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour of the Day

Figure 18. 18. TheThe standard standard profile proﬁle of ofmax max set set-point-point for forthe theTilos Tilos island island HPS. HPS.

More specifically, specifically, following following the the issuance issuance of ofall all the the necessary necessary permits permits in 2018, in 2018, the the wind turbine was was commissioned commissioned in inmid mid-September-September and and was was fully fully activated activated on September on September 19, 2018, 2018, i.e. i.e.,, when when it started it started generating generating power power as part as of partthe integrated of the integrated HPS of Tilos. HPS With of Tilos. Withregards regards to its anticipated to its anticipated energy energyperformance, performance, the medium the quality medium wind quality potential wind of potential the ofarea the (long area-term (long-term average average wind speed wind of speed ~6.5 m of/sec ~6.5 at 3 m/sec0 m height, at 30 based m height, on on based-site meas- on on-site measurementsurements from froma dedicated a dedicated wind windmast) mast) produces produces an annual an annual capacity capacity factor factor of 27.5 of–30% 27.5–30% concerning the the theoretical theoretical output output of ofthe the machine, machine, which which is equivalent is equivalent to an to annual an annual power power generation between 1.9 GWhe and 2.1 GWhe. However, during the examined period, the generation between 1.9 GWhe and 2.1 GWhe. However, during the examined period, actual wind energy production falls below these numbers, owing primarily to the satura- the actual wind energy production falls below these numbers, owing primarily to the tion levels of the overall Kos and Kalymnos system noted earlier, with the ex-post capacity saturation levels of the overall Kos and Kalymnos system noted earlier, with the ex-post factor estimated in the area of 20%. In addition, and during the demonstration stage of the capacity factor estimated in the area of 20%. In addition, and during the demonstration TILOS project, that coincided also with the very early period of HPS operation, trial tests stage of the TILOS project, that coincided also with the very early period of HPS operation, for the HPS dictated the application of certain set-point profiles, which further compro- trialmised tests the forexploitation the HPS of dictated local wind the energy application production. of certain The set-pointcombined profiles, impact of which the two further compromisedis better reflected the in exploitation the figures (Figure of local 1 wind9a,b), comparing energy production. the actual Theand combinedtheoretical impactpro- of theduction two of is betterthe wind reflected turbine in till the the figures end of (Figure January 19 2019a,b), (end comparing-month of the TILOS actual project). and theoretical production of the wind turbine till the end of January 2019 (end-month of TILOS project).

Energies 2021, 14, 1336 17 of 22 Energies 2021, 14, x FOR PEER REVIEW 18 of 24

Actual vs Theoretical Wind Energy Production (1/9/2018-25/1/2019) 900 Actual Theoretical 800

700

600

500

400

300 Wind Power (kW) Wind Power

200

100

0 1/9/18 8/9/18 5/1/19 15/9/18 22/9/18 29/9/18 6/10/18 3/11/18 1/12/18 8/12/18 12/1/19 19/1/19 26/1/19 13/10/18 20/10/18 27/10/18 10/11/18 17/11/18 24/11/18 15/12/18 22/12/18 29/12/18 (a)

Absorbed vs Curtailed Wind Energy Production (1/9/2018-25/1/2019) 900 800 Absorbed Curtailed 700 600 500 400 300 200 100 0 -100 1/9/18 8/9/18 5/1/19

-200 15/9/18 22/9/18 29/9/18 6/10/18 3/11/18 1/12/18 8/12/18 12/1/19 19/1/19 26/1/19 13/10/18 20/10/18 27/10/18 10/11/18 17/11/18 24/11/18 15/12/18 22/12/18 29/12/18

Wind Power (kW)Wind -300 -400 -500 -600 -700 -800 -900

(b)

FigureFigure 19. 19. (a)( aActual) Actual vs vstheoretical; theoretical; (b) (absorbedb) absorbed vs curtailed vs curtailed wind wind energy energy over the over period the period of of TILOS TILOSdemonstration. demonstration.

ToTo this this end, end, the the levels levels of ofwind wind energy energy curtailments curtailments encountered encountered are arebetter better illustrated illustrated in inFigure Figure 19 19b,b, while while another another wayway toto expressexpress andand quantifyquantify the difference between between the the po- potential tentialand the and actual the actual wind wind power power generation generation is is given given in in thethe results of of Figure Figure 20a, 20a,b,b, where where the theactual actual and and theoretical theoretical capacity capacity factors factors areare compared.compared. In this context, context, the the resulting resulting av- average erage difference is in the order of 10%, i.e., the expected capacity factor on the basis of the difference is in the order of 10%, i.e., the expected capacity factor on the basis of the available wind potential and operational curve of the wind turbine used is almost 30%, available wind potential and operational curve of the wind turbine used is almost 30%, while the finally absorbed wind power by the local grid corresponds to capacity factor while the ﬁnally absorbed wind power by the local grid corresponds to capacity factor value equal to 20%. This real-world situation means that 1/3 of the energy yield of the value equal to 20%. This real-world situation means that 1/3 of the energy yield of the wind turbine during the given period was curtailed by the local network operator due to limitedwind turbineelectricity during demand the and given the period restrictions was curtailedimposed by by the the local local grid network stability operator and thedue to existinglimited thermal electricity units demand technical and minima. the restrictions imposed by the local grid stability and the existing thermal units technical minima.

Energies 2021, 14, 1336 18 of 22 Energies 2021, 14, x FOR PEER REVIEW 19 of 24

Wind Turbine Theoret. Capacity Factor (19/9/2018-25/1/2019) 100% Hourly CF Average CF 90%

80%

70%

60%

50%

40%

30% Wind Turbine Hourly CF TurbineWind Hourly 20%

10%

0% 2/1/19 9/1/19 19/9/18 26/9/18 3/10/18 7/11/18 5/12/18 16/1/19 23/1/19 10/10/18 17/10/18 24/10/18 31/10/18 14/11/18 21/11/18 28/11/18 12/12/18 19/12/18 26/12/18 (a)

Wind Turbine Actual Capacity Factor (19/9/2018-25/1/2019) 100% Hourly CF Average CF 90%

80%

70%

60%

50%

40%

30% Wind Turbine Hourly CF Hourly Wind Turbine 20%

10%

0% 2/1/19 9/1/19 19/9/18 26/9/18 3/10/18 7/11/18 5/12/18 16/1/19 23/1/19 10/10/18 17/10/18 24/10/18 31/10/18 14/11/18 21/11/18 28/11/18 12/12/18 19/12/18 26/12/18 (b)

FigureFigure 20. 20. (a)( aTheoretical;) Theoretical; (b) (actualb) actual wind wind capacity capacity factor factor over overTILOS TILOS demonstration demonstration stage. stage.

SimilarSimilar to tothe the case case of ofthe the wind wind turbine turbine installation, installation, commissioning commissioning of the ofPV the station PV station waswas also also successfully successfully completed completed in September in September 2018, 2018, enabling enabling power power generation generation on the on the 14th14th of of September September 2018. 2018. To this To thisend, end,and as and far as the far expe as thected expected energy performance energy performance of the of 2 PVthe station PV station is concerned, is concerned, the excellent the excellent solar potential solar potentialof the area of (~180 the area0 kWh/ (~1800(m .a) kWh/(m at the 2.a) horizontalat the horizontal plane) produces plane) producesan annual ancapacity annual factor capacity of ~18– factor19% concerning of ~18–19% the concerning theoreti- the caltheoretical output of the output station, of thewhich station, is equivalent which to is an equivalent annual power to an generation annual powerbetween generation 250 and 270 MWhe. Although subject to considerable curtailments as well, the PV station is between 250 and 270 MWhe. Although subject to considerable curtailments as well, the considered to be less vulnerable in comparison to the wind turbine, mainly owing to its PV station is considered to be less vulnerable in comparison to the wind turbine, mainly more dispatchable character as a renewable energy source. In this context, energy gener- owing to its more dispatchable character as a renewable energy source. In this context, ation results from the operation of the PV station during the period examined are shown energy generation results from the operation of the PV station during the period examined in the following Figure 21. As one may see in this figure, the maximum power output is limitedare shown in the in order thefollowing of 130 kW, Figure owing 21 to. the As winter one may season see inencountered this ﬁgure, during the maximum the refer- power output is limited in the order of 130 kW, owing to the winter season encountered during the reference period. On the other hand, it is for almost 80% of the time (including night-time periods) that the PV generation output falls below 40 kW, which compares unfavorably with the local load demand and also challenges the predictability of the RES resource, especially with regards to the obligation of the HPS to provide day-ahead guaranteed Energies 2021, 14, x FOR PEER REVIEW 20 of 24

Energies 2021, 14, 1336 ence period. On the other hand, it is for almost 80% of the time (including night-time pe- 19 of 22 riods) that the PV generation output falls below 40 kW, which compares unfavorably with the local load demand and also challenges the predictability of the RES resource, especially with regards to the obligation of the HPS to provide day-ahead guaranteed energy offers.energy Concerning offers. Concerning the capacity the factor capacity of the factorPV station of the over PV the station demonstration over the period demonstration of TILOS,period it ofwas TILOS, estimated it was at estimated12%, which at for 12%, the whichspecific for (September the speciﬁc to January) (September season to January)of theseason year is of considered the year is satisfactory. considered satisfactory.

PV Station Power Generation (14/9/2018-25/1/2019) 160 150 140 130 120 110 100 90 80 70 60 PV Power (kW) Power PV 50 40 30 20 10 0 1/9/18 8/9/18 5/1/19 15/9/18 22/9/18 29/9/18 6/10/18 3/11/18 1/12/18 8/12/18 12/1/19 19/1/19 26/1/19 13/10/18 20/10/18 27/10/18 10/11/18 17/11/18 24/11/18 15/12/18 22/12/18 29/12/18 (a)

Duration Curve of the PV Production (14/9/2018-25/1/2019) 100%

90%

80%

70%

60%

50%

Probability 40%

30%

20%

10%

0% 0 5 -5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 155 160 PV Power Class (kW)

(b)

FigureFigure 21. 21. (a)( aPV) PV station station power power generation; generation; (b) relevant (b) relevant duration duration curve. curve.

Recapitulating,Recapitulating, in Figure in Figure 22, 22we, wedemonstrate demonstrate the daily the daily energy energy balance balance between between the the HPSHPS production production and and the the local local island island electricity electricity consumption, consumption, together together with a with breakdown a breakdown ofof the the HPS HPS production production to its to itsmain main components components (i.e., (i.e.,wind wind turbine, turbine, PV station PV station and battery and battery operation).operation). Note Note that that during during the thefirst ﬁrst two twoweeks weeks of September of September the HPS the had HPS not had received not received thethe energy energy production production licence, licence, thus thus electrical electrical energy energy was consumed was consumed (imported (imported by the bylocal the local grid)grid) in in order order to tokeep keep in operation in operation (at desired (at desired temperature) temperature) and test and the test two the battery two battery banks. banks. As one may see, for the ﬁrst two months of operation, the HPS contribution was moderate, directly dependent on both the performance of the wind turbine and the hourly cap on the basis of the adopted schedule (Figure 18). Following the end of that initial period, which coincides also with a considerable reduction of the load demand and an improvement of the wind turbine performance, contribution of the HPS to the local consumption was found Energies 2021, 14, x FOR PEER REVIEW 21 of 24

As one may see, for the first two months of operation, the HPS contribution was moderate, Energies 2021, 14, 1336 20 of 22 directly dependent on both the performance of the wind turbine and the hourly cap on the basis of the adopted schedule (Figure 18). Following the end of that initial period, which coincides also with a considerable reduction of the load demand and an improvement of the wind turbine performance, contribution of the HPS to the local consumption towas increase found to remarkably, increase remarkably, especially especially during during December December 2018, even2018, even under under these these sub-optimal conditionssub-optimal ofconditions operation of operation imposed imposed by the local by the network local network management. management.

Comparison between the Daily HPS Generation & the Island Electricity Consumption (1/9/2018-8/01/2019) 15 14 Wind 13 PV 12 BESS 11 HPS 10 Consumption 9 8 7 6 5 4 3 Daily Energy (MWh) Daily Energy 2 1 0 -1 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106 111 116 121 126 -2 Day of the Period

Figure 22. 22. TheThe daily daily energy energy balance balance analysis analysis for Tilos for Tilosisland. island.

WhatWhat is is worth worth noticing noticing is that is that for December, for December, local-only local-only RES shares RES exceeded shares exceeded 70%, 70%, which,which, impressively impressively enough, enough, even even reach reach 90% 90%of the of local the consumption local consumption if RES imports if RES and imports and exports are are also also taken taken into into account. account. On the On other the otherhand, hand,RES shares RES sharesduring the during summer the summer period drop in the order of 25–30%, which is due to both the lower wind energy yield and period drop in the order of 25–30%, which is due to both the lower wind energy yield and the higher electricity demand appearing during that period (see also Figure 23). Note that theeven higher these extremely electricity low demand RES participation appearing values during are that quite period higher (see than also the one Figures appear- 23). Note that evening in these Figure extremely 1, not exceeding low RES 22% participation in any occasion. values As are one quite may highersee from than Figure the 22, ones the appearing ininclusion Figure 1of, notRES exceeding exports makes 22% inlittle any difference, occasion. which As one reflects may seeon fromthe increasedFigure 22curtail-, the inclusion ofments RES faced exports over makesthe period little of study. difference, On anwhich annual reﬂectsbasis, the on energy the increased provided of curtailments the Tilos faced overisland the HPS period is slightly of study. less than On 150 an0 annual MWhe, basis,mainly the due energy to the local provided network of curtailments the Tilos island HPS Energies 2021, 14, x FOR PEER REVIEWis and slightly the poor less exploitation than 1500 profile MWh dictatede, mainly by the due local to thenetwork local manager, network completely curtailments22 ne- of 24 and the poorglecting exploitation the values provided proﬁle dictated by the existing by the forecasting local network algor manager,ithms. completely neglecting the values provided by the existing forecasting algorithms.

FigureFigure 23. 23.Actual,Actual, monthly monthly energy energy output (absorbed) output (absorbed) of Tilos HPS of Tilosvs. Tilos HPS island vs. electricity Tilos island con- electricity sumption,consumption, as recorded as recorded from HEDNO. from HEDNO.

Energies 2021, 14, 1336 21 of 22

5. Conclusions and Proposals The implementation of the integrated TILOS energy solution under the current local legislative frame is a great success story introducing several important innovative characteristics in the European market. More specifically, the combined operation of a wind turbine and a PV installation (hybrid power station), the application of new technology battery energy storage, the installation of DSM network/platform and the development of a large number of reliable forecasting algorithms are among the main results of the proposed strategy. Undoubtedly, the core component of the integrated TILOS solution, i.e., the local HPS, carries a bundle of innovative features synopsized below: • The HPS of Tilos comprises the first-ever, fully licensed, MW-scale, battery-based HPS in Greece. As such, it disrupted the local energy market and introduced new aspects in the Greek energy agenda, especially with regards to non-interconnected island electricity systems. • Although rather sophisticated, the Greek regulation for HPSs was tailored to the operation of pumped hydro, concerning the storage technologies employed. A series of advancements were introduced to this end to the local regulations within the context of TILOS in order to accommodate the operation of battery-based HPSs, positively affecting the early-stage storage market of Greece as well. • The HPS of Tilos has been the first to comply with the requirements of a day-ahead dispatch operation that is based on the declaration of guaranteed energy (power) offers and which entails the need for the exploitation of forecasting means as well. • The integrated BESS is able to offer a bundle of services including RES balancing and ancillary services to the local grid, like frequency and voltage regulation, while, under a set of conditions, it can also support black-starting of Tilos; thus standing as a novel, multiple-service storage system. To our knowledge, the microgrid of Tilos is one of the most advanced island microgrids in Europe with smart aspects and many novel technologies and components. All these novel elements, following the TILOS project demonstration stage, are gradually evolving in order to support the operation of a mature energy ecosystem, fostering in the course of time the addition of new agents and actors such as the envisaged pool of prosumers, towards the full-scale decarbonisation of the island of Tilos. As far as the energy performance is concerned, evidence of operations that are drawn from the core period of TILOS project demonstration stage (i.e., May 2018 to January 2019) is provided. The specific period is exploited in order to give a preliminary performance assessment of individual assets comprising the overall Tilos energy solution, which, although determined by sub-optimal operational conditions during the examined span, does provide a first set of indications on the potential of the system to provide increased shares of RES for the island of Tilos. According to the results obtained, the innovative integrated Tilos island solution can be definitely improved by adopting already existing forecasting techniques in order to improve the energy balance between the local HPS and the main grid of Kos-Kalimnos, ameliorating also the battery bank energy management. Moreover, remarkable RES- based energy production curtailments should be minimized, introducing the opportunities of clean electromobility. In any case, the real-world operating example of Tilos island may be equally well applied in several other remote islands all over Europe with very promising results.

Funding: This research received no external funding. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: All data are based on our Lab research efforts. Energies 2021, 14, 1336 22 of 22

Conﬂicts of Interest: The author declares no conﬂict of interest.

References 1. Kuang, Y.; Zhang, Y.; Zhou, B.; Li, C.; Zeng, L. A review of renewable energy utilization in islands. Renew. Sustain. Energy Rev. 2016, 59, 504–513. [CrossRef] 2. Cross, S.; Padfield, D.; Ant-Wuorinen, R.; King, P.; Syri, S. Benchmarking island power systems: Results, challenges, and solutions for long term sustainability. Renew. Sustain. Energy Rev. 2017, 80, 1269–1291. [CrossRef] 3. Anagnostatos, S.D.; Halevidis, C.D.; Polykrati, A.D.; Bourkas, P.D.; Karagiannopoulos, C.G. Examination of the 2006 blackout in Kefallonia Island, Greece. Int. J. Electr. Power Energy Syst. 2013, 49, 122–127. [CrossRef] 4. Kaldellis, J.K.; Zafirakis, D. Present situation and future prospects of electricity generation in Aegean Archipelago islands. Energy Policy 2007, 35, 4623–4639. [CrossRef] 5. Chalvatzis, K.J.; Ioannidis, A. Energy supply security in the EU: Benchmarking diversity and dependence of primary energy. Appl. Energy 2017, 207, 465–476. [CrossRef] 6. De Vos, K.; Petoussis, A.G.; Driesen, J.; Belmans, R. Revision of reserve requirements following wind power integration in island power systems. Renew. Energy 2013, 50, 268–279. [CrossRef] 7. Kaldellis, J.K.; Kapsali, M.; Tiligadas, D. Presentation of a stochastic model estimating the wind energy contribution in remote island electrical networks. Appl. Energy 2012, 97, 68–76. [CrossRef] 8. Kaldellis, J.K.; Kavadias, K.A.; Papantonis, D.E.; Stavrakakis, G.S. Maximizing the Contribution of Wind Energy in the Electricity Demand Problem of Crete Island. Wind Eng. J. 2006, 30, 73–92. [CrossRef] 9. Kaldellis, J.K. The Wind Potential Impact on the Maximum Wind Energy Penetration in Autonomous Electrical Grids. Renew. Energy J. 2008, 33, 1665–1677. [CrossRef] 10. Papathanassiou, A.S.; Boulaxis, N.G. Power limitations and energy yield evaluation for wind farms operating in island systems. Renew. Energy J. 2005, 31, 457–479. [CrossRef] 11. Tzanes, G.; Zafeiraki, E.; Papapostolou, C.; Zafirakis, D.; Moustris, K.; Kavadias, K.; Chalvatzis, K.; Kaldellis, J.K. Assessing the Status of Electricity Generation in the Non-Interconnected Islands of the Aegean Sea Region. In Proceedings of the Applied Energy Symposium and Forum, Renewable Energy Integration with Mini/Microgrids—REM 2018, Rhodes, Greece, 29–30 September 2018. 12. Kaldellis, J.K.; Tzanes, G.T.; Papapostolou, C.; Kavadias, K.; Zafirakis, D. Analyzing the Limitations of Vast Wind Energy Contribution in Remote Island Networks of the Aegean Sea Archipelagos. Energy Procedia 2017, 142, 787–792. [CrossRef] 13. Hellenic Electricity Distribution Network Operator, Issues of the Operator for the non-Interconnected Islands, Athens, Greece. 2018. Available online: https://www.deddie.gr (accessed on 17 September 2020). 14. Technology Innovation for the Local Scale Optimum Integration of Battery Energy Storage (TILOS), Horizon 2020 Program. Available online: http://tiloshorizon.eu (accessed on 17 September 2020). 15. Cabrera, P.; Lund, H.; Carta, J.A. Smart renewable energy penetration strategies on islands: The case of Gran Canaria. Energy 2018, 162, 421–443. [CrossRef] 16. Kaldellis, J.K.; Zafirakis, D.; Kavadias, K. Techno-economic comparison of energy storage systems for island autonomous electrical networks. Renew. Sustain. Energy Rev. 2009, 13, 378–392. [CrossRef] 17. Greek Public Power Corporation. Annual Report on the Operation of Autonomous Power Stations; Greek Public Power Corporation: Athens, Greece, 2016. 18. Kaldellis, J.K.; Zafirakis, D. Optimum Energy Storage Techniques for the Improvement of Renewable Energy Sources-Based Electricity Generation Economic Efficiency. Energy J. 2007, 32, 2295–2305. [CrossRef] 19. Kaldellis, J.K.; Zafirakis, D. Prospects and challenges for clean energy in European Islands. The TILOS Parad. Renew. Energy 2020, 145, 2489–2502. [CrossRef] 20. Moustris, K.; Kavadias, K.A.; Zafirakis, D.; Kaldellis, J.K. Medium, Short and Very Short-Term Prognosis of Load Demand for the Greek Island of Tilos Using Artificial Neural Networks and Human Thermal Comfort-Discomfort Biometeorological Data. Renew. Energy 2020, 147, 100–109. [CrossRef] 21. Zafirakis, D.; Tzanes, G.T.; Kaldellis, J.K. Forecasting of wind power generation with the use of Artificial Neural Networks and Support Vector Regression models. Energy Procedia 2019, 159, 509–514. [CrossRef] 22. Stephanides, P.; Chalvatzis, K.; Li, X.; Mantzaris, N.; Zafirakis, D. Sustainable energy solutions for the Aegean Archipelago Islands: What is the public attitude? Energy Procedia 2019, 159, 243–248. [CrossRef]