Techno-Economic Analysis of a Hydrogen-Based Power Supply Backup System for Tertiary Sector Buildings: A Case Study in Greece

Abstract

In view of the European Union’s strategy on hydrogen for decarbonization and buildings’ decarbonization targets, the use of hydrogen in buildings is expected in the future. Backup power in buildings is usually provided with diesel generators (DGs). In this study, the use of a hydrogen fuel cell (HFC) power supply backup system is studied. Its operation is compared to a DG and a techno-economic analysis of the latter’s replacement with an HFC is conducted by calculating relevant key performance indicators (KPIs). The developed approach is presented in a case study on a school building in Greece. Based on the school’s electricity loads, which are calculated with a dynamic energy simulation and power shortages scenarios, the backup system’s characteristics are defined, and the relevant KPIs are calculated. It was found that the HFC system can reduce the annual CO₂ emissions by up to 400 kg and has a lower annual operation cost than a DG. However, due to its high investment cost, its levelized cost of electricity is higher, and the replacement of an existing DG is unviable in the current market situation. The techno-economic study reveals that subsidies of around 58–89% are required to foster the deployment of HFC backup systems in buildings.

1. Introduction

In order to mitigate climate change, the European Union (EU) has set an ambitious target of a 55% reduction in CO₂ emissions by 2030 and carbon neutrality by 2050 through the European Green Deal [1]. In this framework, the building sector can play a significant role, as it is characterized by high energy consumption and emissions of greenhouse gases. In Europe, buildings consume 40% of the total energy and emit 36% of the greenhouse gases [2]. Therefore, to reduce the environmental footprint of buildings, the EU has issued directives regarding the improvement of the energy efficiency of buildings [3,4]. In those directives, the importance of the reduction in CO₂ emissions from buildings and the consumption of energy from renewable energy sources (RESs) is highlighted [4]. In the case of Greece, these directives were adopted in the national regulation with laws regarding the energy performance of buildings (refer to L.3661/2008 and L.4122/2013). Moreover, a challenging target of the member states’ national long-term strategies for renovating the building stock and the EU’s building Renovation Wave strategy is the decarbonization of the existing building stock [5,6]. This target is transposed in the national long-term renovation strategy 2020 of Greece for renovating the stock of public and private buildings [7]. Therefore, the use of new technologies that could contribute to the reduction in the emissions from buildings, even if they are not directly related to the improvement of the buildings’ energy performance, could be promoted in the near future, both at the national and at the European level, to contribute to the achievement of the decarbonization targets for buildings.

Furthermore, to achieve the emissions reduction targets of the European Green Deal, hydrogen is expected to play an important role by contributing to the decarbonization of several sectors, such as the industry, transport, power and building sectors [8]. According to the EU’s strategy on hydrogen, renewable hydrogen production is expected to gradually increase, and renewable hydrogen is expected to be integrated in several sectors. In that way, hydrogen is expected to become an integral part of the energy system by 2050 [8]. Renewable hydrogen is defined as the hydrogen produced from electrolysis using electricity from RESs [8]. Additionally, according to the EU’s REPowerEU plan, an increase in the production and imports of renewable hydrogen is planned, as a means of decarbonization, by substituting natural gas, coal and oil in specific industries [9]. Therefore, hydrogen technologies, including hydrogen fuel cells, are expected to be deployed at a large scale in the future [8]. Last but not least, due to the expected increase in the efficiency of and reduction in the cost of electrolyzers and a decrease in the cost of solar electricity, renewable hydrogen is expected to be available at lower prices in the future and to be competitive with hydrogen from fossil fuels [8,10]. This will further contribute to the more extensive use of hydrogen in the energy system.

In the case of Greece, hydrogen is expected to play a significant role according to Greece’s national long-term strategy for 2050, as the production of renewable hydrogen and its penetration in several sectors are expected to increase significantly by 2050 [11]. Regarding national legislation, L. 4439/2016 transposes the European Directive 2014/94 concerning specifications for hydrogen refueling stations, while, so far, there is no legislation on the use of hydrogen in buildings. Finally, regarding the renewable hydrogen market and implementation of renewable hydrogen applications, both are still in their infancy in Greece.

In this framework, the use of hydrogen technologies, and specifically fuel cells, in buildings could be promoted both at the national and the European level. This could contribute both to the increase in the penetration of hydrogen and fuel cells in the energy system and to the achievement of the building sector’s decarbonization targets. Moreover, according to the EU, public bodies’ buildings can play an exemplary role in the integration and demonstration of new technologies [12]. Therefore, tertiary sector public buildings could play an important role, boosting the use of hydrogen technologies in buildings.

Hydrogen technologies, and specifically hydrogen fuel cells, have the ability to decarbonize several sectors due to hydrogen’s very low or zero CO₂ emissions [8]. Hydrogen fuel cells can play a significant role in the decarbonization of the transport sector, together with electrification, especially in the case of heavy-duty vehicles or buses [8,10,13]. Railway applications with hydrogen fuel cells are also possible [10,13]. Moreover, hydrogen fuel cells can be used in electricity storage applications, combined with electrolyzers and hydrogen storage. These systems can store large amounts of electricity for long periods and can contribute to the balancing of RES electricity production [8,13,14]. Another application of hydrogen fuel cells is combined heat and power (CHP) applications in buildings, aiming at the decarbonization of heat production [8,14]. CHP systems offer increased efficiency and reduction in energy consumption and emissions [14]. Last but not least, hydrogen fuel cells can be used as backup power supply systems in several applications, such as buildings [13], telecommunication infrastructure [13,15], data centers [10] or other infrastructure. An alternative to the latter is the operation of fuel cells as power generation systems for off-grid infrastructure, where fuel cells are often combined with electrolyzers and hydrogen storage [13].

The installation of backup power supply systems for the provision of power and increased resilience in case of power shortages, either planned or caused by exceptional events such as natural disasters ([16]) or cyberattacks ([17]), is an issue concerning several infrastructures, including buildings, where the loss of power is important. Therefore, the specification of an appropriate system that has the necessary characteristics is important. Usually, conventional generators, such as diesel generators (DGs), are chosen over batteries due to their ability to produce backup power for long periods without increasing the capital expenses significantly [18]. However, these generators are also characterized by high emissions and noise pollution. The former is an important drawback in the framework of the decarbonization of buildings, while the latter can a problem in special-purpose buildings such as hospitals. To tackle the aforementioned issues, hydrogen fuel cell (HFC) backup systems have been proposed as another alternative to conventional generators. HFCs have numerous advantages compared to conventional generators, such as lower emissions of CO₂ and other pollutants, lower maintenance and operation cost and lower noise emissions [18,19]. At the same time, they exhibit similar characteristics to conventional generators, such as long runtime capability and high resilience [18]. Therefore, HFC systems arise as a pertinent technology for backup applications. The use of HFCs as backup systems has been studied in the literature. In Ref. [19], the use of a fuel cell system for the provision of backup power in the telecommunication industry was studied. A polymer electrolyte membrane (PEM)-type fuel cell was considered, and the system configuration was analyzed in the study along with the system’s costs, revealing excellent techno-economic feasibility. Extensive comparisons of an HFC system with DGs and batteries specifically for purposes of backup power provision were presented in Ref. [18] regarding a telecommunication application. Although the DG was found to have the lowest cost, it is accompanied by high annual maintenance cost as well as high noise and emissions. Even though fuel cells are admittedly more expensive, financial incentives were proposed as a way to increase cost competitiveness in relation to the DG. Concerning batteries, they were found to have the highest cost compared to the other systems, as the duration of the system’s yearly runtime increased. In Ref. [20], a fuel cell system, a DG system and their combination were studied as a backup system for the case of an energy hub in a sports facility, consisting of photovoltaics (PV) and wind energy. The fuel cell system was found to have the lowest emissions of CO₂ while also achieving a significant reduction in other pollutants such as SO₂, CO and NO_x. The combined DG and fuel cell backup system provided a good compromise of investment cost (IC) and emissions reduction compared to the case of diesel-only or fuel cell-only systems without increasing the cost of energy significantly. In Ref. [21], the optimal sizing of a backup system consisting of a fuel cell, a battery or a combination of them for a telecommunication system was investigated. It was found that the fuel cell system had a higher cost than the combined battery and fuel cell system, while the system consisting only of a battery was proved to be techno-economically prohibitive. One issue of backup systems in general is the few hours of utilization per year. Regarding this, in Ref. [22], the further utilization of a fuel cell, which is designed as a backup system for telecommunication infrastructure, for the provision of grid ancillary services was studied as a way of utilizing the system for more hours per year. The study concluded that it is possible to achieve profit, but the increased hydrogen delivery cost and the reduction in the fuel cell’s lifetime due to its longer utilization could offset that profit. Furthermore, the operation of fuel cells as backup systems in remote applications has been studied. For example, in Ref. [23], the operation of a fuel cell in a remote residence powered by PVs was studied. In that case, the production of hydrogen took place on-site, with the use of an electrolyzer, a pump and a hydrogen storage tank. Similarly, the use of fuel cells in a remote telecommunication infrastructure, which is powered by PVs and includes batteries for short-term storage, was studied in Ref. [24], where hydrogen was produced and stored on-site. Regarding building applications, the sizing of a backup system consisting of a fuel cell and battery for a university building was studied in Ref. [16]. The study concluded that a combined fuel cell–battery system is cheaper than a battery-only system, especially as the outage duration extends. In this study, the optimal supply of hydrogen to the building was also investigated. It was concluded that the external hydrogen supply is more adequate because the installation of an electrolyzer and reservoir tank to produce hydrogen would make the system economically unviable. The combination of fuel cells with PV and batteries in a backup system was studied in Ref. [25] to provide the necessary loads to telecommunication infrastructure during blackouts. The fuel cell system was proposed due to its long runtime capability along with low emissions and cost of maintenance. The proposed system was able to provide power even during blackouts with a duration of up to 30 h, thus increasing the resilience of the infrastructure.

The investigated aspects of HFCs found in the studies cited above are summarized in

Table 1. Summary of investigated aspects and outcomes from the identified studies.

Investigated Aspect of HFCs	Outcome	References
Technical feasibility of HFCs as backup system in interconnected applications	The HFC can operate in backup power provision applications either as a standalone backup system or in combination with other systems such as PVs or batteries.	[16,18,19,21,22,25]
Technical feasibility of HFCs as backup system in remote applications powered by RESs	The HFC can operate as a backup system in applications powered by RESs	[20,23,24]
Economic performance of HFCs	The HFC system has a higher IC and lower operation and maintenance cost than a DG.	[16,18,19,20,21,22,25]
Environmental performance of HFCs	The HFC has important environmental benefits compared to a DG, as it reduces CO2 and other pollutant emissions	[18,20,25]

. As can be seen, although the technical feasibility and the economic and environmental performance of the HFCs have been addressed in the literature, no in-depth comparison of the economic performance of the HFCs with the economic performance of DGs in terms of the cost of power supply for backup applications over the systems’ lifetime was found. The same is true regarding the assessment of the economic viability of the replacement of existing DGs with HFCs in the current economic environment. Finally, given the presented policy framework regarding the promotion of the use of hydrogen, the issue of identifying the required policy measures for fostering wider use of HFCs as backup systems has not yet been thoroughly addressed in the literature.

Based on the discussed studies and the research gaps that were identified, the key deliveries of the current study are:

• To compare the cost of the backup power supply of an HFC system with a DG system over the systems’ lifetime.
• To conduct a techno-economic analysis of the replacement of existing DG systems with HFCs.
• To investigate the conditions under which (a) the HFC system can achieve a better economic performance than the DG and (b) the replacement of existing DG systems with HFCs would be viable.

In that sense, the paper aims at providing insights about measures, such as subsidies, that could be employed in order to promote the use of HFCs for backup power supply and that would in general promote the use of hydrogen in building applications. To that goal, a techno-economic analysis of an HFC system for backup power supply is conducted for a case-study building. Furthermore, the performance of the HFC system is compared to the performance of a DG, both from an environmental and economic point of view, and the levelized cost of electricity (LCOE) of each system is calculated. Finally, the replacement of an existing DG with an HFC backup system is assessed through the calculation of relevant KPIs.

2. Materials and Methods

The main objective of this study is to compare the operation of the HFC and DG backup systems and to investigate the techno-economic feasibility of the replacement of a conventional DG with an HFC backup system. The case study that is chosen is the potential integration in a school building. A dynamic building energy simulation model was developed for the calculation of the hourly variation of the loads of the building for a year. In that way, the electricity loads of the building could be calculated on an hourly basis, which is essential for assessing backup power supply systems. Then, several scenarios of power shortages were assumed, and in combination with the electricity load calculations, the necessary characteristics of the backup system, such as the required size and required stored fuel, were defined for each scenario. Initially, the operation of the two systems was compared in terms of energy consumption and emissions, and then we calculated the emissions reduction from the replacement of a DG with the proposed system. Additionally, the cost for each of the backup systems was estimated for each scenario and compared. A techno-economic analysis was conducted for each scenario, considering an investment for replacing an existing DG backup system with an HFC system with the same characteristics, and the key performance indicators (KPIs) for the assessment of this investment were calculated. Finally, the conditions that would make the HFC system competitive with the DG and the conditions that would make the replacement of the DG with an HFC viable were investigated by means of a sensitivity analysis.

2.1. The Building Case Considered

A common building type in the tertiary sector is educational buildings. Especially at the local level, school buildings are dominant among the ones owned or managed by local public authorities such as municipalities. That said, and also considering the exemplary role of public buildings regarding the testing and demonstration of innovative energy technologies (see Ref. [12]), educational buildings are perhaps the most suitable substrate to test the feasibility of novel solutions. Moreover, school buildings are also interesting because energy and environmental novelties have a direct impact on the environmental awareness of the younger generation. Serving the aforementioned purpose, a school building was chosen as a case study herein. More specifically, the building presented in Refs. [26,27] is used, since in the references, a thorough technical presentation of the building together with dynamic energy simulation results validated over experimental data is provided. The building is an elementary school building located in Athens, Greece. It was built in 1984, and its design is typical for Greek schools. It has two floors—the ground floor consisting of classrooms, offices and computer rooms and the first floor consisting only of classrooms. Both floors share the same heating system. Regarding cooling, an air conditioning (AC) unit is installed only in the computer room, while in the rest of the rooms, ceiling fans are installed for cooling purposes. The building has the following characteristics:

• Total area of the building: 2169.6 m².
• Building geometry: a common ATHINA-type geometry for Greek school buildings, based on Ref. [26]
• Building envelope characteristics: the thermophysical and optical properties of the opaque and glazed surfaces were taken from [26].
• Building use: school building.
• Number of occupants: a typical value for school buildings in Greece was considered, equal to 0.5 occupants per m² of floor area, based on the Greek national technical guidelines for the calculation of the energy performance in buildings [28].
• Operating schedule:

  ○

  From 9 September to 15 June: 8 h/day (08:00–16:00), 5 days/week. Some of the classrooms are not used after 14:00. However, the school is closed during Christmas and Easter holidays, i.e., 15 days at the end of December and the beginning of January and 15 days in April.

  ○

  From 1 August to 9 September: the school remains closed.

• The systems of the buildings and their characteristics are presented in

  Table 2. Characteristics of heating, cooling and lighting system and devices of the building.

  System	Subsystems	System Properties	Operation Schedule
  Heating	Central natural gas boiler, radiators in each room, distribution system consisting of two pumps	Boiler heating capacity: 220 kW Boiler efficiency: 89.8% Setpoint: 20 °C	From 1 November to 15 April, from 07:30 to 11:30
  Cooling	AC split unit (only in the computer room)	AC cooling capacity: 6.7 kW AC energy efficiency ratio: 2.6 Setpoint: 26 °C	From 15 May to 30 July, during working hours
  Lighting	24 fluorescent lamps in each room	36 W per lamp	All year, during working hours
  Fans	2 ceiling fans in each room	82 W per fan	From 15 May to 30 July, during working hours. They operate only if the ambient temperature is above 29 °C
  Electrical devices	Computers and other devices	0.5 W/m2 of floor area for each room	All year, during working hours

  .

Finally, the building is connected to the electricity grid and does not have any installed RESs.

2.2. Dynamic Energy Analysis of the Studied Building

The dynamic energy simulation model for the building was constructed in the DesignBuilder v.4.2 software, as presented in Ref. [26]. The DesignBuilder software is based on the scientifically validated EnergyPlus calculation engine and has been extensively used in building energy simulation applications [29]. A visualization of the building’s model is presented in Figure 1.

The characteristics of the building described in the previous subsection were processed according to the Greek national technical guidelines for the calculation of the energy performance in buildings (Ref. [28]) for the extraction of the proper input conditions, which were fed in the simulation model later on. Climate data for the area of Athens were used from a weather file (.epw file) available in the software. The simulation was conducted for producing the following results on an hourly basis:

• Energy consumption of the building per end use (heating, cooling, lighting and devices);
• Energy consumption of the building per energy carrier (electricity and natural gas).

From these results, the electricity consumption results are relevant for the analysis of the power supply backup system. This hourly energy simulation of the building was necessary as both the peak electricity load and the daily variation of the electricity load must be known to properly size the backup system.

2.3. Description of the Considered Backup Systems

The proposed HFC-based backup system consists of the following main components:

• Fuel cell;
• Battery (internal);
• Hydrogen tank;
• DC/DC converter;
• DC/AC inverter;
• Controller.

A schematic representation of the system is shown in Figure 2. It should be mentioned that the representation is indicative and corresponds to a typical HFC-based backup system configuration [19].

The most appropriate type of fuel cell for this system is the PEM fuel cell [19,30]. As for the battery, it is a necessary component due to the required startup time of the fuel cell. During the fuel cell’s startup, the internal battery provides the required energy to the load and to the fuel cell, which requires energy for its startup. The battery capacity is defined so that it is enough to provide the required energy during the startup of the fuel cell [19,21]. The fuel cell is connected to the battery through a DC/DC converter. When the fuel cell reaches a steady-state operation, one option is that the battery is charged by the fuel cell, and when it is full, the controller decouples it from the fuel cell and the load so that the fuel cell directly provides the required power to the load. Alternatively, the battery can continue to provide power to the load, and it is charged by the fuel cell [19]. In the case of low loads, the fuel cell operates with low efficiency, or it is not able to operate at all. In this case, the battery can directly supply the load, while the fuel cell can be operated only at specific time intervals, so that the battery remains charged [19]. It must be mentioned that regardless of the operating strategy that will be chosen, the calculations within the scope of this paper are not affected. Finally, since the system produces DC electricity, an inverter is necessary in the case of AC loads, as in the case of buildings.

As for the hydrogen supply, it was assumed that the required hydrogen is bought and stored in the facility and not produced on-site, considering that the required stored quantities of hydrogen are relatively low. The on-site production of hydrogen through the integration of an electrolyzer could be possible; however, as discussed in Ref. [16], this would increase the IC of the system significantly. Regarding the storage of hydrogen, hydrogen storage tanks are considered, where hydrogen is stored at a pressure of 200 bar. As for the tanks, there are two options [18]:

• Leasing hydrogen tanks, which can be replaced.
• Purchasing hydrogen storage tanks, which can be refilled.

Usually, the first option is chosen in the case of small quantities of stored hydrogen. As the quantities increase, the rental becomes more expensive, and therefore, the purchase of a tank becomes a more viable option [18].

As for the DG backup system that is also considered, a typical DG is assumed, and the required quantity of diesel fuel is assumed to be stored in a fuel tank.

Regarding the sizing of the backup system, the nominal power of the system was calculated considering that the backup system must always be able to supply the building’s electricity loads. Thus, the nominal power of the backup system must be higher than the peak electricity load that occurs in the building. Therefore, from the results of the dynamic energy analysis of the building, the peak electricity load of the building was specified, and the nominal size of the backup system was defined by applying a safety factor of 1.1 on the peak load. Furthermore, regarding the size of the fuel storage (hydrogen or diesel fuel), the energy that is consumed annually must be stored on-site in the case of a power shortage. Therefore, the considered shortage duration is important for the definition of the storage size. For that, different scenarios were defined (see next subsection). To calculate the required stored energy at each scenario, engineering practice dictates that the worst case in terms of energy demand should be taken into account, i.e., for each power shortage duration scenario, the respective number of hours with the highest electricity demand of the year was specified based on the results from the dynamic energy analysis of the building. Finally, considering the HFC’s and DG’s efficiency, the storage system size was defined such that it could store the specified amount of energy in the form of hydrogen or diesel fuel.

2.4. Power Shortage Duration Scenarios

The operating scenarios for the backup system were defined based on different shortages’ durations per year. The shortages that are assumed could take place either more than once and last shorter or just once and last for a period equal to the presumed duration. Initially, the electricity supply continuity indicators, namely the system average interruption duration index (SAIDI) and system average interruption frequency index (SAIFI) were considered. The former represents the average total period in a year that a customer experiences a power shortage, while the latter represents the average number of shortages that a customer faces in a year. These indicators are calculated for the case of Greece from the Hellenic Electricity Distribution Network Operation [31]. In Ref. [32], data for Greece for the years 2012–2016 are presented, related to the planned and unplanned power shortages, including exceptional events. These are presented in

Table 3. Electricity supply continuity indicators for Greece from 2012–2016 [32].

Year	SAIFI (Interruptions per Customer)	SAIDI (Minutes per Customer)
2012	3.3	299
2013	3.2	289
2014	2.8	258
2015	2.9	248
2016	2.8	244

.

From Table 3, it can be concluded that the indicators are improving over the years. In the definition of the power shortage scenarios, the case of 2012 is considered the worst case among the available data. The annual shortage duration in this case is 299 min per year. Since this represents the average value for the whole country, more scenarios were defined for longer shortage durations. A scenario of shortage duration of 24 h per year and a scenario of 72 h per year were considered, with the latter representing a scenario of a major shortage, which is likely to occur, for example, under extreme weather conditions. The considered scenarios are presented in

Table 4. Power shortage duration scenarios considered.

Scenario	Duration of Power Shortage (min)
1	299
2	1440
3	4320

.

2.5. Techno-Economic Analysis Methodology

A techno-economic analysis regarding the operation of the HFC and DG system for the backup power supply was conducted, and the levelized cost of electricity and the emissions of the two systems were calculated and compared. Then, an economic feasibility study of the replacement of a conventional DG backup system with an HFC system was conducted in terms of the net present value (NPV), internal rate of return (IRR) and payback period (PBP). To calculate those indicators, initially, the results from the building energy simulation were processed, and the requirements for the backup system were defined. Then, for the HFC and DG backup systems, the indicators were calculated as follows:

• Emissions reduction

The use of an HFC-based system contributes to a significant reduction in emissions compared to a DG. The reduction in CO₂ emissions depends on the manufacturing of hydrogen, i.e., if the hydrogen comes from RESs. For renewable hydrogen, the emissions reduction is significant. Furthermore, the emissions of hydrocarbons (HC) and nitrogen oxides (NO_x), particulate matter (PM) and carbon oxide (CO) that come from DGs can be eliminated. The emission factors for the two systems are presented in

Table 5. Emission factors from diesel engines and renewable hydrogen.

Emission Factor (g/kWh)	Diesel Engines, 0 < p < 8 kW 1	Diesel Engines, 8 < p < 19 kW 1	Diesel Engines, 19 < p < 56 kW 1	Renewable Hydrogen
CO2 (g/kWh) (Data from [33])	267	267	267	0
HC + NOx (g/kWh) (Data from [34])	7.5	7.5	4.7	-
PM (g/kWh) (Data from [34])	0.4	0.4	0.015	-
CO (g/kWh) (Data from [34])	8	6.6	5	-

¹ p refers to the power of the diesel engine.

. For DGs, the standards set by the European Commission for non-road mobile machinery were used, which also depend on the engine’s power range. In Table 5, the emission factors for each power range are presented.

• LCOE for the two systems.

The LCOE is calculated from the equation below in Equation (1) [35]. It represents the cost of each kWh of electricity that is produced from the system over its lifetime.

$$\mathrm{LCOE}=\frac{\mathrm{IC}+{{\displaystyle \sum }}_{\mathrm{t}=1}^{\mathrm{LT}}\frac{{\mathrm{Fuel}}_{\mathrm{cost}}+\mathrm{O}\&{\mathrm{M}}_{\mathrm{cost}}}{{\left(1+\mathrm{r}\right)}^{\mathrm{t}}}}{{{\displaystyle \sum }}_{\mathrm{t}=1}^{\mathrm{LT}}\frac{{\mathrm{E}}_{\mathrm{gen}}}{{\left(1+\mathrm{r}\right)}^{\mathrm{t}}}}$$

(1)

where:

• IC: investment cost, in EUR.
• LT: lifetime of the system, in years.
• ${\mathrm{E}}_{\mathrm{gen}}$: annual generated electrical energy in year t, in kWh.
• ${\mathrm{Fuel}}_{\mathrm{cost}}$: annual fuel cost of the system in year t, in EUR.
• $\mathrm{O}\&{\mathrm{M}}_{\mathrm{cost}}$: annual operation and maintenance cost of the system in year t, in EUR.
• $\mathrm{R}$: discount rate. The discount rate is set equal to 2%, as proposed in the official ministerial guide for conducting the energy efficiency plans for regional and municipal buildings in Greece [36].

• Cost reduction and calculation of KPIs regarding the replacement of a conventional DG-based backup system with an HFC-based system.

• Due to the HFC system’s lower operation and maintenance cost and its higher efficiency, a cost reduction in the annual operation and maintenance cost of the backup system is expected because of the replacement of the DG. This can be considered as an annual cash flow. By taking into account the IC of the HFC system, KPIs particularly related to the investment can be calculated, i.e., the following:

  ○

  PBP: it is calculated as the year in which the cumulative cash flow (which comes from the annual cost reductions) becomes positive.

  ○

  NPV: assuming a constant annual cash flow, it is calculated as follows:

  $$\mathrm{NPV}=-\mathrm{IC}+\mathrm{NCF}\times \frac{1}{\mathrm{r}}\times [1-\frac{1}{{\left(1+\mathrm{r}\right)}^{\mathrm{N}}}]$$

  (2)

  where:

  • $\mathrm{NCF}$: annual net cash flow, in EUR.
  • $\mathrm{N}$: lifecycle of the investment, in years.
  • A positive NPV value suggests a viable investment.

  ○

  IRR: The internal rate of return, which is the discount rate at which the NPV is zero. An IRR greater than the discount rate suggests a viable investment.

To calculate the KPIs that were mentioned, data regarding the techno-economic characteristics of the two backup systems are necessary. The realistic data adopted herein are presented in

Table 6. Techno-economic data for the HFC system.

Techno-Economic Characteristic	Value
Fuel cell IC	6000–7000 EUR/kW, if Pnom ≤ 3 kW (Data from [37])
	2500–4000 EUR/kW, if 5 ≤ Pnom ≤ 10 kW (Data from [15,16,19])
	2000–3000 EUR/kW, if 10 ≤ Pnom ≤ 100 kW (Data from [15,19])
Battery IC	200–400 EUR/kWh (Data from [38])
Inverter IC	200 EUR/kW (Data from [16])
Converter IC	283 EUR/kW(Data from [37])
Hydrogen storage cost	Hydrogen tank leasing: 100 EUR/year/tank (Data from [18])
	Hydrogen tank IC: 9 EUR/kWh (Data from [15])
Operation cost	0.03 EUR/kWh (Data from [39])
Maintenance cost	65–200 EUR/year (Data from [18,21,37])
Renewable hydrogen fuel cost 1	12 EUR/kg or 0.36 EUR/kWh
Fuel cell efficiency	50–60% (Data from [19,30,37])
Hydrogen lower heating value	33.3 kWh/kg
Hydrogen density	0.0141 kg/L t (at 200 bar and 25 °C)
System’s lifetime	15 years (Data from [18])

¹ Renewable hydrogen supply in retail is not available in Greece. Therefore, an estimation of the cost is given based on the retail price given in [40] and on a production cost of 4–9 EUR/kg [10].

for the HFC system and in

Table 7. Techno-economic data for the DG system.

Techno-Economic Characteristic	Value
Generator IC	800–1100 EUR/kW (Data from [18,41])
Fuel tank	6.5 EUR/kg (Data from market study)
Operation cost	0.01 EUR/kWh (Data from [39])
Maintenance cost	35 EUR/kW (Data from [41])
Diesel fuel cost 1	1.9 EUR/l t or 0.18 EUR/kWh
Generator efficiency	25% (Data from [18] 2)
Diesel fuel lower heating value	12.67 kWh/kg
Diesel fuel density	0.846 kg/L t
System’s lifetime	15 years (Data from [18])

¹ Average value in Greece from September 2022 to February 2023, according to [42], and ² also from specification sheets from manufacturers, e.g., Ref. [43].

for the DG.

Finally, in order to investigate the effect of a change in the parameters that are presented in Table 6 and Table 7 on the studied indicators, a sensitivity analysis was conducted. Among those parameters, two are the most likely to change in the future:

• The IC of the fuel cell: The production of fuel cells in larger volumes will result in a decrease in the cost. In IEA’s hydrogen technology roadmap, a reduction in the cost of the PEM fuel cells for stationary applications below 800 USD/kW by 2030 is set as a recommendation [15]. Moreover, in Ref. [19], a cost reduction of up to 61% and 58% is expected for 5 kW and 10 kW PEM fuel cells, respectively, for backup applications, if the production volumes increase up to 50,000 units per year. Apart from cost reductions, the IC could be decreased through subsidies.
• The cost of renewable hydrogen: A reduction in the cost of renewable solar electricity combined with a decrease in the capital cost and an increase in the efficiency of electrolyzers is expected to reduce the renewable hydrogen price from 4–9 USD/kg, which is the estimated price today, down to 1 USD/kg and below 1 USD/kg by 2030 and 2050, respectively [10]. Again, apart from cost reductions, the hydrogen cost could be decreased through subsidies.

In view of the above, a sensitivity analysis was performed by calculating the LCOE and the NPV for different combinations of the HFC systems in terms of the size of the storage IC and the renewable hydrogen cost. Through the analysis, the sensitivity of those indicators in the change in the IC and the cost of hydrogen was investigated. Finally, the sensitivity analysis was performed based on scenarios 1 and 3, as these are the scenarios with the least and the most operating hours, respectively (refer to Table 4). The comparison between the results for the two scenarios could allow the assessment of the effect of the hours of operation of the backup system.

3. Results

3.1. Building Energy Demand

The dynamic energy simulation provided a realistic view of the variation of the building’s loads on an hourly basis for a whole year. The simulated annual energy consumption of the building per energy carrier and end-use is as follows:

• A total of 22.48 MWh of electricity for lighting, cooling, fans and electrical equipment.
• A total of 30.69 MWh of natural gas for heating.

Adopting the CO₂ emission conversion factor for electricity and natural gas in Greece, which is equal to 0.42 kg/kWh [44] and 0.196 kg /kWh [28], respectively, the annual energy-induced CO₂ emissions of the building are:

• A total of 9442.15 kg of CO₂ due to electricity consumption.
• A total of 5830.81 kg of CO₂ due to natural gas consumption.

The results regarding the hourly variation of the electricity loads of the building are presented in Figure 3. It should be mentioned that the simulation model also provides the heating loads of the building, which is served by natural gas (see Table 2), and therefore, they are not included in Figure 3, as they are not related to the studied (electricity) backup system.

As seen in Figure 3, the lighting and devices’ loads are constant every day throughout the year. Therefore, the daily variation of the electricity load in winter is the same every day. On the other hand, in summer, the total electricity load is higher, and its variation is different from day to day due to the combined effect of the cooling and the fan loads, which depend on the weather conditions as described in Ref. [26]. According to the annual results presented in Figure 3, the highest loads occur in July, and more specifically from 9 to 26 July. This is expected since the highest ambient temperatures occur in this month, while no loads occur from 1 August to 9 September, as the school remains closed during that period.

In Figure 4, the daily variation of the electricity load for a typical winter day and summer day is presented. Since the electricity load in the winter is the same every day, one typical winter day is chosen (1 March) and the load variation is presented in Figure 4a. As for summer, in Figure 4b, the daily variation of the electricity loads of one of the days when the highest loads occur (22 July) is presented.

From Figure 4a, it can be seen that in winter, the load is constant for 6 h of the day, while it decreases in the last 2 h of the operation schedule, since some of the classrooms are not used after 14:00. From Figure 4b, it can be seen that in the presented summer day, the lighting and devices’ loads are the same as for the winter day. The cooling load is constant for 6 h of the day, and the cooling system operates at its maximum capacity. The load of the ceiling fans is very low compared to the other loads because in the previous study (see Ref. [26]), a rather high operation set-point for ceiling fans was imposed (i.e., 30 °C) as a result of interviews with teachers who explained that ceiling fans are rarely used, and only under extreme heat, for students’ safety reasons. In the last two hours of the operation schedule, the electricity loads of the building, including the cooling and fans’ loads, decrease, as some of the classrooms are not used after 14:00. It must be mentioned that in summer days before 9 July, the cooling loads are slightly lower due to the ambient conditions, and the cooling system’s output is lower than the maximum. Finally, the highest electricity load of the year occurs from 9 to 26 July from 08:00 to 14:00 and is equal to 17.85 kW (Figure 4b), while the minimum electricity load occurs in winter days from 14:00. to 16:00 and is equal to 7.49 kW (Figure 4a).

Through inspection of the daily electricity demand presented in Figure 5, it can be observed that the daily demand is constant in the winter, while in the summer, higher demand occurs and changes from day to day due to the cooling loads. The highest demand occurs from 9 to 26 July and is equal to 127.58 kWh.

3.2. Energy and Environmental Performance of the Backup Scenarios Considered

The sizing of the backup system was performed based on the results presented in Figure 3, Figure 4 and Figure 5 and considering that the backup system should be able to cover the building’s load in any case of power shortage. Thus, the backup system must be able to provide a maximum electricity load of 17.85 kW. Therefore, by applying a 10% safety factor, a backup system with a nominal power of 20 kW is considered.

As for the size of the storage, for Scenario 1, the 5 h of the year with the highest electricity demand should be considered. This is expected to occur during summer due to the identified higher loads (see Figure 3). By taking into account the daily variation of the loads in the summer, as presented in Figure 4b, it is expected that the hours with the highest loads are from 09:00 to 14:00, as afterward, the loads decrease. Therefore, the loads for every working day in July from 09:00 to 13:00 (5 h) were compared, and it was found that the days with the highest loads for 5 h are from 9 to 26 July when the electricity load from 09:00 to 13:00 is equal to 89.28 kWh. As for the 24 h case for the 2nd scenario, the highest daily electricity demand is considered, which is equal to 127.58 kWh. Finally, regarding the 72 h case for the 3rd scenario, the required energy is considered equal to 3 times the energy of the 24 h scenario. Due to the building’s operating schedule, the 24 h power shortage scenario includes 8 h of electricity loads, as the building is not operating for the rest of the hours. A comparison of the energy and environmental performance of the two considered backup systems is presented in

Table 8. Energy and environmental performance of the two systems.

Performance Indicator	HFC	DG
Efficiency	55%	25%
Fuel consumption (kWhfuel/kWhel)	1.81	4
CO2 emissions (g/kWh)	0	267
HC + NOx emissions (g/kWh)	0	4.7
PM emissions (g/kWh)	0	0.015
CO emissions (g/kWh)	0	5

, based on the data from Table 5, Table 6 and Table 7. Furthermore, the estimated annual energy demand for the backup system is presented in

Table 9. Energy and fuel demand in each scenario.

Scenario	Annual Total Shortage Duration (h)	Annual Energy Demand (kWh)	Annual Hydrogen Demand (kWh)	Annual Hydrogen Demand (kg)	Annual Diesel Demand (kWh)	Annual Diesel Demand (lt)
1	5	89.28	162.33	4.87	357.12	33.32
2	24	127.58	231.96	6.97	510.32	47.61
3	72	382.74	695.89	20.90	1530.96	142.83

. The required hydrogen and diesel fuel demand in each scenario is calculated based on the energy demand, the performance of the systems as presented in Table 8 and the data regarding the fuel properties of the two systems from Table 6 and Table 7. The demand is also presented in Table 9.

As can be observed in Table 9, the energy demand for hydrogen is lower than the demand for diesel due to the fuel cell’s higher efficiency (see Table 8). The calculated fuel demand determines the size of the fuel storage of the two systems, as it is considered that the required fuel quantity must be stored on-site in the case of a power shortage.

As mentioned before, one of the advantages of the HFC-based backup system is the ability to produce power without CO₂ emissions, in the case renewable hydrogen is used, as well as without HC and NO_x and PM and CO emissions. To assess the emissions reduction potential in the case of the replacement of a DG with the hydrogen system, the emissions from the DG in each scenario are calculated based on the emission conversion factors of Table 5 and presented in

Table 10. Annual emissions from a DG backup system for each scenario.

Scenario	DG CO2 Emissions (kg)	DG HC + NOx Emissions (kg)	DG PM Emissions (kg)	DG CO Emissions (kg)
1	94.64	1.68	0.01	1.79
2	135.23	2.40	0.01	2.55
3	405.70	7.20	0.02	7.65

.

It may be fairly concluded that the replacement of a DG with the HFC-based backup system could result in the avoidance of around 100–400 kg of CO₂ emissions, which corresponds to 1–4.3% of the building’s emissions from electricity. If the lifetime of the backup system is considered (15 years, as reported in Table 6 and Table 7), a cumulative avoidance of 1.4–6 tons of CO₂ in 15 years is possible. Moreover, a significant avoidance in the emissions of HC and NO_x and PM and CO is possible, i.e., up to 7.2 kg of HC and NO_x, 0.02 kg of PM and 7.65 kg of CO emissions. Finally, it should be mentioned again that noise emissions will also be avoided, something that is not quantified here, but it is indeed an advantage of the fuel cell system in the case of buildings as mentioned in other studies (e.g., [18,19]).

3.3. Economic Results

For the calculation of the required indicators, the costs of the two systems, i.e., the IC, operation and maintenance (O&M) cost and fuel cost, were calculated based on the data from Table 6 and Table 7. The results are presented in

Table 11. Costs of the two backup systems for each scenario.

Scenario	HFC IC (EUR)	HFC Annual O&M Cost (EUR)	HFC Annual Fuel Cost (EUR)	DG IC (EUR)	DG Annual O&M Cost (EUR)	DG Annual Fuel Cost (EUR)	Annual Cost Reduction from the Replacement of DG with HFC (EUR)
1	61,460.95	102.68	58.50	18,184.62	700.89	63.30	603.02
2	62,087.67	103.83	83.59	18,263.82	701.28	90.46	604.32
3	66,263.02	111.48	250.77	18,791.46	703.83	271.38	612.95

. It should be mentioned that in the case of the HFC-based system, the IC of the system excluding the hydrogen storage (i.e., total cost of the fuel cell, battery, converter and inverter) was considered equal to 3000 EUR/kW for all scenarios based on Table 6. The hydrogen storage cost was calculated for each scenario based on the required stored energy and the cost of the storage (see Table 6).

From Table 11, it can be observed that there is a significant difference in the annual cost between the two systems due to the much lower O&M cost of the HFC system. Furthermore, the higher efficiency of the fuel cell results in a lower annual fuel cost, even though the cost per kWh of fuel is higher for hydrogen than diesel fuel. As a result, an annual cost reduction of around EUR 603–613 is possible in case the HFC system is used instead of the DG. However, the IC of the HFC system is currently approximately 3.5 times higher than the cost of the DG cost. Finally, the difference in the HFC system’s IC among the scenarios is due to the different sizes of the hydrogen storage tank.

3.3.1. LCOE

Based on the results from Table 11, the LCOE of the two systems can be calculated for each scenario. The calculation of the LCOE is done based on Equation (1). The LCOE results are presented in

Table 12. LCOE of the two systems.

Scenario	HFC Annual O&M and Fuel Cost (EUR/kWh)	HFC LCOE (EUR/kWh)	DG Annual O&M and Fuel Cost over Lifetime (EUR/kWh)	DG LCOE (EUR/kWh)
1	1.81	55.38	8.56	24.41
2	1.47	39.34	6.21	17.35
3	0.95	14.42	2.55	6.37

for each scenario, along with the O&M and fuel cost of each system.

It can be concluded that in general, the LCOE of the DG is lower than the LCOE of the HFC system, which means that the DG can produce the required electricity at a lower cost. However, the O&M and fuel cost for the HFC system is much lower than those of the DG in every scenario; therefore, it can be concluded that the difference in the LCOE occurs due to the high IC of the HFC system. Finally, it can be observed that in general, the annual O&M and fuel cost and the LCOE for both systems decrease as the operating hours of the system increase.

3.3.2. NPV, IRR and PBP

Based on the results from Table 11, the economic feasibility of the replacement of a DG-based backup system with an HFC-based system can be assessed. The assessment is conducted by calculating the NPV, IRR and PBP for each scenario. For the calculation, the IC and the annual cash flow are necessary. For the latter, the annual cost reduction due to the replacement of the DG (see Table 11) is considered as the annual cash flow. Finally, in the calculation, an investment lifecycle of 15 years is assumed (see Table 6 and Table 7). The results are presented in

Table 13. Economic assessment KPIs.

Scenario	HFC IC (EUR)	Annual Cash Flow (EUR)	NPV (EUR)	IRR	PBP (Years)
1	61,460.95	603.02	−53,712.58	−17.89%	101.9
2	62,087.67	604.32	−54,322.65	−17.95%	102.7
3	66,263.02	612.95	−58,387.07	−18.32%	108.1

.

The tabulated KPIs suggest that the replacement of the DG with an HFC is perceived as non-viable. This is basically due to the relatively low annual cash flow compared to the IC. Moreover, the increase in the shortage duration seems to have a negative effect on the KPIs, as the KPIs become slightly worse. This occurs since the IC is higher due to the larger hydrogen storage that is required, while the annual cash flow only slightly increases.

3.4. Sensitivity Analysis and Conditions for Desired KPI Values

The sensitivity analysis was performed to further investigate the sensitivity of the LCOE and NPV to changes in the IC and the price of hydrogen. As shown in Table 13, the economic KPIs suggest a non-viable investment under the conditions presented in Table 6. Therefore, in addition to the sensitivity of the KPIs to the change in the IC and hydrogen price, the sensitivity analysis will also show the necessary conditions so that the KPIs receive the desired values.

The results from the sensitivity analysis are presented in Figure 6 for the LCOE and Figure 7 for the NPV in terms of contour plots representing the value of the respective KPI for different combinations of IC and hydrogen price. The sensitivity analysis was conducted for Scenario 1 and Scenario 3, as those are the scenarios with the least and the most operating hours, respectively. It should be mentioned that in these plots, the base case (meaning the case that corresponds to the current market situation, i.e., the IC and renewable hydrogen price as presented in Table 6), is included with a point. Additionally, the maximum or minimum desired value for the two KPIs is included with a contour line. This desired value is as follows:

• For the LCOE, the maximum desired value is the LCOE of the DG. If the LCOE of the HFC system is lower than that, the system can produce energy at a lower cost than a DG.
• For the NPV, the minimum desired value is 0, since positive NPV suggests a viable investment.

From inspection of Figure 6 and Figure 7, it can be seen that both indicators are sensitive to the IC, while their sensitivity to the hydrogen price is marginal. This is an expected result, as the operating hours of the system are very short, and so the fuel cost is not an important factor of the total cost compared to the IC. However, it can be observed that in the case of Scenario 3, the sensitivity to the hydrogen price is slightly higher due to the longer operating hours. Consequently, the IC remains the decisive factor in both the competitiveness of the HFC system against the DG and in the viability of the replacement of a DG with the proposed system. Finally, from Figure 6 and Figure 7, it can be concluded that it is possible to achieve the desired value for the LCOE and the NPV under conditions. The necessary conditions for this were calculated for all considered scenarios. Initially, the threshold for the IC, i.e., the maximum value of the IC for achieving the desired LCOE and NPV value, was calculated, followed by the calculation of the necessary initial IC reduction. Since the sensitivity of the KPIs to the hydrogen price is low, the calculation was performed once considering the current hydrogen price and once considering a hydrogen price as low as 1 EUR/kg, which is a price level that could be achieved in the future [10]. The results are presented in

Table 14. IC threshold for LCOE > LCOE_Diesel for each scenario.

	H2 Price = 12 EUR/kg		H2 Price = 1 EUR/kg
Scenario	IC Threshold (EUR/kW)	Reduction in Initial IC	IC Threshold (EUR/kW)	Reduction in Initial IC
1	1290	58.02%	1330	56.72%
2	1300	58.12%	1350	56.51%
3	1330	59.86%	1480	55.33%

for the LCOE and in

Table 15. IC threshold value for NPV > 0 for each scenario.

	H2 Price = 12 EUR/kg		H2 Price = 1 EUR/kg
Scenario	IC Threshold (EUR/kW)	Reduction in Initial IC	IC Threshold (EUR/kW)	Reduction in Initial IC
1	380	87.63%	420	86.33%
2	380	87.76%	430	86.15%
3	390	88.23%	540	83.70%

for the NPV. Moreover, the LCOE and the investment KPIs for each scenario for the respective IC threshold and hydrogen price of Table 14 and Table 15 are presented in

Table 16. LCOE and investment KPIs for each scenario for the IC threshold and H₂ price from Table 14.

	H2 Price = 12 EUR/kg					H2 Price = 1 EUR/kg
Scenario	NPV (EUR)	IRR	PBP (Years)	HFC LCOE (EUR/kWh)	DG LCOE (EUR/kWh)	NPV (EUR)	IRR	PBP (Years)	HFC LCOE (EUR/kWh)	DG LCOE (EUR/kWh)
1	−18,051.63	−10.93%	42.8	24.30	24.41	−18,162.63	−10.44%	40.5	24.39	24.41
2	−18,234.98	−10.98%	43	17.33	17.35	−18,250.41	−10.25%	39.7	17.34	17.35
3	−18,724.06	−11.06%	43.4	6.36	6.37	−18,770.35	−9.14%	35.1	6.36	6.37

and

Table 17. LCOE and investment KPIs for each scenario for the IC threshold and H₂ price from Table 15.

	H2 Price = 12 EUR/kg					H2 Price = 1 EUR/kg
Scenario	NPV (EUR)	IRR	PBP (Years)	HFC LCOE (EUR/kWh)	DG LCOE (EUR/kWh)	NPV (EUR)	IRR	PBP (Years)	HFC LCOE (EUR/kWh)	DG LCOE (EUR/kWh)
1	148.37	2.26%	12.6	8.43	24.41	37.37	2.06%	12.8	8.53	24.41
2	165.02	2.29%	12.6	6.11	17.35	149.59	2.23%	12.6	6.11	17.35
3	75.94	2.13%	12.7	2.53	6.37	29.65	2.04%	12.8	2.54	6.37

, respectively.

From Table 14 and Table 15, it can be concluded that:

• The IC threshold for achieving the desired LCOE value can be much higher than the threshold for the desired NPV.
• At the current hydrogen price, the IC threshold for viable performance essentially differs slightly among the scenarios, being slightly higher in the case of Scenario 3. This is because of the higher annual O&M and fuel cost savings in comparison to the DG that occur in this scenario (see Table 11), which can compensate for a higher IC.
• The change in the hydrogen price has a slight impact on the IC threshold for both indicators, as indicated in Figure 6 and Figure 7. However, it has a higher impact when the operating hours of the system are longer. That means that the sensitivity of the LCOE and NPV in the hydrogen price is higher when the operating hours of the system are longer, as it was observed in Figure 6 and Figure 7.

Furthermore, from the tables above, it can be observed that in order to have the required economic performance, a reduction in the IC of the system is primarily desired. For the LCOE, a reduction of around 58–60% of the IC is necessary so that the LCOE of the HFC system is lower than the LCOE of DG. As for the NPV, a reduction of around 87–89% of the IC is necessary for the replacement of a DG with an HFC to be viable.

Additionally, from Table 16 and Table 17, it can be observed that:

• For each scenario, for the IC threshold presented in Table 14, as expected, the LCOE of the HFC is lower than that of a DG. As for the investment KPIs, they still suggest an unviable investment, but they are significantly improved compared to the values presented in Table 13.
• For the IC threshold presented in Table 15, the investment KPIs suggest a viable investment, as expected. At the same time, the LCOE of the HFC is much lower than the LCOE of the DG.

4. Discussion

In this study, the economic and environmental performance of the HFC-based backup system in comparison to a conventional DG, also discussed in previous studies (e.g., [18,19]), are explored. The environmental benefits of the proposed system are important, as the implementation of this system instead of a DG in the studied building could achieve an annual CO₂ reduction of up to around 400 kg of CO₂, as shown in Table 10. This is equal to 4.3% of the building’s total emissions due to electricity consumption, and a cumulative 1.4–6 tons of CO₂ in a 15-year period. This reduction may be relatively low compared to the building’s total annual CO₂ emissions, but it is in line with the EU goals for decarbonized buildings. The HFC system can also contribute to the avoidance of HC and NO_x and CO and PM emissions (see Table 10). All in all, in the modern grid-connected environment, the replacement of DGs with the proposed HFC system is able to contribute to the sustainable development of the building sector.

Concerning the economic performance dimension, it is assessed in terms of a comparison of the proposed system with a conventional DG. As can be observed from Table 11, the maintenance cost of the HFC system was found to be lower than the costs of the DG. Even though diesel fuel is cheaper per kWh of fuel (see Table 6 and Table 7), the annual fuel cost is lower in the case of the HFC (see Table 11) due to the higher efficiency of the fuel cell (see Table 8). Overall, it was concluded from Table 11 that the HFC system has lower annual operation and fuel costs. The main drawback of the HFC system is its high IC. As a result, the calculated LCOE of the HFC system was higher (see Table 12), which means that the HFC system produces electricity with a higher cost over its lifetime compared to a DG. Furthermore, in the case of an investment for replacing a conventional DG with the proposed system, the economic KPIs suggest a non-viable investment under the ongoing market environment, as reported in Table 13.

One observation from the investigation is that as the operating hours of the system increased, the IC threshold for obtaining the desired LCOE and NPV increased (see Table 14 and Table 15) due to the higher annual O&M and fuel cost savings in comparison to a DG that occur when the operating hours are higher. This suggests that the longer the operating hours of the system are, the higher the annual savings from the replacement of a DG will be. Cases with more operating hours could be, for example, applications where a generator provides power for longer periods, such as off-grid applications (as the applications mentioned in [23,24]), or applications where the backup generator also provides grid ancillary services, as proposed in [22]. On the other hand, the increased hydrogen storage size in such a case would result in worse NPV, IRR and PBP (see Table 13). This means that the hydrogen storage could be sized differently in the case of many hours of operation, and the refueling of the storage more than once per year should be considered. In that way, the storage size could be smaller, and thus the IC could be lower. In this case, the HFC system could have a better economic performance, and the replacement of a DG would be more beneficial. Moreover, the reduction in emissions would be higher. However, if the number of operating hours per year would increase significantly, the issue of hydrogen supply would become important because for large quantities of hydrogen, the refueling cost would be considerable. Hence, the on-site production and possible storage of hydrogen would be a feasible solution, which, however, would increase the IC.

An important finding of the study is that the adoption of the proposed technology in the building sector requires financial support. The IC was found to be the decisive factor in the economic performance of the system, while the hydrogen cost does not have a significant effect on the economic performance, as shown in the sensitivity analysis presented in Figure 6 and Figure 7. As it was presented in Table 14 and Table 15, in order for the proposed system to be competitive with the conventional DG in terms of LCOE, it needs to have a 58–60% lower IC, while the replacement of an existing DG with the proposed system was found to be viable with an 87–89% lower IC. Indeed, this was confirmed in the results presented in Table 16 and Table 17, where for the identified IC thresholds, the LCOE and NPV, respectively, received the desired values. A reduction in the IC is possible in the future if the production volume of this technology is increased. However, in the current market situation, a form of subsidy is necessary. In the case of Greece, where the use of hydrogen and fuel cells is still in its infancy, policies that would create a more extended use of fuel cells and a higher demand for hydrogen are indeed necessary, which could kickstart the more extensive use of fuel cell systems, eventually allowing hydrogen to further penetrate the energy system. Currently, subsidies for HFC systems as backup systems specifically for the building sector are not found. However, initiatives from the governments for the promotion of hydrogen technologies are reported, for instance, in the case of fuel cell electric vehicles (FCEVs). For this application, subsidies for the purchase of FCEVs are reported and are considered necessary to increase the market penetration of FCEVs and boost the replacement of internal combustion engines towards the decarbonization of the mobility sector [45,46]. Subsidies aiming at the development of stationary fuel cell applications and hydrogen infrastructure are also reported [45]. The aforementioned findings indicate that in the framework of promoting the use of hydrogen, subsidies for the proposed HFC backup system could be possible in the near future.

Regarding the methodology that was implemented, the comparison between the two systems and the techno-economic assessment that was performed based on the chosen case study in this work can be replicated for other buildings, as well as other infrastructure, where the installation of a backup system is relevant, given that data about the infrastructures’ electricity load variation exist. The methodology can be replicated for other regions too by taking into account the respective economic environment and the respective input conditions for the simulation model, i.e., climatic conditions, construction materials properties, etc.

Regarding the limitations of the work, the calculations of the system’s properties were based on a dynamic building energy simulation instead of measurements of the electricity load. The latter would provide a more reliable view of the techno-economic study performed for the case studied. However, the model used is validated against experimental data in a previous study, which provides evidence that the fidelity of results is acceptable. Another point that could be perceived as a limitation is the fact that the information regarding the frequency and duration of power shortages refers to the whole country rather than local sources or event statistical data of the specific building studied. Information about shortages in the specific area or even the specific building would allow a more accurate calculation of the annual shortage duration. Moreover, the design and operation of the proposed system and its subsystems were not investigated in detail, as the aim was to focus on the environmental and economic performance of the system. As a result, the assumptions regarding the efficiency of the two systems that were made could have slightly different values.

As for future work, pilot projects investigating the implementation of the proposed system in buildings and other applications would be necessary, including measurements of both the building loads and the system’s power production. The pilot projects should focus on the detailed design of the proposed system, i.e., the sizing of the subsystems, the design of the control system, etc., so that issues such as the startup time of the fuel cell are addressed. They should also identify the challenges and specify the requirements for the integration of the proposed system in several applications. In the end, this could allow the specification of a portfolio of applications where the proposed system could be used for the provision of backup power. Furthermore, exploitation of the methodology presented in this study for various building typologies, such as houses, hospitals, shopping centers, etc., would reveal the required subsidies in those typologies. Last but not least, case studies of the implementation of the proposed technology in other types of applications, such as industrial applications, would be interesting to investigate the techno-economic feasibility of HFCs in those applications as well.

5. Conclusions

In this work, the operation of an HFC-based backup system was compared with a conventional DG backup system in terms of environmental and economic performance, and the techno-economic feasibility of the replacement of an existing DG with the HFC system was assessed for several power shortage scenarios. A school building was chosen as a representative case study, considering the high occurrence frequency of such types of buildings most commonly owned or managed by local public authorities. Building loads were calculated using a valid dynamic building energy simulation model providing a realistic time series of electricity loads on the bases of which the main system specifications are determined. The economic conditions that would ensure the techno-economic feasibility of the implementation of the HFC system were investigated.

The HFC system was found to have a lower environmental footprint than that of the DG, being able to reduce the building’s CO₂ emissions due to electricity consumption by up to 4.3% in the investigated power shortage scenarios, as well as eliminate the emissions of other pollutants that are emitted from DGs. From that, it was concluded that the promotion of the use of this system in the building sector fits with the EU and Greek policies, as it could contribute to the achievement of the EU and Greek decarbonization targets for buildings.

As for the comparison between the economic performance, the HFC system has a lower annual operating and fuel cost. However, due to its high IC, its LCOE was higher, and in the case of replacing an existing DG, the results of the techno-economic analysis suggested an unviable investment. Therefore, subsidies are necessary to foster the application of such technology at least in public tertiary sector buildings. This would gradually make the proposed HFC system economically competitive with the DG and further boost its use in private buildings as well. Currently, it is estimated that subsidies of at least 58–60% are necessary to make the HFC system competitive with a DG. Projects referring to the replacement of an existing DG with the proposed system would require a subsidy of at least 87–89%. At the same time, it was found that a potential subsidy on the cost of hydrogen would not affect the economic performance of this technology significantly.

The adoption of the proposed system could take place in the building sector, in applications such as schools, data centers, hospitals, etc., and could also trigger the adoption in different sectors where backup systems are relevant, such as telecommunications. This would contribute to a broader decarbonization of those sectors. Moreover, extensive adoption of this system in several applications could facilitate the deployment of fuel cell technologies and the integration and more extensive use of hydrogen in the energy system, which is also a principal target of the EU policy.