Techno–economic analysis of green hydrogen production by a floating solar photovoltaic system for industrial decarbonization

Abstract

This study proposes a conceptual design of green hydrogen production via proton exchange membrane electrolysis powered by a floating solar photovoltaic system. The system contributes to industrial decarbonization in which hydrogen blending with natural gas is proposed as an approach to smooth the energy transition. The proposed design addresses the challenge of supplying a continuous flow-rate of green hydrogen, which is typically demanded by industrial end users. This study particularly considers a realistic area required for the installation of a floating solar photovoltaic system. To enable the green hydrogen production of 7.5 million standard cubic feet per day, the required structure includes the floating solar photovoltaic system and Li-ion batteries with the nominal capacities of 518.4 megawatts and 780.8 megawatt-hours. This is equivalent to the requirement for 1 524 765 photovoltaic modules and 3718 Li-ion batteries. The assessment confirms the technical viability of the proposed concept of green hydrogen production, transportation and blending. While the present commercialization is hindered by economics due to a high green hydrogen production cost of USD 26.95 per kg, this green hydrogen pathway is expected to be competitive with grey hydrogen produced via coal gasification and via natural gas steam reforming by 2043 and 2047, respectively.

Graphical Abstract

floating solar photovoltaics, gas blending, green hydrogen, industrial decarbonization, lithium-ion battery, photon exchange membrane electrolysis

Introduction

The reliance of the global society on fossil fuels as primary energy sources for different end-use sectors has caused a threat to energy security and led to negative environmental impacts. To mitigate such consequences, global energy transformation actions are crucial, in which it is essential to rapidly displace fossil fuels with low-carbon energy sources. The utilization of sustainable energy plays an important role in achieving the successful energy transition. Among others, solar is the main sustainable and most readily available resource that supports life on Earth; therefore, there is a strong drive to generate energy by utilizing solar technologies. Solar photovoltaics (PV) are one of the promising technologies used to harness solar energy by directly converting sunlight into electricity. Solar PV systems are praised for their abundant energy source, low-maintenance prerequisites, low emissions of CO₂ and environmental friendliness [1].

The great potential for the utilization of solar PV technology is, however, hindered by the requirement for spacious amounts of land for its installation [2]. Considering such limitations of ground-mounted solar PV technology, solar PV cell installation on non-conventional land-based spaces, such as rooftops, trees and water bodies, has received growing interest over the past few decades. A rooftop PV is highly attractive for densely populated urban areas [3]. A solar PV tree can generate electricity for different purposes, such as street lighting, electronic device and vehicle charging, and agricultural irrigation. However, as they are limited by roof space and the number of solar PV panels that can be supported by an artificial stem, the electricity production capacities of rooftop PV and solar PV tree systems are relatively small [4].

Among the others already mentioned, floating photovoltaic (FPV) systems installed on water bodies, such as lakes, ponds and reservoir, have the potential to generate electricity with a high production capacity [5]. Furthermore, they offer unique advantages besides avoiding land use, including improved PV module cooling systems due to the adjacent water, less dust accumulation and water evaporation reduction [6].

Renewable electricity generated from a solar PV system with a sufficient production capacity, such as from a floating PV system, can be utilized further to produce green fuels and chemicals. Hydrogen is considered a promising carbon-free fuel that can be used in a multitude of end-use applications, including but not limited to fuel-cell electric vehicles, seasonal electrical energy storage, heating, industrial process materials and natural gas mixture composition [7]. Electricity from a solar PV system can be used to synthesize green hydrogen by converting water into hydrogen and oxygen via an electrolysis process. Traditionally, hydrogen is generated via steam methane reforming using natural gas and coal gasification.

The green hydrogen demand during recent years has been driven by the use of fuel-cell electric vehicles and several industrial applications in oil refining and the production of ammonia and methanol [8]. Moreover, the demand from broader users is projected to rise significantly, such as from the steel, cement, chemical, paper, aluminium and other industries requiring hydrogen as an energy carrier [9]. Note that the heavy hydrogen demands from industries cannot be currently be fully fulfilled by stand-alone technologies producing green hydrogen when considering the very high investments and excessive constructions and equipment modifications required. Therefore, to expedite the commercialization of green hydrogen technologies, moderate approaches are needed so that less significant changes in existing industries are required.

Research in hydrogen production using solar energy has been carried out by the scientific community with different points emphasized. Gibson and Kelly [10] performed a study to optimize the efficiency of a PV–electrolysis system for hydrogen production by matching the voltage and maximum power output of the photovoltaic system to the operating voltage of proton exchange membrane (PEM) electrolysers. Dincer and Acar [11] assessed 19 technologies of hydrogen production by considering their environmental impacts, costs, energy and exergy efficiencies. Based on their assessment, electrolysis technology powered by a photovoltaic system gives the highest energy and exergy efficiencies among all the photonic hydrogen production methods. Fereidooni et al. [12] performed both experimental and simulation studies to evaluate hydrogen production through an electrolysis process supported by a 20-kW photovoltaic power station in the city of Yazd, Iran. The designed system is capable of producing 373 tons of hydrogen per annum, but the hydrogen utilization was not taken into consideration in the discussion. Khelfaoui et al. [13] conducted an investigation of solar hydrogen production considering climatic conditions in the Algerian Sahara regions. They performed a parametric study and optimized the proposed solar PV/PEM system to maximize the hydrogen production. Hassan et al. [14] modelled and analysed green hydrogen production by solar energy in four different cities in Iraq. They suggested the best potential location for the generation of green hydrogen based on solar irradiation and ambient temperature characteristics.

A bright prospect among some moderate approaches is hydrogen blending with natural gas, which is a promising pathway to support the ongoing energy transition to minimizing the use of high-carbon energy sources. This concept is applied by injecting hydrogen into natural gas transmission or distribution networks [15, 16]. Studies on hydrogen–natural gas blending have risen recently covering technical, economic and environmental aspects of its application [17–19]. Isaac [17] presented an insight into the HyDeploy project in the UK, which is the first practical project demonstrating hydrogen blending into a natural gas system without requiring devices and equipment changing. Among three planned phases of the project, Phase 1 has been carried out to exhibit that hydrogen is safe to blend into the Keele University private gas network. Pellegrini et al. [18] conducted a preliminary assessment of green hydrogen blending (≤10 vol%) into the Italian natural gas network. Their study suggested that the injection of 8100 tons of green hydrogen per year into the existing network is technically feasible. Wu et al. [19] provided a technical review on the hydrogen-induced failure of high-strength pipeline steels in hydrogen-blended natural gas transmission networks. Having identified existing issues, they proposed three research directions, including the multi-mechanism synergy mechanism, improvement in experimental methods and the establishment of a new interatomic multiscale model. Zhao et al. [20] evaluated the effects of hydrogen addition to pipeline natural gas on combustion and cooking performance for residential end users and their findings suggested that a 15% hydrogen addition by volume does not have a significant influence on the combustion performance. Cristello et al. [21] investigated potential technical issues that may occur when hydrogen is injected into natural gas pipelines by simulating the physical characteristics of transmission and distribution pipelines using a gas hydraulic model. They also examined the behaviour of natural gas and hydrogen with different mixtures in pipelines by using the Real-Time Transient Model. Sorgulu et al. [22] observed the impact of hydrogen injection into natural gas by examining the natural gas consumption, heating times and heating values. Jia et al. [23] suggested some recommendations to improve blended hydrogen–natural gas transportation that include hydrogen embrittlement investigation within specific compositions, the exploration of diverse network system infrastructures and studies on materials resistant to hydrogen embrittlement. Ozturk et al. [24] conducted an experimental study assessing the environmental impacts of the burning of hydrogen and natural gas blends by comparing the burning performance of 10%, 20% and 30% hydrogen blending ratios and by analysing the emissions discharged from gas stoves. Davis et al. [25] evaluated different scenarios to determine the greenhouse gas mitigation potential and cost-effectiveness of blending hydrogen with natural gas for energy consumption for the years 2026–50.

Fig. 1 shows the permissible limits of blended hydrogen in natural gas pipelines governed in different countries. The presented data suggest that those ranges have been proven to be practically acceptable in the application of hydrogen and natural gas blending. Such limits are necessary to avoid potential consequences for end users who implement the blending, i.e. the requirement for larger gas volumes to meet the energy needs due to the lower energy density of hydrogen compared with natural gas and a higher risk of pipeline brittleness [23, 26].

Figure 1.

Permissible limits of blended hydrogen in natural gas pipelines in different countries, adapted from [26, 27]

From the perspective of sustainability, hydrogen and natural gas blending application is expected to be much more attractive when green hydrogen is utilized. Research in the integration of green hydrogen production and its blending with natural gas began to emerge a few years ago. Jin et al. [28] studied green hydrogen blending scenarios in existing natural gas pipelines in Italy by considering diverse renewable energy sources and several gas blending ratios. But the technical aspects of the hydrogen production process and its transportation to the blending location were not discussed in detail. Ozturk and Dincer [29] evaluated green hydrogen production generated via electrolysis using wind, solar and wave energies, and its blend with natural gas for residential applications. Their findings were limited to technical aspects, while the required spacious land for solar PV was not discussed in detail. Currently, there is a research gap in the evaluation of green hydrogen production that is specifically aimed to fulfil 24/7 continuous demand, with a constant flow-rate, for industries, while the utilization of green hydrogen for blending purposes with natural gas is crucial for the act of industrial decarbonization. Furthermore, a realistic design for technologies that empower renewable sources needs to be addressed, in which it is required to consider the available areas for installation and operation as a constraint. Besides, technical and economic aspects need to be assessed together, as both aspects drive the feasibility of a conceptual design.

This study addresses the research gap by aiming to propose a conceptual design for green hydrogen production for the purpose of industrial utilities. The objectives of this study are as follows:

• to design a generation system of green hydrogen for the purpose of blending with natural gas;

• to assess green hydrogen production utilizing stand-alone solar energy technology for electrolysis instead of mixed energy resources as proposed by Ozturk and Dincer [29];

• to evaluate both technical and economic aspects of green hydrogen production by an FPV system that is specifically design for industrial decarbonization.

In this proposed design, green hydrogen is produced via proton exchange membrane electrolysis (PEMEC) powered by FPV. As discussed by Kumar et al. [6], FPV for green electricity generation brings prospective benefits over other solar PV configurations; therefore, this technology is assessed in this study. The novelty of this study lies in addressing the challenge of supplying the continuous flow-rate of green hydrogen typically demanded by industrial end users, by accounting for the available and required areas for FPV installation and operation.

1 Research methodology

The study is limited to an assessment based on simulation and calculation using different software including HOMER Pro, Aspen Plus, Aspen HYSYS and Microsoft Excel, and the Google Earth engine. There were no on-site surveys that took place to obtain primary data, such as route elevations. Fig. 2 depicts the schematic system proposed in this study. The proposed system consists of floating solar photovoltaic modules installed on the Jatiluhur reservoir, operating devices including an electrolyser and batteries units placed next to the reservoir and a hydrogen pipeline system delivering hydrogen to the entrance gate of the Jababeka industrial estate—a centralized industrial area in West Java. As the natural gas and hydrogen blending takes place at the entrance gate of the industrial estate, the blended gas is ready to be distributed by other parties for end users in Jababeka, including those from steel and iron, petrochemical, ceramic, glass, and oleochemical industries [30]. Additionally, in this system, excess electricity generated during the daytime is sent to the local electrical grid and bought by the state-owned electricity company, Perusahaan Listrik Negara (PLN).

Figure 2.

The schematic system proposed in this study.

Fig. 3 provides the spatial connections of the elements of the system. The Jatiluhur reservoir—the largest reservoir in Indonesia—is located on Java Island, which is one among five big islands in Indonesia. The reservoir has a surface area of 83 km² [31] and it currently serves several activities for different purposes, including fish ponds, water irrigation and hydroelectric power generation. Disregarding the occupied water spaces for those existing activities and considering a relatively distant space to the hydroelectric power generation location, which is shown by the triangle marker in Fig. 3, it was interpreted from the Google Earth engine that there is a maximum surface area of 5 km² on the Jatiluhur reservoir that can be utilized for the installation of floating PV modules. Thus, this value became the upper limit of the available space considered in this study. The ratio of the estimated maximum surface area to the total surface area of the Jatiluhur reservoir, i.e. 0.06, is within the practical range of the FPV installation on water bodies, which is 0.01–0.1 [6].

Figure 3.

Map of the assessed system. (a) The Jatiluhur reservoir, (b) the potential space for the FPV installation, (c) the potential space for supporting facilities, (d) hydrogen pipeline route and (e) final hydrogen delivery point for industrial end users.

The assessment in this study was initiated by estimation of the green hydrogen demand for the purpose of blending with natural gas. Then, simulation work in the Aspen Plus v.12.1 process simulator was performed to calculate the amount of power required to operate the electrolyser unit for producing green hydrogen. This power requirement became the input information for the simulation by using the HOMER Pro v.3.14.5 software to obtain the FPV capacity and number of batteries. Subsequently, the number of PV modules and the required surface area on the Jatiluhur reservoir were quantified. The calculated area was then compared with the maximum available surface area defined earlier. Next, simulation work using the Aspen HYSYS v.12.1 process simulator was carried out to model the green hydrogen transportation via the pipeline system. Finally, simulation work was done to evaluate the hydrogen–natural gas blending effects by using the Aspen Plus v.12.1 process simulator.

The hydrogen demand was estimated based on the actual natural gas demand for utility purposes and the acceptable range of hydrogen–natural gas blending ratios. The actual natural gas demand in the Jababeka industrial estate was 100 million standard cubic feet per day (MMSCFD) or 2.83 million standard cubic metres per day, and the value was obtained from the Indonesian government document Keputusan Menteri Energi dan Sumber Daya Mineral Republik Indonesia No. 134.K/HK.02/MEM.M/2021 [30].

To determine the percentage of hydrogen blending into natural gas pipeline networks considered in this study, the current permissible limits given in Fig. 1 are taken as a base. The practicality of blending percentages within the defined limits was reported by previous publications, such as Altfeld and Pinchbeck [32], in which they confirmed that modified burners of premixed gas appliances are not needed when the hydrogen blending into the natural gas is <10 vol%. As the surface area availability for the PV installation was essentially taken into account as a constraint in this study, hydrogen blending at 7.5 vol% was set as the assessed scenario. Thus, the amount of hydrogen to be distributed along the pipeline from the Jatiluhur reservoir to the entrance gate of the Jababeka industrial estate was 7.5 MMSCFD (0.21 million standard cubic metres per day). Note that other studies performed analyses on cases with higher blending percentages of hydrogen for future potential applications. For instance, Sorgulu et al. [22] conducted experimental investigation for the combustion performance of blended gases when 10%, 20% and 30% of hydrogen volumetric fractions were considered. They found an increased heating time by 15.87% with 20% of hydrogen addition. Koo et al. [33] evaluated the hydraulic performance of hydrogen blending in high-pressure natural gas pipelines. They suggested that there is no significant hydraulic impact on pipelines when ≤20 vol% of hydrogen is blended.

Green hydrogen, interchangeably called renewable hydrogen, is produced through clean processes using renewable sources, such as solar, wind, water or biomass, as the feedstock and utilities [34]. In this study, an electrolysis process that converts water from the reservoir into hydrogen was taken into account, with the required electricity directly generated from the FPV system. The PEM electrolyser was selected as a viable technology for performing electrolysis in this study, as it is well known for its high efficiency at producing pure hydrogen, fast dynamic response and ability to operate at relatively high pressures [35–37]. Furthermore, the PEM electrolyser was proven to be suitable for the utilization of renewable power sources, as it is less sensitive to the fluctuating behaviour of the power supply [38]. The PEM was considered to be made of a semipermeable polysulfonate material to allow proton exchange and avoid electron exchange between the electrodes [39]. The polysulfonate material was taken into account because of its advantageous characteristics, including its relatively low gas permeability, high proton conductivity, low thickness and good durability for high-pressure operations [40].

Typically, a PEM electrolyser unit operates at a mild temperature, i.e. within the range of 70–90°C [41]. Therefore, the PEM electrolyser unit in this study was simulated with an operating temperature of 90°C. Employing the PEM electrolyser unit at a mild temperature could promote proton exchange across the membrane and improve the electrical conductivity of the electrolyte [42]. The set pressures at the anode and cathode sides were 1.5 and 15 bar, which are based on the values suggested by Zaccara et al. [43].

Equations (1–3) exhibit oxidation and reduction reactions in the anode and cathode sides, respectively, and the overall reaction occurring in the electrolyser, which are:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{A}\mathrm{n}\mathrm{o}\mathrm{d}\mathrm{e}\left(\mathrm{o}\mathrm{x}\mathrm{i}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\right):2{\mathrm{H}}_2{\mathrm{O}}_{(\mathrm{l})}\to {\mathrm{O}}_{2(\mathrm{g})}+4{{\mathrm{H}}^{+}}_{(\mathrm{a}\mathrm{q})}+4{\mathrm{e}}^{\text{~}\text{-}}}\end{array}}$$

(1)

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{C}\mathrm{a}\mathrm{t}\mathrm{h}\mathrm{o}\mathrm{d}\mathrm{e}\left(\mathrm{r}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\right):4{{\mathrm{H}}^{+}}_{(\mathrm{a}\mathrm{q})}+4{\mathrm{e}}^{\text{~}\text{-}}\to 2{\mathrm{H}}_{2(\mathrm{g})}}\end{array}}$$

(2)

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{O}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{l}\mathrm{l}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}:{\mathrm{H}}_2{\mathrm{O}}_{(\mathrm{l})}\to {\mathrm{H}}_{2(\mathrm{g})}+1/{}_2{\mathrm{O}}_{2(\mathrm{g})}.}\end{array}}$$

(3)

The formulae for calculating the amount of power needed for hydrogen generation were given by Zhang et al. [44]:

$${\displaystyle \begin{array}{l}{\displaystyle E=I\cdot V}\end{array}}$$

(4)

$${\displaystyle \begin{array}{l}{\displaystyle I=j\cdot A}\end{array}}$$

(5)

$${\displaystyle \begin{array}{l}{\displaystyle V=V_0+V_{act}+V_{ohm}+V_{con}}\end{array}}$$

(6)

where E is the amount of power (W), I is the electric current (A), j is the current density (A/m²), A is the active area, V is the cell voltage (V), which is the summation of the reversible voltage (V_o), the activation overpotential losses (V_act), the ohmic potential losses (V_ohm) and the concentration overpotential losses (V_con). All values of the cell voltage elements were determined according to the formula and graph provided in detail by Zhang et al. [44], where V_o = 1.22 V, V_act = 0.822 V, V_ohm = 26.32 V and V_con = 8353.84 V. A was estimated iteratively so that several essential parameters, i.e. the required energy per kg of hydrogen, the water-to-hydrogen ratio and the electrolyser efficiency, were almost identical to those given by Zaccara et al. [43]. Furthermore, the iterated A was employed in the Aspen Plus simulation to quantify the permeation of the produced gas through the membrane so that the desired flow-rate of hydrogen could be achieved. The permeated gases calculation was based on the formulae given in Zaccara et al. [43]:

$${\displaystyle \begin{array}{l}{\displaystyle H_2^{perm}=(0.0009\cdot e^{0.025T})\cdot \mathrm{\Delta }P\cdot A}\end{array}}$$

(7)

$${\displaystyle \begin{array}{l}{\displaystyle O_2^{perm}=0.5\cdot H_2^{perm}}\end{array}}$$

(8)

where H₂^perm and O₂^perm are permeated hydrogen and oxygen, respectively, while ΔP is the pressure difference between the cathode and anode and the active area. The electrolyser efficiency (η) was calculated by adapting the formula presented in Ozturk and Dincer [29], which is expressed as:

$${\displaystyle \begin{array}{l}{\displaystyle \eta =\frac{{{\displaystyle \dot{n}}}_{H2}.LH{V}_{H_2}}{{{\displaystyle \dot{W}}}_{electrolyzer}\cdot 3.6}}\end{array}}$$

(9)

where ṅ_H2, LHV_H2 and Ẇ_electrolyzer are the hydrogen production rate (kg/h), the hydrogen lower heating value (MJ/kg) and the power input to the PEMEC, respectively, and 3.6 is the conversion factor from kWh to MJ.

The process flow-sheeting for the PEM electrolysis was obtained from simulation in Aspen Plus by adapting those suggested by Pratama et al. [42] and Zaccara et al. [43]. The flowsheet, as shown in Fig. 4, can be divided into four sections: (i) process preconditioning, (ii) stack, (iii) cathode and (iv) anode sections. In the process preconditioning section, the pressure and temperature of the water inlet were adjusted to the specified operating conditions at 80 bar and 90°C assuming that the fresh water was under atmospheric pressure at the ambient temperature. Then, the conversion of water into hydrogen and oxygen occurred in the stack section (an RSTOICH reactor in Aspen Plus) obeying the overall reaction mechanism given in Equation (3). The rest of the process sequence shown in Fig. 4 depicts the cathode and anode sections in the PEM electrolyser.

Figure 4.

The PEM electrolyser flowsheet

There were three manipulator blocks employed in the simulation consisting of one design-spec block and two calculators. The ‘H2CALC’ block calculated the amount of permeated hydrogen (Stream 8) to enter the anode section and the ‘O2CALC’ block computed the amount of permeated oxygen (Stream 9) to enter the cathode section, which employed Equations (8) and (9), respectively. The ‘H2OIN’ block was employed to obtain the flow-rate of fresh water in which the hydrogen flow-rate was set at 7.5 MMSCFD (0.21 million standard cubic metres per day) as desired.

The calculation of the FPV capacity and the number of batteries using HOMER Pro has the main objective of providing a continuous supply of hydrogen during the day and night to industrial end users. Such a steady incoming supply is typically required by industries to maintain the smooth running of their processes. Furthermore, the simulation work performed in this study using HOMER Pro resulted in a dispatch profile that contains strategies for fulfilling the renewable-energy demand to produce green hydrogen continuously.

As the design input for the simulation in HOMER Pro, the solar radiation data for the Jatiluhur reservoir were obtained from the RenewablesNinja website [45–47] and NASA’s meteorological database provided in HOMER Pro. As depicted in Fig. 5, the solar radiation varies from 3.23 to 5.24 kWh/m²/day, resulting in an annual average solar radiation of 4.81 kWh/m²/day. The area receives more solar radiation from May to October due to the dry season and vice versa from November to April due to the rainy season.

Figure 5.

Monthly average solar global horizontal irradiance at the Jatiluhur reservoir

Monocrystalline silicon-based solar cells were considered due to their higher efficiency compared with other materials [48]. The commercial Canadian Solar CS6U-340M modules were selected as the solar PV panels in this study due to their satisfying energy and economic performance [49, 50]. The technical and economic data of the solar PV modules are provided in Table 1. The listed capital cost has been adjusted for floating solar PV installations on water bodies, as suggested by Ramasamy and Margolis [51].

Table 1.

Technical and economic data of FPV [52–54]

Parameter (unit)	Value
Module rated capacity (Wp)	340
Nominal efficiency (%)	17.49
Module dimension (inch × inch × inch)	77.20 × 39.10 × 1.57
Lifetime (years)	25
Capital cost (USD/kW)	1343.85
Operating and maintenance cost (USD/kW/year)	26.88

Table 2 lists the technical and economic data of the battery considered in this study. In this study, Tesla Powerpack 2 lithium-ion batteries were used to store some of the electricity generated by the FPV system so that the PEM electrolyser could be operated to serve the continuous hydrogen demand for industrial end users. Such an energy storage mechanism is able to overcome the major limit of solar PV technology caused by its cyclic time-dependent energy source [1]. Among other battery technologies, the lithium-ion battery was selected, as it is features a higher efficiency and a larger energy density compared with the others [55].

Table 2.

Technical and economic data for the Tesla Powerpack 2 lithium-ion battery [56]

Parameter (unit)	Value
Nominal voltage (V)	380
Nominal capacity (kWh)	210
Maximum charge/discharge current (A)	131
Lifetime (years)	10
Capital cost (USD)	118 020
Replacement cost (USD)	84 000
Operating and maintenance cost (USD/year)	1180.2

The hydrogen pipeline route shown in Fig. 3 was designed with the aim of obtaining as many straight lines as possible and fewer obstacles such as rivers and road crossings that could be encountered. Such considerations are the most prominent factors to influence pipeline route selection for economic and installation reasons [57]. Simulation of the green hydrogen transportation considered the route altitude and therefore pressure drop along the route. The simulation considered a correlation between the volumetric hydrogen flow-rate and the pipeline size, which was suggested by Menon [58] and Singlitico et al. [59]. The correlation is expressed as:

$${\displaystyle \begin{array}{l}{\displaystyle {{\displaystyle \dot{V}}}_{H_2}(T_b,p_b)=\frac{1.1494}{24}\cdot \left({10}^{-3}\right)\cdot \left(\frac{T_b}{p_b}\right)\cdot 2\sqrt{\frac{D^5\cdot (p_{PIPE,IN}^2-p_{PIPE,OUT}^2)}{Z_{MEAN}\cdot T_{MEAN}\cdot G_{H_2}\cdot L\cdot \mathit{\lambda }}}}\end{array}}$$

(10)

where ${\dot{V}}_{H_2}(T_b,p_b)$ is the volumetric hydrogen flow-rate under standard conditions (T_b = 15°C and p_b = 1 bar); D is the inner pipeline diameter; p_PIPE,IN and p_PIPE,OUT are the inlet and outlet pressures along the pipeline, respectively; T_MEAN is the mean temperature along the pipeline; Z_MEAN is the dimensionless compressibility factor that was assumed to be 1; G_H2 is the gas gravity of hydrogen valued at 0.0696; L is the pipeline length; and λ is the linear friction coefficient. The linear friction coefficient was calculated using the Prandtl–Nikuradse formula [60], which is:

$${\displaystyle \begin{array}{l}{\displaystyle \frac{1}{\sqrt{\lambda }}=-2\cdot \log \frac{\epsilon }{3.71}}\end{array}}$$

(11)

The hydrogen–natural gas blending effects were investigated by accounting for the Wobbe indexes and the heating values of gases before and after the hydrogen injection into the natural gas pipeline network. The Wobbe index is a value given to a specific gas that indicates its combustion performance [61], which calculated as:

$${\displaystyle \begin{array}{l}{\displaystyle IW=\frac{HHV}{\sqrt{G_s}}}\end{array}}$$

(12)

where IW, HHV and G_s are the Wobbe index, the higher heating value of a gas and the gas relative density. When two gases have an identical Wobbe index and they are using the same burner nozzle and the same nozzle pressure, they will generate the same amount of heat [62].

Table 3 provides the natural gas compositions considered in this study, which are the typical compositions that are distributed in natural gas networks in Indonesia. As gas blending took place at the entrance gate of the Jababeka industrial estate, hydrogen was injected into the available distribution natural gas pipeline system. The typical ranges of operating pressures and temperatures for distribution pipelines are 0.017–20.68 bar and –6.7 to 60°C, respectively [16, 63]. In this study, the considered operating conditions under which the gas blending took place were 18 bar and 27°C.

Table 3.

Natural gas compositions [64]

Composition	Mol fraction (%)
Methane	90.12
Ethane	3.38
Propane	1.03
i-Butane	0.24
n-Butane	0.27
i-Pentane	0.12
n-Pentane	0.07
Hexane	0.13
Nitrogen	0.41
Carbon dioxide	4.23

The economic assessment was carried out by applying the cash flow method, assuming that the lifetime of the system is 25 years and the annual tax rate is 25%. The capital expenditures (CAPEX) for the pump and heater integrated with the PEMEC were calculated using the formulae given by Seider et al. [65] while the CAPEX and operating expenditure (OPEX) components for the PEM electrolyser unit were calculated using the approaches suggested by Khan et al. [66]. The expenses for the floating PV modules and batteries were quantified using the economic data given in Tables 1 and 2. The expenses for the hydrogen pipeline were taken into account by referring to the formulae given by Khan et al. [67]. Finally, the costs for the compressor and cooler attached to the pipeline system were obtained by using the formulae given by Seider et al. [65]. The OPEX of the system consists of fixed and variable costs throughout each year. The fixed costs consist of the operating and maintenance costs of all units, non-energy costs associated with the pipeline system, the stack replacement of the PEMEC, tax and loan interest. The variable cost is the cost of electricity for operating the compressor(s) dedicated to the green hydrogen transportation while there is no cost for energy for the pump and heater integrated with the PEMEC, as the electricity required is produced by the FPV.

The hydrogen production cost was calculated by using the add-in Solver program in Microsoft Excel, aiming to obtain an internal rate of return (IRR) that was equal to the weighted average cost of capital (WACC). In that case, the net present value was expected to be zero. The WACC formula is expressed as:

$${\displaystyle \begin{array}{l}{\displaystyle WACC=\frac{E}{V}\cdot Re+\frac{D}{V}\cdot Rd\cdot (1-T_c)}\end{array}}$$

(13)

where E, D and V are the market values of the firm’s equity, debt and total, respectively; Re and Rd are the costs of the equity and debt, respectively; and Tc is the corporate tax rate. The cost of equity considered in this study was 12%, while the cost of debt in Indonesia as a developing country was 6.5%. It was assumed that the shares between the equity and debt are 30% and 70%, respectively. Additionally, the period of lending was assumed to be 10 years. It is worth mentioning that the excess electricity is to be sent to the local electrical grid and sold to the PLN. The set price for produced green electricity is 88 USD cents per kWh, which is 20% lower than the electricity tariff for national customers determined by the PLN for the period of July–September 2023 [68].

As technologies for renewable resources become more mature and the cumulative installed capacity rises, the investment cost is expected to decline over time [69]. Such a reduction was calculated by using the following equations:

$${\displaystyle \begin{array}{l}{\displaystyle lr=1-2^{-a}}\end{array}}$$

(14)

$${\displaystyle \begin{array}{l}{\displaystyle C_t=C_0{\left(\frac{P_t}{P_0}\right)}^{-a}}\end{array}}$$

(15)

where lr is the learning rate listed in Table 4 and a is the learning index; C_t and C_o are the investment costs in year t and in the base year, respectively; and P_t and P_o are the cumulative installed capacities in year t and in the base year, respectively.

Table 4.

Learning rate of each technology

Technology	Learning rate	References
Solar PV	39%	[70, 71]
Li-ion battery	18%	[42, 72]
PEM electrolyser	13%	[42, 52]

2 Results and discussion

The process performance of the PEM electrolyser obtained in this work, as listed in Table 5, is then compared with that given in the literature, in which the attained values are close to those suggested by Zaccara et al. [43]. Those practical values were achieved with a PEMEC active area of 0.3 m². Input of the active area into Equations (4) and (5) suggested by Zhang et al. [44] results in power of 40.66 MW, which becomes the base for the determination of the electricity that needs to be continuously supplied by the FPV system.

Table 5.

PEM electrolyser process performance

Parameter	Value
	This work	Zaccara et al. [43]
Required energy (kWh/kg H2)	54.57	54.80
Water inlet (kg H2O/kg H2)	9.39	9.80
Efficiency (%)	61.1	62.6

Fig. 6 displays the system configuration simulated in HOMER Pro. Electricity produced from the floating PV system passes to the PEMEC in direct-current form, while it is also simultaneously stored in the battery system during the daytime to enable a continuous electricity supply for the PEMEC despite the intermittency of the solar PV. Based on the simulation performed for the PEMEC system, as shown in Fig. 4, the electricity loads for the heater and pump in the process preconditioning section are 417.21 and 8.62 kW, respectively. The total electric load for the process preconditioning section and the PEMEC device was input into the HOMER Pro software. Considering the random variabilities of 5% for day-to-day and time steps during weekdays and weekends, the total estimated demand load for the PEMEC system is 975.85 MWh/day.

Figure 6.

Simulation in HOMER. Image from HOMER Pro modelling software used with permission from UL Solutions.

The design architecture, as given in Table 6, consists of the FPV system with a capacity of 518.4 MW and 3718 batteries. With the suggested design, the electricity demand can be provided from a 100% renewable resource, i.e. solar energy, with an unmet electric load of 0.045%. The simulation result suggests that the cycle charging strategy is more suitable for the proposed system due to the continuous electricity demand behaviour during the day and night. There are 1 524 765 PV modules required for the FPV installation, for which the value was obtained by dividing the capacity of the FPV system by the module rated capacity. Considering the dimension module shown in Table 1 and assuming that there is a 5-inch gap between two solar PV modules, the required water body space for the floating PV installation is 3.57 km², which is under the maximum value constrained.

Table 6.

HOMER simulation results

Configuration	PV (MW)	Number of batteries	Capital cost (USD)	O&M cost (USD/year)	Renewable fraction (%)	Excess electricity (%)	Unmet electric load (%)
PV/BAT	518.4	3718	1.14 B	18.32 M	100	49.4	0.045

Fig. 7a demonstrates the distribution of electricity production per month. The electricity generation peaks in July and August, while it reaches its lowest amount in February. The overall trend in Fig. 7a corresponds to the solar radiation profile shown in Fig. 5. Overall, the FPV system produces 755 671 MWh/year of total electrical energy, with a capacity factor of 16.6%. 161 751 MWh/year of electricity is directly utilized for the PEM electrolyser, while 220 619 MWh/year of electricity is being stored during the daytime so that the PEM electrolyser can be operated continuously, even when the sunlight intensity is low. Additionally, there are 373 301 MWh/year of excess electricity generated throughout a year, which is equal to 49.4% of the total electricity generated. This is because of the intermittent nature of solar energy, which results in the difference between the total energy produced and the electricity demand [73].

Figure 7.

Electricity generated by the floating PV system. (a) Monthly load profile and (b) system dispatch.

Fig. 7b depicts the system dispatch including the profiles of total electricity production and usage during a day. During the day from 06:00 to 18:00, the floating PV modules generate electricity, with the highest activity at 11:00–12:00. From 07:00 to 11:00, battery charging takes place whereas, from 00:00 to 07:00 and 17:00 to 00:00, battery discharging occurs, supplying electricity for the PEM electrolyser. Surplus electricity is gained between 09:00 and 17:00 as the rate of battery charging decreases, while the floating PV still gets sufficient sunlight to produce electricity.

In addition to the simulation for hydrogen and electricity generation, this study also modelled hydrogen transportation along the pipeline system. In total, the length of the pipeline following the route suggested in Fig. 3 is 36.74 km. Fig. 8 shows the altitude profile along the pipeline route. The elevation difference between the initial point and the hydrogen injection location is 50.5 m, so the end point is known to have a lower land level.

Figure 8.

Altitude profile along the hydrogen pipeline route

Fig. 9 displays the flow-sheeting of the hydrogen pipeline transportation obtained using Aspen HYSYS. The minimum inner diameter of the pipeline, which was calculated by using Equations (10) and (11), is 5.58 inches. The inner diameter considered in the simulation is slightly higher than that, at 7.98 inches. This is the standard inner diameter for pipes with standard thickness, with a nominal diameter of 200 mm. The hydrogen inlet pressure is 15 bar, while, due to the transportation route with certain elevations, the hydrogen outlet pressure at the end point decreases to 14.43 bar. Therefore, a set of compressor and cooler units were employed to adjust the temperature and pressure of the flow. The set outlet pressure from the compressor is 18 bar, while the outlet temperature coming out of the cooler was adjusted to 27°C.

Figure 9.

The hydrogen pipeline transportation simulation

The Wobbe indexes and the heating values of the gases obtained from the mixing simulation in Aspen Plus are presented in Table 7. Those parameters for hydrogen and natural gas are similar to those presented in the literature [20, 74] while the Wobbe indexes for both pure hydrogen and natural gas are inside the standard Wobbe index bandwidth commonly specified for the exit points of gas transportation networks for high calorific gas, which is from 47.0 to 55.7 MJ/m³ [75]. Based on the simulation, it is observed that the existence of hydrogen in the blended gas only results in a marginal decrease of <2% in the Wobbe index compared with that of natural gas. On the contrary, the addition of hydrogen leads to a slight increase in the heating values on a mass basis by <2%, as a result of the smaller density of hydrogen compared with natural gas.

Table 7.

Wobbe indexes and heating values of gases before and after the gas blending

Parameter	Hydrogen	Natural gas	Blended gas
Wobbe index (MJ/m3)	48.49	50.81	49.94
Higher heating value (MJ/m3)	12.79	40.60	38.61
Higher heating value (MJ/kg)	142.19	49.18	49.95
Lower heating value (MJ/m3)	10.78	36.59	34.74
Lower heating value (MJ/kg)	119.83	44.32	44.95

The energy flow through the overall system is shown in the Sankey diagram demonstrated in Fig. 10. To provide hydrogen for industrial end users with energy of 596 MWh, the solar energy required is equivalent to 11 837 MWh. The major PV losses are caused by the solar PV nominal efficiency that is typically encountered in solar PV systems [42].

Figure 10.

Sankey diagram of the proposed system

The green hydrogen production cost was calculated by primarily calculating the CAPEX and distribution of OPEX throughout the lifetime of the system. The CAPEX of the system is 1.53 billion. The expenses for the floating solar PV modules and batteries dominate the CAPEX fractions, in which they account for 45.6% and 49.2%, respectively, of the total capital expenditure. The other fractions are composed of the CAPEX values of the PEM electrolyser and hydrogen pipeline, which are 3.7% and 1.5%, respectively.

Fig. 11 displays the green hydrogen production cost, which is USD 26.95 per kg of hydrogen. The production cost is composed of costs of the elements of the system, including the floating solar PV modules, Li-ion batteries, PEM electrolyser and pipeline system. Overall, the CAPEX and the fixed operating cost significantly contribute to the production cost, with shares of 51.4% and 48.5%, respectively, while the contribution from the variable cost is marginal, at 0.1%. Such costs are spent only on the energy required to operate the compressor connected to the hydrogen pipeline system. The green hydrogen production cost was obtained when the net present value was set to zero. With that set, the calculated IRR was identical to the value of WACC, which is 7.0%.

Figure 11.

Elements of hydrogen production costs

The challenge from an economic aspect is caused by two essential factors: (i) heavy electricity requirement as a result of the continuous demand for hydrogen from industries, which is even heavier during the daytime, as electricity must be partially stored for the night-time necessity; and (ii) current capital investments for solar PV and batteries that are still expensive. The result presented in Fig. 11 clearly shows that the expenses for Li-ion batteries comprise the largest portion of the green hydrogen production costs, at 40%. This leads to a cost that is far beyond the green hydrogen production costs, which are in the range of USD 3.5–8.5 per kg, from systems that do not rely heavily on the use of batteries because of the continuous 24/7 hydrogen production [9, 76–79].

As stated in the ‘Introduction’ section, studies on industrial-scale, continuous green hydrogen production have been rarely presented to date. Lee et al. [80] are among the few who have discussed an outlook for green hydrogen production for industries when considering seasonal solar radiation for photovoltaic technology and the utilization of alkaline water electrolysis technology and lithium-ion batteries. Their finding suggests that the green hydrogen production costs could reach values of above USD 25 per kg when a higher electrolyser scale and decreased battery capacity are applied. In their study, the assessed ranges for the electrolyser scale and battery capacity were limited to 0.5–3 MW and 1–5 MWh. When much higher electrolyser power is needed, with the utilization of an energy storage system such as batteries, one could anticipate a hike in the green hydrogen production cost.

Currently, there is no existing system of green hydrogen production for industrial decarbonization in Indonesia, and the use of grey hydrogen for feedstocks and utilities still dominates. When a system without FPV and batteries is proposed, the local grid needs to continuously supply electricity to the PEM electrolyser unit in order to produce hydrogen. When a system with FPV and without batteries is presented, it can only rely on green electricity generated from the PEMEC powered by FPV during the daytime, while it requires electricity from the local grid during the night. Neither of these design scenarios generates green hydrogen as an end product because, currently, electricity from the Indonesian local grid is mainly generated from coal-fired power plants.

From an environmental perspective, the proposed system of green hydrogen via a PEMEC powered by FPV and equipped with batteries outperforms the other two design scenarios because it does not rely on fossil fuel utilization, therefore improving the carbon footprint of the process. As depicted in Fig. 12, from an economical perspective, the design scenario without FPV and batteries requires the lowest production cost among other scenarios utilizing PEMEC because Indonesia’s electricity tariff is relatively low, i.e. 0.11 USD per kWh [68] while the design scenario of green hydrogen proposed in this study generates the highest production cost due to the current renewable technology prices. All three design scenarios using a PEMEC to produce hydrogen require production costs that are higher than those of the existing conventional technologies using coal gasification (CG) and natural gas steam reforming (SMR).

Figure 12.

Hydrogen production costs based on technologies. The costs for grey hydrogen from SMR and GC were extracted from Zhen et al. [81].

Additionally, the sensitivity analysis regarding the changes of some economic parameters, i.e. lending rates and the prices of FPV modules and Li-ion batteries, is presented in Fig. 13. The values of these economic parameters for the base cases are given in the ‘Research methodology’ section, in which the lending rates and the prices for FPV modules and Li-ion batteries are 6.5%, USD 1343.85 per kW and USD 562 per kWh, respectively. The assessed changes in each economic parameter are based on potential shifts in current practice. The lending rate in developing countries such as Indonesia can reach up to 10%. Therefore, the relative difference in the lending rate taken into account is ±50%. The relative changes in battery and FPV prices are 20% and 25%, respectively, considering the ranges of realistic prices for other types of Li-ion batteries and FPV panels. The sensitivity analysis shows that the hydrogen production cost changes by 4–7% when each economic parameter is reduced or increased.

Figure 13.

Sensitivity analysis of economic parameters in the green hydrogen production system

The projection of the hydrogen production costs until the year 2060 is presented in Fig. 14. The cost projection for green hydrogen produced from the proposed system was calculated by considering the learning rates for the solar PV, Li-ion battery and PEM electrolyser technologies. The cost projection for hydrogen produced from fossil fuels, i.e. CG and SMR, was obtained from Zhen et al. [81], considering the prices of bituminous coal and natural gas of USD 75.4/ton and USD 0.41/m³, respectively. The production cost of green hydrogen produced from the proposed system already decreases by 69.4% in the year 2030, while the green hydrogen production cost is expected to be equal to that of grey hydrogen produced via CG and via natural gas steam reforming in 2043 and 2047, respectively. In 2060, the green hydrogen production cost gets reduced by 91.9% to USD 2.19/kg. The significant decrease in the green hydrogen product cost can be expected, as the expenses for PV modules contribute the most to the overall CAPEX, and solar PV technology has a relatively high learning rate, as given in Table 4.

Figure 14.

Projected hydrogen production costs. The costs for grey hydrogen from SMR and GC were extracted from Zhen et al. [81].

3 Conclusion

This study has successfully assessed green hydrogen production for blending with natural gas for the purpose of industrial decarbonization. Despite the intermittent nature of the solar energy source, this work suggests the technical viability of the proposed system for producing green hydrogen with the continuous flow-rate of 7.5 MMSCFD (0.21 million standard cubic metres per day). Such continuous demand is typically required by industrial end users. The floating solar photovoltaic and batteries system designed in this study can provide the required electricity for the utilization of PEM electrolyser technology, in which the system consists of the FPV with a capacity of 518.4 MW and 3718 lithium-ion batteries, each with a nominal capacity of 210 kWh. The water body space requirement for the floating solar PV modules is 3.57 km², which is feasible considering the actual available area of the water body evaluated in this study. Furthermore, one can expect marginal effects on the gas combustion performance due to the blending of green hydrogen with natural gas, as indicated by the Wobbe index and heating values of the blended gas obtained in this work.

The economic assessment suggests challenges in the application of this technology due to the current excessive capital investment and high green hydrogen production cost. The cost of green hydrogen produced from the proposed system is USD 26.95/kg, which is far from the costs of green hydrogen generated from systems that do not rely heavily on batteries for continuous 24/7 hydrogen production. However, these challenges are expected to be minimized as a result of decreasing investment costs over time, which is indicated by the technology learning rates. By 2030, the green hydrogen production cost is expected to decrease by 69.4%, while, by 2060, the green hydrogen production cost specifically designed to fulfil the continuous demand from industries is projected to be already lower than grey hydrogen production costs. Furthermore, national government interventions will help in expediting the deployment of green hydrogen technology. Although the Indonesian government has set a strategy to escalate the contribution of renewable energy in the Indonesian energy mix strategy, as declared in the General Plan of National Energy, RUEN [82], stronger and more directive policies regarding renewable-energy incentives, especially those for green hydrogen, are crucially needed. Currently, most existing green hydrogen incentives apply in Europe, with Japan and South Korea as the leading countries in Asia [34, 83].

Nomenclature

Nomenclature
 
• a

  learning index

 
• $A$

  active area (m²)

 
• C_o

  investment costs in the base year (USD)

 
• C_t

  investment costs in year t (USD)

 
• D

  inner pipeline diameter (m)

 
• $E$

  amount of power (W)

 
• G_H2

  gas gravity of hydrogen

 
• $G_s$

  gas relative density (kg/m³)

 
• $H_2^{perm}$

  permeated hydrogen (bar)

 
• HHV

  higher heating value of a gas (MJ/kg)

 
• $I$

  electric current (A)

 
• IW

  Wobbe index

 
• $j$

  current density (A/m²)

 
• L

  pipeline length (km)

 
• LHV_H2

  hydrogen lower heating value (MJ/kg)

 
• lr

  learning rate (%)

 
• ṅ_H2

  hydrogen production rate (kg/h)

 
• $O_2^{perm}$

  permeated oxygen (bar)

 
• $p_b$

  standard pressure (bar)

 
• P_o

  cumulative installed capacities in the base year (kton)

 
• p_PIPE,IN

  inlet pressures along the pipeline (bar)

 
• p_PIPE,OUT

  outlet pressures along the pipeline (bar)

 
• P_t

  cumulative installed capacities in year t (kton)

 
• Rd

  costs of debt (%)

 
• Re

  costs of equity (%)

 
• $T_b$

  standard temperature (°C)

 
• Tc

  corporate tax rate (%)

 
• T_MEAN

  mean temperature along the pipeline (°C)

 
• $V$

  cell voltage (V)

 
• $V_{act}$

  activation overpotential losses (V)

 
• $V_{con}$

  concentration overpotential losses (V)

 
• ${\dot{V}}_{H_2}(T_b,p_b)$

  volumetric hydrogen flow-rate under standard conditions (T_b = 15°C and p_b = 1 bar)

 
• $V_o$

  reversible voltage (V)

 
• $V_{ohm}$

  ohmic potential losses (V)

 
• Ẇ_electrolyzer

  power input to the PEMEC (MJ)

 
• Z_MEAN

  dimensionless compressibility factor

 
• $\mathrm{\Delta }P$

  pressure difference between cathode and anode and the active area (bar)

 
• $\epsilon$

  relative pipeline roughness

 
• $\eta$

  electrolyser efficiency

 
• λ

  linear friction coefficient.

Acknowledgements

R.M. thankfully acknowledges Alya Nurul Shafira for fruitful discussion during the manuscript preparation.

Author contributions

RM: data collection and analysis, writing—original draft, conceptualization, funding acquisition. ASPP: data collection and analysis, visualization, writing—original draft. MRM: data collection and analysis, writing—original draft. WWP: conceptualization, writing—original draft, review, validation.

Conflict of interest statement

None declared.

Funding

This work was funded by the Osaka Gas Foundation of International Cultural Exchange Year 2022/2023 (PKS-1813/UN2.F4.D/PPM.00.00/2022).

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.