1. Introduction
According to the IEA report “World Energy Outlook 2020” [
1], world energy demand has continuously increased over the last decades, except for a slight decrease in 2020 because of the COVID-19 pandemic. The growing tendency has returned in 2021 even though the COVID-19 pandemic has not yet ended [
2]. Most of this produced energy comes from fossil fuels, and similar figures are shown when regarding power generation. In this case, approximately two-thirds are generated through fossil fuels [
3].
This situation of generation based on fossil fuels is unsustainable, fundamentally posing, on the one hand, a more than foreseeable depletion of fossil fuels in the medium term if the current rate of consumption is maintained [
4,
5]. The second problem focuses on pollutant emissions, particularly greenhouse gas emissions, which are gases produced in large quantities when energy uses are centered on fossil fuels as combustibles or raw materials [
6,
7].
Therefore, there are many reasons for renewable energies to be present or even to be the only generation sources used, aiming to reduce or eliminate fossil fuels [
8]. Concentrating on electricity generation, the use of renewable energies is a must, since otherwise it is impossible to reach the ambitious objectives of CO
2 emissions reduction, combined with the sharp increase in the use of electricity, which has the final energy consumption of all countries [
9], expecting to exceed 30% of the total amount in a short time in many of them [
10].
The described problem is further complicated in remote regions, such as islands [
11]. This is because their small size and inaccessible location make connecting to a large grid difficult or even technically or economically impossible. Therefore, in the case of islands, the typical solution is to have a fossil fuels-based energy system (coal, gas, and diesel), since these systems have the advantage of high reliability. However, as disadvantages, these systems have a huge emissions problem and also a strong dependency on a big and complex supply chain. In addition, the countries producing these fuels are in many cases unstable, with the consequent risk of shortages that can reduce system reliability, added to the inconvenience that the prices of these fuels suffer large oscillations with frequently and unexpected rising of prices due to cartel decisions [
11].
Consequently, from both environmental and strategic (energy autonomy) points of view, renewable energies are an option that will be implemented in many places [
12,
13,
14,
15]. However, when renewable energies take on an important weight in the generation of power, a series of challenges arise, mainly associated with the intrinsic variability that this type of source presents [
16,
17,
18]. Specifically, the two renewable sources that are currently capable of covering the existing energy needs, solar photovoltaic and wind power, present a strong variability, with longer or shorter periods of low or strong wind, solar cycles, cloudy or rainy days, etc. All this means that to cover the energy needs with these sources, the system needs to be oversized and must include large storage systems to be able to absorb at least part of the excess to have them available in case of need. Even so, there will inevitably be excesses, but in this case, they will be more limited. Thus, the optimal generation and storage should, therefore, be sized, usually from an economic point of view, i.e., installing a generation power and storage capacity able to cover the energy needs at the lowest cost. Storage systems (pumping stations, mega-batteries) and large-scale generation (wind and solar PV) present an added problem in many cases on islands, caused by the scarcity of appropriate sites that would be required [
19].
Focusing specifically on the case of the Canary Islands, this archipelago consists of seven islands separated by less than 100 km from the Moroccan coast at its closest point and about 300 km at its furthest one, being about 1500 km from the mainland European coast. All 7 islands are inhabited by more than 2 million people [
20], and the 2 major islands, Gran Canaria and Tenerife, have around 950 and 850 thousand inhabitants, respectively. Lanzarote and Fuerteventura have approximately 150 and 120 thousand, respectively; finally, La Palma, La Gomera, and El Hierro have a population of slightly above 80, 20, and 10 thousand inhabitants, respectively (official data of 2020). The total final energy demand for the whole archipelago in the year 2019 was about 9.4 TWh (values which had remained almost constant during around the last 10 years and that have been reduced in 2020 and 2021 mainly because of the COVID-19 pandemic) [
20]. Only slightly below 1.5 TWh of the total energy demand were generated through renewable sources, i.e., around 16% of the total generation. The energy demand broken down by islands and expressed in GWh per year (also expressed in percentages in brackets) is 3711 (40%), 3582 (38.4%), 906 (9.7%), 717 (7.7%), 281 (2.9%), 76.9 (0.8%), and 62.4 (0.5%) for Tenerife, Grand Canary, Lanzarote, Fuerteventura, La Palma, La Gomera, and El Hierro, respectively.
This study proposes an autonomous generation system of electric power based on renewable energy and storage technologies (reversal pumping and mega-battery systems) with no CO2 emissions. The optimizing criteria are based mainly on economic reasons, but energy waste optimization and land occupation have also been considered. The analysis has been applied to the islands aiming to be energetically autonomous, estimations for each island have been carried out, and the comparison to a future scenario with the electric system interconnected between them has been made.
In view of the aforementioned situation, renewable energy generation systems have to be based on solar photovoltaic and wind technologies, because are the uniquely mature-enough technologies that currently might substitute conventional fossil technologies [
8,
21]. This is because other mature technologies that could also help in this process, such as hydroelectric power stations, are not available due to the absence of water resources in the Canary Islands. The solution to these feasibility problems, in all likelihood, involves the installation of high-energy storage capacities [
22,
23]. Consequently, the combined use of a renewable energy system backed by a high-capacity storage system will satisfy the local energy demand much more efficiently than a stand-alone renewable energy facility, due to the inherent variability of renewable sources [
24].
All countries of the European Union must face the decarbonization of the economy by 2050. In this sense, the Canary Islands are working with a clear time limit in their strategy to reduce fossil fuels dependence and, particularly in the case of the Canary Islands, they have the advantage of the rich natural resources of the archipelago, such as the wind and the sun. The Canary Islands Government and the Spanish national government are yet more optimistic and have scheduled to bring forward the end of the decarbonization process by 10 years (PTECan project) [
25,
26].
As a vanguard of this ambitious project, it is worth mentioning that in recent years all of the Canary Islands have been expanding their wind farms and solar parks to a greater extent. In addition to the projects that are already in the planning or execution phase, many other efforts are necessary to achieve the ambitious results that are being sought. For instance, the Chira-Soria pumping reversal station project is being carried out on the Grand Canary Island [
27], a facility that involves a high storage capacity, between 3.2 and 3.6 GWh, and a total generation capacity of 200 MW, with an estimated cost of around EUR 400 million. Between Lanzarote and Fuerteventura (Playa Blanca–La Oliva), there currently exists a submarine interconnection of approximate 16 km, 132 kV, and 121 MVA of transport capacity [
28], while there is a project for another interconnection between Tenerife and the La Gomera islands (Chío-El Palmar) of 42 km at 66 kV with a capacity of 50 MVA, with a planned cost of EUR 103 million [
29]. Consequently, many other projects are needed, whether for the installation of renewable generation sources, storage facilities, or interconnection grids between islands.
Various applications throughout the world have proven that electrical connections between islands are useful, since they provide several advantages, such as reliability, greater sustainability, cost reduction, and increased efficiency [
30,
31]. Examples of these implementations are in Greece, the Philippines, or Ecuador, and in Spain (between the peninsular territory and the Balearic Archipelago, as well as the abovementioned interconnection in the Canary Archipelago between Fuerteventura and Lanzarote). In fact, there are also some studies that analyze the Canary Islands themselves [
30,
32], but these usually deal with the interconnection between two islands and/or without making future estimates, although there are some studies that speak of the feasibility of interconnections between the islands of the archipelago [
33,
34]. This means less energy (forecasts of almost double the demand by 2040), with less variability (now generation is almost 90% based on fossil fuels, which allows for easy regulation), and only between two islands (approximately 75% of the demand but with the simplest grid, more concentrated generation, and storage, etc.).
To close this preliminary part, it can be pointed out that the research tackles the goal of a zero-emission generation system based entirely on renewable energies, meeting both technological and economic criteria, and achieving 100% demand coverage, given the need for high system reliability in the case of an autonomous network. This study addresses the energy autonomy of an archipelago of considerable size, which is in a relatively dispersed layout with seven different islands. Population and energy demand has also been considered with the added difficulty of considering long-term future demand forecasts. The islands have a total surface area of 7472 km
2, will have around 2,500,000 inhabitants by 2040, and an electricity demand estimated at up to 16–18 TWh per year also by the year 2040 [
25].
To address the present study, we used the software developed by the National Renewable Energy Laboratory (NREL), called the Hybrid Optimization of Multiple Energy Resources (HOMER) [
35]. The method followed by the software is mainly economic, since the code estimates the optimal size of a system based on the investment to be made, the LCOE (levelized cost of energy), and the amortization based on the energy sources to be installed [
35]. The HOMER software has been extensively tested in the simulation of hybrid renewable systems, with several generation sources and storage technologies all around the world [
36,
37]. These research works include some studies related to the Canary Archipelago, but all deal with problems of a minor nature, such as for a single island, or at most for the two largest islands [
21,
30,
38].
In order to complete the previous objectives,
Section 2 describes the current electricity supply of the Canary Archipelago.
Section 3 focuses on the methodology developed to carry out the current analysis. The contextualization of the problem is discussed in
Section 3, and the three scenarios under consideration are described.
Section 4 below briefly describes the characteristics and the demanded information required for all the systems necessary to make the simulations within the desired horizon, both for generation and storage technologies. The principal results of the simulations conducted are presented in
Section 5, with the corresponding analysis and discussion.
Section 6 is dedicated to summarizing the conclusions of the current study in terms of the generation system in the three scenarios under consideration. Some possible future work is also mentioned in this section.
3. Methodology
In order to carry out the code estimations, a rigorous methodology has been followed, which includes a detailed introduction of the entered information required to execute the simulation and a scheme of the steps used, as shown in
Figure 1. Among the necessary input data, the following can be mentioned: annual information on the energy demand to be included or, for future estimations, their forecasts; technical and cost information on the generation system to be considered (in the case of the current study, wind and photovoltaic power plants); technical information and cost of the storage system (mega-batteries and reversible pumping); the energy resources available for each generation system (wind and solar resource available at the selected sites); other additional economic data (for example, the annual interest rate and the useful life of the project). Based on all these data as input in the HOMER software, the most efficient combination of these generation systems can be estimated, so that all the energy demanded by the system is supplied. In particular, the nominal power to be required, the generation of power from each system, and the necessary storage capacity, etc., are determined. Economic information, such as LCOE, initial capital, net present cost (NPC), payback, and internal rate of return (IRR), are also provided by HOMER. The selection criteria in the methodology are the economical ones previously mentioned but keeping CO
2 gas emissions at zero.
Economic criteria implicitly imply a trade-off between the sizing of the generation and storage facilities to satisfy the demand needs. Since the reached solution is the one in which the cost is the lowest, this largely means that the size of the system is as small as possible while also always covering the energy needs. In other words, the optimum point between the oversizing of the generation and storage systems is reached. To make these estimates, economic data for the year 2019 have been used as the background, as it is assumed that the cost variation of the technologies used by 2040 will remain constant, which a priori should be true, since the technologies used are all at a fairly high level of maturity. The methodology has been tested in two distinct scenarios, both for the year 2040. These are coverage of the demand in each island separately and of an aggregated electric system for all of them, always considering the electrification of the economy.
The HOMER software performs simulations of the operation of each system analyzed through an energy balance at every time step defined in the code, the most common being the hourly balance, which is the one used in the current study. For this, it compares the energy demanded at each time step with that which can be distributed by the generation system under analysis, as well as how to operate the generators and whether it is required to make use of the batteries. Subsequent to simulating all system configurations, HOMER provides a list of feasible systems sorted by net present cost (NPC). The software displays a list of feasible solutions that meet all requirements. The global optimal solution tops the list, although other options can also be considered. For instance, other criteria can be considered, such as lower total installed power, minimization of energy wastages, the weighted combination of several of these factors, etc. In this case, as is usually carried out by other researchers [
21,
30,
35,
38,
40], the global optimum has been considered as the better solution.
The uncertainties that must be considered to properly estimate future demands, in this case for the year 2040, will be associated with the evolution of electricity demand, electricity costs, wind/solar resources, and storage technologies. For this, different possible future scenarios have been contemplated and evaluated based on different aspects, such as the uncertainty in the evolution of demand, the degree of electrification of the economy, the degree of penetration of the electric vehicle, the possibility of implementing policies demand management, favoring self-consumption, etc. Finally, the optimal solution is provided for what is considered the most likely future scenario, which is fundamentally based on a work carried out recently focused on the Canary Archipelago itself [
41] and on the Canary Islands Government [
42,
43,
44]. To quantify the cost savings of electricity itself that would be achieved with this autonomous and totally renewable system, the current cost of electricity has been considered (that of the year 2019 has been taken, as it is a value not affected by the strong existing uncertainties that have multiplied the price by more than double), that is, it has been conservative, since it is not expected that the future cost of electricity will return to the figures considered, so the feasibility obtained from the alternative generation systems would still be elderly. The savings due to the elimination of greenhouse gas emissions have also been considered; for this the “cost per ton of CO
2” emitted into the atmosphere has been considered (also taken from the emission figures for the year 2019). Wind and solar radiation resources could also present uncertainties, but for the current analysis period of about 20 years, no substantial changes are assumed, and the typical values (from the Global Wind Atlas and PVGIS) are considered to be representative
3.1. Scenario of Autonomous Islands Generation
In this first scenario, the implementation of the necessary inputs for the software will be carried out for each of the archipelago islands separately and for the most probable scenario in the year 2040. On each island, the particular conditioning aspects for the different technologies are taken into account. As anticipated above, this scenario is based on the estimates carried out in different studies for the Canary Islands themselves; specifically, an in-depth analysis focused on the Gran Canaria Island is of particular interest [
25,
41,
42]. This scenario consists of the total electrification of the economy with the implementation of strong efficiency measures and favoring self-consumption. Thus, in the end, translates into an increase of approximately 100% compared to the current values of electricity demand.
Figure 2 shows the forecasted aggregated value of all the islands while considering that, by 2040, the percentage of consumption of each island will remain approximately constant to the current one over these years, so it is possible to obtain that percentage for each of them.
The estimates of each island have been carried out by considering that the average hourly demand will have the same shape as the current one, with shape changes only being caused by the distribution resulting from the recharging of electric vehicles. Then, the contribution of electric vehicles has been estimated through the weighted ponderation of six profiles of electric vehicle recharge methods (homes and public roads, workplaces, hotel parking, shopping centers, and regular recharging points) for each island [
42]. For the remaining demand, the current hourly demand profiles have been taken, but broken down by sector (residential, commercial, industrial, public administration, lodging and other uses) [
43] and multiplied by the factor that contemplate the estimated increase until 2040. In other words, it has been considered that the current profiles and consumption proportions of each island are maintained, and a multiplying factor has been applied to include the economic growth until 2040. As an example,
Figure 3 shows the shape of the estimated demand curve of an actual typical January day and for the year 2040 in this scenario of total electrification for the island of Lanzarote (for the remaining islands, a similar procedure has been carried out for the estimation, and the shape is also quite similar in all of them).
3.2. Scenarios of Interconnected Grid between Islands
In the current case, the calculations are made for the aggregate system, but the most suitable locations for the different generation and storage systems have been sought. That is to say, for example, for the installation of the pumping systems, the appropriate locations existing on the different islands will be used, while for the wind turbines two locations have been sought close to the islands with the highest consumption so as not to have to oversize the interconnection between the two. In this scenario, the aggregate values shown in
Figure 2 are used, as well as the aggregate demand curve. As an example,
Figure 4 shows the aggregated curves for a typical January day for the current and forecasted 2040 demand values.
In relation to the interconnections, these have been dimensioned in such a way that they can allow energy exchanges between islands under any circumstances, as seen in
Table 2. To this end, for the islands located at the ends of the grid (Lanzarote and La Palma) the interconnection allows the passage of peak demand plus a 10% safety margin to allow that no extensions should be made (the demand in 2040 will be stabilized, as in fact it is currently, but it will change fundamentally due to the contributions of electric vehicles). For the island of La Gomera, the interconnection currently in the design phase has been assumed, since it has a transport capacity greater than the peak demand foreseen for 2040, plus the 10% safety margin. For the island of Fuerteventura, the interconnection has been sized so that it can transport its demand plus that of Lanzarote. For the two central islands, those with the maximum demand, the interconnection is sized to be capable of transporting the highest demand peak plus the 10% safety factor; it is not oversized to a greater degree, given that these two islands are where the bulk of both the generation and storage systems will be located.
In the estimations of the interconnection lengths between the different islands, the cable distances buried in the land from the substations to the coast have been considered, as well as the length of the underwater route, taking into account the particular underwater orography of each case. As an example, the profile of the interconnection between the islands of Tenerife and La Gomera is shown in
Figure 5. This figure shows that the distance between both islands is around 29 km, but the additional distance to bring the cable through the seabed brings the distance up to about 35 km, which together with the distance to reach the substations at both ends leads to the total length of 42 km, as shown in
Table 2. Therefore, the same procedure has been followed for the remaining proposed interconnections between the islands.
Qiblawei et al. [
30] have proposed a regression model of the cost of interconnections using Bloomberg New Energy Finance (BNEF) data. In the fitting process, they tested several parameters, such as rated power, distance, voltage, and year of commissioning; finally, the authors concluded that the combination of power and distance produces good results, with an R
2 = 0.85. The authors estimated that one has a maximum error of 27.8% of the costs predicted by the correlation versus the actual costs based on the cost analysis of 52 HVDC projects. The proposed expression is as follows:
where
is the cost expressed in million USD (applied exchange ratio 1 USD = 1 EUR),
is the maximum power transported through the interconnection in MW, and
is the interconnection length expressed in km. According to Gils and Simon [
34], the installation has a life service of about 40 years and annual fixed operating costs equivalent to 0.6% of the initial investment.
The interconnections between islands are shown in
Figure 6, while
Table 2 summarizes the main characteristics of the analyzed interconnections between the Canary Islands. Finally, the total losses associated with the interconnection must be considered. These are basically composed of the converter substation losses plus those of the cable itself, which are set at around 1% and 0.4%/100 km, respectively [
30]. Gils and Simon [
34] assume DC transmission losses of 0.27%/100 km in the marine cables, plus additional losses of 0.7% in the conversion from an AC–DC and DC–AC. Finally, assuming that less than 25% of the total energy generation needs to be transported to other islands (most of the generation is produced, almost equally, between the two islands of greatest importance, namely Gran Canaria and Tenerife), then the total energy losses are around 0.25–0.35% of the generated electricity.
The first of the interconnected scenarios consists of considering the aggregate demand, generation, and storage of all the Canary Archipelago islands, so that the software can be used to estimate the optimum mix, highlighting that the costs of the interconnections and their associated losses have been taken into account. In the second interconnected scenario, the different characteristics of the systems used have been analyzed and optimized; for example, among other considerations, the designs of the pumping and biomass plants have been reviewed, and the possibility of increasing the installed capacity of solar PV above the available roof area has been explored. All these aspects will be described and discussed in more detail in the results section.
3.3. Simulation Inputs
The generation sources considered are wind and solar photovoltaic, in combination with the unavoidable use of storage (battery systems and reversible pumping storage have been considered) to achieve the optimal design of the generation system. The final results are a compromise solution between cost reduction, excess energy minimization, and affordable land occupation, always covering the demand. In the current cases, the system is conceived to supply 100% of the Canary Archipelago’s energy demand. Additionally, the previously mentioned criterion of zero CO
2 emissions has also been met, since the nature of all used technologies leads to zero emissions during their operational lifetime. However, if the emissions during their whole life cycle are considered, then there are emissions, in any case, much lower than those currently existing.
Figure 7 presents a scheme of the sources analyzed to cover the energy demand for the considered scenarios.