Green hydrogen and wind synergy : Assessing economic benefits and optimal operational strategies

Volatile electricity prices have raised concerns about the economic feasibility of wind projects in Finland. This study assesses the economic viability and optimal operational strategies for integrating wind-powered green hydrogen production systems. Utilizing modeling and optimization, this research evaluates various wind farms in Western Finland over electricity market scenarios from 2019 to 2022, with forecasts extending to 2030. Key economic metrics considered include internal rate of return, future value, net present value (NPV), and the levelized cost of hydrogen (LCOH). Results indicate that integration of hydrogen production with wind farms shows economic benefits over standalone wind projects, potentially reducing LCOH to € 2.0/kgH 2 by 2030 in regular and low electricity price scenarios, and to as low as € 0.6/kgH 2 in high-price scenarios. The wind farm with the highest capacity factor achieves 47% reductions in LCOH and 22% increases in NPV, underscoring the importance of strategic site selection and operational flexibility.


Introduction
Human-caused global warming has led to increasing temperatures all around the world due to large-scale exploitation of fossil fuels and greenhouse gas emissions.To mitigate climate change, countries have agreed to reduce greenhouse gas emissions, most notably in the 2015 Paris climate agreement.As a party to this agreement, Finland, a northern European country, has pledged to become carbon neutral by 2035 [1].Wind power is one of the most important sources of renewable energy for Finland to achieve its climate goals, given the country's good wind conditions.At the end of 2022, Finland's wind power capacity was 5677 MW, with an expected increase to over 8500 MW by 2025.In 2022, wind power generated 14% of Finland's total electricity consumption [2].However, with increasing shares of wind power, the electricity price volatility will also increase as prices during windy hours can drop to low or even negative values.As more and more power production will occur during windy hours, if demand cannot match production, wind power plants may need to be shut down or curtailed.
One solution to tackle renewable energy curtailment is Power-to-X (P2X), converting surplus renewable electricity into energy forms or chemicals like green hydrogen and synthetic fuels or products, which can be stored, transported, or used in various sectors [3,4].By utilizing excess renewable power production, P2X helps to balance the energy system, ensuring that the generated power does not go to waste.In comparison to other possible use cases of excess electricity, such as heat, hydrogen and its derivatives offer higher value for the producer and are therefore more attractive economically.This process significantly reduces renewable curtailment, optimizing the energy system and supporting the transition towards a more sustainable and flexible energy landscape [5].The European Union(EU) has identified green hydrogen as a key component in its goal of decarbonization, outlined in the Hydrogen Strategy in 2020.As part of the REPowerEU initiative, the EU has set a target of 10 Mt of annual green hydrogen production by the year 2030 [6].The Finnish government later announced a target of producing 1 Mt of hydrogen annually by the same year, covering 10% of the whole EU's production target [7].
If renewable electricity is used, the produced hydrogen is so-called green hydrogen [8].Currently, almost all of world's hydrogen is produced from fossil fuels, mainly with steam methane reforming from natural gas, which produces carbon dioxide [5].Capturing the CO 2 produced by steam methane reforming allows mitigating some of the emissions from the process but would still rely on fossil fuels for hydrogen production.Alternative hydrogen production methods include methane pyrolysis, which also produces carbon black as a byproduct and thermochemical water splitting driven by nuclear reactions [9].The main water electrolysis technologies to produce green hydrogen are alkaline, proton exchange membrane (PEMEC) and solid oxide electrolyzers.Alkaline electrolysis is the most mature technology with already commercial equipment available [10].Solid oxide electrolyzers offer high production efficiency due to high operating temperatures but are still under development, with further research on material stability needed [11].PEMECs offer higher efficiencies compared to alkaline electrolyzers due to higher current density.In addition, they produce high purity hydrogen gas and have a compact design [12].Importantly, PEMECs can operate dynamically, making them ideal when coupled with fluctuating renewable energy production [13].However, PEMECs use scarce metals such as platinum and iridium as catalyst material.Research on PEMEC technology is focused on developing alternative catalyst materials or alloys that would lower the need for these expensive materials [14].Water consumption and quality are important factors in the electrolysis process with varying requirements for different technologies.PEMECs are sensitive to cationic and anionic impurities in feedwater such as Na + , Ca 2+ and Cl − , which are prevalent in both tap water and seawater, latter having significantly higher ions concentrations.AEL is more resistant to ionic impurities, but organic and inert impurities act as catalyst poisons in the cell.Therefore, sufficient water pre-treatment is required to prolong lifetime of the electrolyzer [15].
Many studies on the dedicated production of hydrogen in hybrid renewable energy systems can be found in literature.Case studies based on geographical location have been analyzed for example in Uruguay [16], Saudi Arabia [17] and western China [18].The focus of these studies was to optimize a hydrogen energy system economically.However, the case studies often assumed the efficiency of an electrolyzer to be constant when coupled with intermittent renewable energy sources.For example, Ghenai et al. [19] analyzed an off-grid system for a residential area based on solar power, hydrogen production and fuel cells.The hydrogen production rate is simply calculated on a single efficiency value regardless of the load.Hernandes-Gomez et al. [20] compared different modelling approaches for PEMECs and suggest that the widely used approach is simplistic and does not take into account delays in ramping hydrogen production up and down dynamically.Kojima et al. [21] studied the effects of fluctuating power sources on water electrolysis and found that electrolyzer degradation was accelerated with changing currents.Experimental work conducted by Rakousky et al. [22] showed that while dynamic changes in current can have even positive effects on degradation, short and sharp variations decrease the lifetime and efficiency of electrolyzers.

Novelty and contributions of this study
Dedicated wind power for hydrogen production with a PEMEC has been studied for example in the Sultanate of Oman [23].Wind speed data were used to assess both the levelized cost of electricity and hydrogen in the case study.In the Finnish context, only a few studies can be found in the literature on green hydrogen production from wind power.Elberry et al. [24] analyzed the role of hydrogen storages in the Finnish electricity system as long-term energy storage.Koivunen et al. [25] also analyzed the whole Finnish energy system to understand the impact of power-to-hydrogen on CO 2 emissions based on different scenarios.Korpås et al. [26] studied optimal hydrogen production and storage in a power market context in Norway.The study used hourly electricity prices for only one year as a case study, excluding the effect of price levels on the results.
Despite the various previous studies, a research gap exists regarding the economic feasibility of wind-electrolytic hydrogen production across various energy market scenarios, including periods of low and high electricity prices, such as those experienced in 2020 and 2019, and during the 2022 EU energy crisis.This gap highlights the need for further investigation into how changing energy market dynamics affect the viability of such systems, a vital aspect of renewable energy economics that remains, to the authors' best knowledge, understudied.In response to this gap, this study examines the implications of the growing hydrogen economy and the significant fluctuations in electricity prices in Europe, from the lows of 2020 to the highs during the 2022 EU energy crisis.With the phasing out of wind power subsidies in Finland, which requires projects to operate on a market basis [27], there is a clear necessity for wind power producers to assess the relative profitability of selling electricity directly to the grid or converting it into hydrogen.This research provides an analysis of the operational and economic feasibility of on-site green hydrogen production linked to a Finnish wind farm under various current and future energy market conditions.It simulates the system's optimal operation to maximize revenue from hydrogen and day-ahead electricity markets, analyzing economic metrics such as the internal rate of return (IRR), net present value (NPV), payback period, and levelized cost of hydrogen (LCOH).Unlike previous studies, this research integrates the dynamic operation of a PEM electrolyzer with variable efficiency, covering historical market conditions from 2019 to 2022 and projecting future scenarios up to 2030, considering technological advancements and cost reductions.This study considers a variable electrolyzer efficiency, improving the accuracy of economic feasibility assessments.It also conducts a sensitivity analysis to evaluate the impact of variables such as electrolyzer size and fluctuating electricity and hydrogen market prices on the system's operations.The contributions of this research are twofold: it enhances the understanding of the economic feasibility and operation of windelectrolytic hydrogen production systems under different market conditions, and it assesses the competitiveness of these systems against standalone wind farms.By offering a holistic examination of optimal operations, future trajectories, economic metrics, and sensitivity factors, the study provides stakeholders in the sustainable energy sector with valuable insights for making informed decisions regarding the integration of renewable energy sources and hydrogen production technologies in varying energy market conditions.

Methods
The methodology of this research encompasses modeling of the key components involved in wind-electrolytic hydrogen production, as depicted in Fig. 1, including wind turbines, a PEM electrolyzer, and compressor, along with detailed current and future cost assumptions for each.Initially, these components are modeled to reflect their real-world operational characteristics and economic factors.Following this, an objective function is formulated aiming to maximize revenue from both electricity and hydrogen markets, while adhering to operational constraints inherent to the system.Subsequently, key economic metrics are calculated to assess the system's economic feasibility.The methodology further extends to conducting a sensitivity analysis, assessing the impact of a range of input parameters on the system's performance.Fig. 1 illustrates the schematic of the proposed system.The design assumes that all the units are adjacent to ensure optimal functionality.Furthermore, it is presumed that a single operator retains ownership and control over the entire integrated system.
Wind power can be sold directly to the day-ahead electricity market, especially during times when electricity prices peak.During periods of lower electricity prices, it is often more cost-effective to divert wind power towards hydrogen production via the electrolyzer.Additionally, when wind generation is low or electricity prices dip, extra power can be sourced from the main grid to continue hydrogen production with cheap electricity, thereby decreasing the LCOH.The primary goal of this study is to optimize operational strategies in both the hydrogen and electricity markets to maximize investor profitability.Key decision points include determining the ideal times to market the wind power versus using it for hydrogen production.The produced hydrogen can be transported as compressed gas in bottles or pipelines, as a liquid, or processed into another compound.For energy storage applications, hydrogen is stored in large facilities and used later for power production [28].However, large-scale storage systems are costly and still under development, thus they are beyond the scope of this study.Hydrogen gas is compressed to 180 bar or higher for bottling, with some tank types capable of withstanding up to 700 bar [29].Individual tanks or cylinders are loaded onto trucks, commonly known as tube trailers, for market transportation.
Additionally, the alternative of injecting produced hydrogen directly into a hydrogen pipeline is considered; however, since no hydrogen pipelines yet exist in Finland, this study considers hydrogen gas bottling as the export method from the electrolyzer.The oxygen byproduct generated during the electrolysis process is earmarked for sale for medical applications.It is assumed that the transportation of the harvested hydrogen and oxygen occurs without any losses.It is feasible to harness the waste heat generated by the electrolyzer, potentially integrating it into district heating networks, although such a proposition requires financial assessment and is location-specific.The subsequent subsections explain the modeling process and mathematical equations used to simulate each individual component.

Wind farm
This subsection discusses the modeling of the wind farm and its specific characteristics, situated in Finland.The aggregated capacity and other important attributes, such as turbine hub height and model, are set to match real-life projects in the selected regions [2].To facilitate comparisons, all wind farms are assigned the same size at each of the designated study locations.Table 1 details the locations and capacities of the wind farms considered in this study.
Hourly wind speeds for each year and site are measured at the specified hub height, as detailed in Table 2 [31].The power produced by a wind farm is determined using the following equation [32], where P wind (t) represents the hourly wind power output at time t, and P rated is the rated power of the wind turbine.V wind (t), V ci , V co , and V rated denote the hourly wind speed, wind turbine cut-in, cut-out, and rated speeds, respectively.
In evaluating the economic parameters associated with wind farms, a range of current and future estimates has been reported in the literature, reflecting both CAPEX and O&M costs.Table 2 lists the techno-economic parameters for the selected wind turbines in this study.To enhance accuracy and ensure fairness in comparing results, this study uses the average of the values reported in various references.While the most reliable reference was given priority in some instances, values from other references were also included to provide additional information.
The 2023 values are based on the mentioned references, and projections for 2030 are derived from estimations within these sources, reflecting anticipated trends and technological advancements in the industry.The assumed CAPEX includes wind turbine cost, cabling costs, and other plant-related costs.It is important to note that all financial values, denominated in euros (€), have been normalized to 2023 values.This normalization ensures that the economic outcomes remain comparable.A Vestas turbine is selected for this study due to its widespread use in the Nordic region, including Finland, where it is a preferred choice for many current and upcoming wind projects [33].

Table 1
The location and area of considered case study wind farms.Techno-economic parameters of the wind turbines used in this study [33][34][35][36].

Hydrogen production
The modeling and techno-economic parameters of the selected electrolyzer, a PEMEC in this study, are discussed in this section.The reported efficiency range for commercial PEMECs, including balance-ofplant components, is 50-83 kWh/kgH 2 [37].Efficiencies of 55.6 kWh/kgH 2 for 2020 and 49.70 kWh/kgH 2 for 2030 were reported by Sens et al. [36].Efficiency improvements are limited due to thermodynamic principles that cap maximum efficiency [38].It is important to note that the cited efficiency values refer to full-load operation of the electrolyzer.Unlike conventional thermal power plants, where efficiency improves with load, electrolysis systems see a decline in efficiency as the load increases [39].Fig. 3 illustrates how the efficiency of the PEMEC varies with the load, based on the assumed full-load efficiency for the 2023 model of the electrolyzer (55 kWh/kgH 2 , as shown in Table 3) [39][40][41].For the 2030 projections, a full-load efficiency of 50 kWh/kgH 2 is adopted, and the curve adjusts accordingly.Note that the mentioned efficiencies are based on the lower heating value (LHV) of hydrogen.
Regarding the economic parameters associated with electrolyzers, a range of estimates is available in the literature, reflecting the emerging nature of the technology.Christensen et al. [42] projected an investment cost of €1100/kW for 2020, reducing to €650/kW by 2030, with a wide uncertainty range of €385-2068/kW [43].The International Renewable Energy Agency reported CAPEX estimates ranging from USD700 to 1400/kW [37].Similarly, the International Energy Agency's report for G20 Japan provided current estimates of 1100-1800 USD/kW, with projections ranging from 650 to 1500 USD/kW by 2030 [5].Hancke et al. estimated the stack replacement cost at 35% of the initial PEMEC CAPEX [43].
The operational lifetime of contemporary PEMECs was reported to be between 60,000 and 100,000 h [5,36,37,44].Table 3 provides a summary of the techno-economic parameters for the electrolyzer evaluated in this study.In this research, the primary focus is not on optimal sizing.Instead, the initial capacity of the electrolyzer is set at 80% of the wind farm's size to manage the inherent intermittency of wind power production effectively.This strategic capacity setting aims to optimize the system's economic efficiency by reducing investment costs and avoiding the inefficiencies associated with potential oversizing.To thoroughly examine the implications of this sizing strategy, a sensitivity analysis, detailed in Fig. 12, explores a range of electrolyzer sizes relative to the wind turbine capacity.This analysis provides insights into the system's performance and resilience under various configurations.
PEMECs are influenced by the quality of inlet water.Various prevalent impurities can impact their performance, the purity of the hydrogen produced, and their overall lifespan.Purification of feed water introduces additional expenses, operational intricacies, and design constraints.However, if the purification systems fail, this can result in the deterioration or contamination of the membrane electrolyte or catalyst materials [46].In this study, it is assumed that all selected case studies have access to pristine, fresh water from nearby lakes and rivers at no cost, thus avoiding the expense of water desalination.Nonetheless, the cost of water purification is considered since deionizing the water for the electrolyzer feed notably extends its lifetime [46].The transportation of this water from the sea to the location of the electrolyzer is neglected for the sake of model simplification, as water transportation costs per pipeline over distances up to 1000 km have been calculated to be less than €0.06/kgH 2 [47][48][49].
The current market price of green hydrogen is reported to be between €5-6/kgH 2 [50] and €3-8/kgH 2 [51].The Hydrogen Valley platform, a global collaborative hub for large-scale hydrogen flagship projects, categorizes average market sales prices from under €1-2/kgH 2 to above €10/kgH 2 [52].According to this platform, approximately 32% of all projects report hydrogen sales prices between €4-6/kgH 2 .For this study, the assumed market price of hydrogen is set at €4/kgH 2 for the years 2019-2022 and is projected to decrease to €3/kgH 2 by 2030, in line with Hydrogen Europe's strategy for reducing hydrogen costs [53].Additionally, the price of produced oxygen is assumed to fall within the range of €20-40/tonO 2 [32].Table 4 details the assumed market prices for both hydrogen and oxygen.
If the system draws electricity from the main grid, in addition to the hourly day-ahead market prices [54], electricity transmission costs and electricity taxes [55] must also be paid.The specific amounts of these costs are in Table 5.
After production, hydrogen is compressed using a reciprocating compressor with a piston mechanism before being stored in a highpressure hydrogen tank [56].Energy requirements for compressors vary, with reports indicating a range from as low as 0.4 kWh/kg [42] to as high as 2.2 kWh/kg [39].The energy demands of compressors differ based on their specific application.This study focuses on hydrogen transport using trucks, so the compressor is sized to meet this requirement.This approach ensures a more precise model representation than relying on generic values from literature intended for different applications.The energy required by the compressor unit (kJ/kg) is Fig. 3.The assumed variation of efficiency (specific electricity consumption) of the PEMEC with load in the simulations.
calculated using the equation for ideal conditions [56]: where k (1.4) is the ratio of the specific heats (cp and cv), R ʹ H2 is the hydrogen gas constant (4.12 kJ/kg K), T in (K) is the hydrogen inlet temperature (25 • C) and P in and P out are inlet and outlet pressures, respectively, listed in Table 6.The inlet pressure of hydrogen from a PEMEC, entering the compressor, can range from 10 to 30 bar for similar applications [56].Manufacturer data indicates that inlet pressure can vary between 30 and 100 bar if pressurized hydrogen is required [57].The required compressor electric power is calculated with Eq. ( 3), where ṁH2 is the hydrogen production rate, and ƞ is,c , ƞ m , and ƞ e are the compressor isentropic efficiency, the mechanical efficiency, and the electric generator efficiency, respectively [56].It is assumed that the compressor also runs on the same electricity that powers the electrolyzer.

Problem formulation and scenarios
This subsection explains the modeling process of the system, including the objective function and constraints, using EnergyPRO [58], a software known for modeling complex energy systems [59].The objective is to maximize the total annual revenue for the entire system, calculated using Eq. ( 4).R H2 , R O2 , and R DA denote the revenues from selling hydrogen, oxygen, and power in the day-ahead electricity market.The total annual revenue is calculated by summing these revenue streams over an entire year.
The annual variable O&M cost (C varOM ) expressed in Eq. ( 5), is calculated by summing the operational costs of the system, including the cost of electricity import from the grid (C DA PEM ), the variable O&M cost of wind turbines (C OM wind ), and the water consumption cost (C water ).The electricity consumption cost of the compressor is taken into account in the electricity cost of the electrolyzer.
The objective function is expressed via Eq.( 6), which aims to maximize the net profit of the system: The operational problem is transformed into a mixed-integer linear programming (MILP) model and solved using the software.Eqs. ( 7)-( 9) represent the constraints of the system.Eq. ( 7) indicates that the total wind power delivered to the electrolyzer (P wind PEM ) plus the wind power directly sold in the day-ahead market (P wind DA ) should be equal to the total wind power production from the wind farm (P wind ).Eq. ( 8) states that the total electricity consumption of PEMEC and compressor (P PEM (t)) equals the electricity imported to the PEMEC from the grid (P DA PEM (t)) and the wind farm (P wind PEM (t)).Also, the electrolyzer is bound to operate within its minimum and maximum load at each time step, indicated by Eq. ( 9).P wind DA (t) + P wind PEM (t) = P wind (t) (7) The evaluated economic metrics of the system in this study are LCOH, IRR, NPV, and payback period.Each metric provides a different perspective on profitability, investment return, and overall value.The LCOH takes into account a broad array of factors, not just the initial CAPEX, but also operational and maintenance costs, electricity costs, and other recurrent expenses throughout the plant's lifetime.This makes it an effective metric for comparing different scenarios featuring various plant sizes and energy mixes.LCOH is calculated via Eq.( 10) [32]: C fixedOM and C varOM are the annual fixed and variable OM costs of the system, calculated by Eq. ( 5).R DA is the revenue gained from selling excess wind power directly to the electricity market (electricity export).When calculating LCOH, only the associated costs should be considered.However, as all of the wind power does not go through the electrolyzer and a part of it is sold in the market, the revenue from selling this part of the wind power should be considered while calculating LCOH.C CAPEX,a and C rep,a are the annual capital and replacement costs to replace all components and parts that wear out during the plant's lifetime.These parameters are calculated using annuity method, converting a one-time investment cost, into annual fixed costs over the lifetime of the plant.Eqs.(11) and ( 12) express the annualized capital and replacement costs [56]: C CAPEX and C rep indicate total capital investment and replacement costs a Winter days (December-February) consumption fee (7:00-21:00)/other time consumption fee.

Table 6
Techno-economic parameters of the hydrogen compressor used in this study [36,39,42,56]. of all units.r and n represent interest rate, which is considered as 6% in this study and lifetime of the system, whereas t in the above equations denotes the year in which the replacement costs are paid.In addition to LCOH, the IRR, payback period, and NPV are also calculated.NPV value of the system is expressed with Eq. ( 13) as follows [60]: Where Revenue is calculated using Eq. ( 4) and includes all the revenues from hydrogen and electricity markets.In this study, the year 2023 is chosen as the reference year (present year), and all the monetary values, including costs and revenues, are converted to this year using Eq. ( 14), where FV and PV denote the future and present values, respectively [60]: To account for different energy market conditions, different scenarios are analyzed for the years 2019-2022, each with distinctive characteristics as follows.
• 2019 served as a baseline, with moderate wind speeds and electricity prices averaging €44/MWh in the Finnish bidding area [54].
• 2020 stood out for its extremely low electricity prices, averaging only €28/MWh, influenced by strong wind conditions and a noticeable drop in industrial electricity demand due to COVID-19 lockdown measures [54].
• 2021 witnessed an average price of €72/MWh, partly driven by the post-pandemic economic recovery and compounded by low wind and prevailing dry conditions across the Nordic regions [54].
• 2022 was marked by an unprecedented spike in electricity prices, with an average of €154/MWh, attributed to the European energy crisis and the cessation of natural gas supply from Russia [54].
Analyzing performance across these years allows for a comprehensive understanding of the fluctuations in energy market prices and wind conditions, aiding in the evaluation of the system's operation and feasibility under diverse circumstances.For simplicity and ease of reference, each scenario is labeled by combining the wind farm number and the year; e.g., WF1,19 refers to wind farm 1 analyzed for 2019.For 2030 scenarios, the study utilizes a decade-long average of wind speeds recorded at the sites.To streamline scenario analysis, only wind farm 4, which has the best wind conditions, is selected.Three hypothetical price structures for day-ahead electricity in 2030 are developed: WF4,30-low reflects the low electricity prices of 2020.WF4,30-reg and WF4,30-high represent the regular and elevated electricity prices observed in 2019 and 2021, respectively.

Results
This section presents the results from the simulation and technoeconomic analysis of case studies in different scenarios.

Optimal operation
In this subsection, the simulation results of the optimal operation of the proposed hydrogen production system are detailed.Fig. 4 illustrates the annual wind energy output from each wind farm and their corresponding average annual capacity factors.
Fig. 4 shows that wind farm 4 yields the highest wind production.In contrast, wind farms 1, 2, and 3 exhibit comparable production values, with capacity factors fluctuating between 30% and 40%.Notably, 2021 presented the most favorable wind conditions among the studied years, contributing, in part, to that year's low electricity prices.Fig. 5 illustrates the optimal operation of the system, highlighting the balance between electricity imports (positive values) and exports (negative values) to and from the main grid across various scenarios and years.The figure also details the annually produced hydrogen for each scenario.
It is noteworthy from Fig. 5 that the wind farms imported and exported electricity every year.This dynamic operation demonstrates that, at times, it is beneficial to purchase electricity to produce hydrogen, while at other times, it is advantageous to sell the excess wind-generated power.In years with high electricity prices, such as 2022 and 2030-high, there is a pronounced shift towards exporting more wind-generated electricity to the day-ahead market, rather than importing electricity from the grid for hydrogen production.For example, 55% and 40% of the annual wind production were exported to the market in the WF4,2022 and WF4,2030-high scenarios, respectively.Consequently, the annual hydrogen production in these scenarios is considerably lower than in others.
Conversely, in years with moderate to low electricity prices, such as 2019, 2020, 2030-low, and 2030-reg, it is more economical to use the majority of wind power for hydrogen production and import inexpensive electricity from the main grid.In the WF4,20 scenario, only 4% of the wind production was sold in the electricity market, with the remainder used for hydrogen production.This accounts for the higher hydrogen production observed in these scenarios.Interestingly, even in    2021, despite high electricity prices, more electricity was imported to produce hydrogen.When comparing the 2030 scenarios to their 2019-2022 counterparts with similar electricity pricing trends, such as WF4,19 and WF4,30-reg, it is clear that more electricity is exported and less imported due to lower hydrogen market prices in the 2030 scenarios.The exported electricity in WF4,2030-reg accounts for 8% of the total wind production, compared to 3% in WF4,2019.
Fig. 6 presents the hourly fluctuations in power flows, as described by Eqs. ( 7)-( 9), observed during optimal system operation on March 12 and 13, 2019, under varying wind and electricity price conditions.
On March 12, favorable wind conditions resulted in wind production of 732 MWh from wind farm 4, with electricity prices maintaining a modest average of €12/MWh.Under these conditions, the PEMEC operated continuously at full capacity to maximize hydrogen production.During the first 10 h, the wind power was insufficient to meet the PEMEC's full-load electricity demands, necessitating the procurement of additional electricity from the grid.Any excess wind energy was sold to the electricity market in the subsequent hours.
The scenario on March 13 was markedly different due to reduced wind speeds, resulting in a total wind energy production of 469 MWh.
Furthermore, electricity prices rose to an average of €62/MWh, requiring a more adaptive operation of the PEMEC.For the first 8 hours, lower electricity prices made hydrogen production economically viable, utilizing all generated wind power supplemented by grid electricity.
However, from hours 9 to 14, as electricity prices soared above €100/ MWh, the electrolyzer was shut down, and all wind energy was directed to the electricity market.The PEMEC resumed operations later in the day when prices decreased, drawing power from both wind energy and the grid.Notably, during hours 18 and 19, the electricity cost was neither low enough to warrant imports from the grid nor high enough to profitably sell the wind energy, prompting the exclusive use of wind power for hydrogen production.This strategic decision was influenced by the PEMEC's higher efficiency at approximately 30% of its full-load, as illustrated in Fig. 3.
Figs. 7 and 8 provide a detailed breakdown of various revenue streams, as calculated by Eq. ( 4), and the system's variable operational costs, as indicated by Eq. ( 5), across different scenarios.The data reveal nuanced financial dynamics shaped by fluctuating electricity prices and operational adjustments.
In the years 2019, 2020, 2021, and the 2030 scenarios with low and regular electricity prices, hydrogen sales constituted the primary revenue source, while earnings from the day-ahead electricity market were relatively minor.This pattern was particularly notable in 2019 and 2020, making electricity consumption the primary variable cost during these years.In contrast, the 2022 and 2030-High scenarios witnessed a significant shift, with increased electricity exports generating substantial revenue from the electricity market, surpassing revenues from hydrogen sales.Fig. 5 underscores this trend by highlighting reduced hydrogen production in these scenarios, which indicates lower electricity consumption for hydrogen production and consequently lower operational costs, as detailed in Fig. 8.The reduction in hydrogen sales revenue in the 2030 scenarios is attributed to the decreased market price for hydrogen compared to previous years.

Economic analysis
Building on the insights from earlier sections, this part delves into the crucial aspects of investment and fixed O&M costs, which are vital for any feasibility study.Fig. 9 offers a breakdown of both investment and fixed O&M costs for different units within the system, providing a clear view of the financial implications associated with maintaining and operating the integrated wind-hydrogen system.
Wind turbines and the PEMEC primarily account for the bulk of the investment costs.By 2030, PEMEC is anticipated to achieve a significant reduction in both CAPEX and fixed costs by 41%.Conversely, wind farms are projected to see a more modest reduction of 17%.
To provide a clear understanding of the system's financial feasibility, Fig. 10 illustrates key economic indicators, namely NPV and LCOH.The left side of the figure displays the integrated wind-hydrogen system's NPV, while the right side presents the NPV assuming no hydrogen production, a scenario where the electrolyzer is absent and all wind energy is directly sold to the electricity market.This comparison effectively demonstrates the economic implications of integrating hydrogen production versus exclusively selling electricity.The hydrogen production system consistently outperforms the alternative in all scenarios, generating higher NPV values.However, its resilience becomes particularly evident during economically challenging times, such as in 2020 when electricity prices plummeted.In contrast, wind producers who only sell their output in the day-ahead market would experience financial losses, as evidenced by a negative NPV in 2020.Even during periods of regular electricity pricing, such as in 2019 and as projected for 2030-Reg, strategies without hydrogen production are less financially viable, producing lower NPV values.Overall, the data strongly supports the integration of hydrogen production into wind farm operations as a financially viable strategy that can help secure revenue against the volatility of electricity market prices.
In 2022, the hydrogen production system not only achieved the Fig. 7. Revenue breakdown by scenarios.
highest NPV but also recorded negative LCOH values, suggesting that the system could profit from hydrogen production even before considering its sale, due to potential revenue from selling excess electricity back to the grid.LCOH ranged from €2.5 to €3.5/kgH 2 from 2019 to 2021.However, by 2030, LCOH is expected to decline significantly; for instance, in the WF4,2030-Reg and WF4,2030-Low scenarios, it will diminish by 31% and 29% respectively, compared to WF4,2019 and WF4,2020.This reduction is largely due to a more favorable investment landscape in 2030, characterized by lower unit costs as illustrated in Fig. 9. Despite this reduction in LCOH, the system's NPV is expected to decline, mainly due to the anticipated decrease in hydrogen market prices during that period.Wind farm 4, with its optimal wind conditions, consistently outperforms its peers, achieving the highest NPV among the locations analyzed.Table 7 presents a comparison of IRR and the payback period across various scenarios, distinguishing between cases with and without hydrogen production.
In scenarios characterized by low and regular electricity prices, the merits of the hydrogen production system are pronounced, as it offers a significantly reduced payback period, which is shorter by 5-20 years, in contrast to the stand-alone wind farm.However, during periods of elevated electricity prices, the stand-alone wind farm exhibits a slightly   expedited payback period, shorter by a maximum of 1-2 years, compared to the hydrogen production system.Yet it's crucial to note that the stand-alone wind farm achieves a lower NPV, as illustrated in Fig. 10.This underscores that while the stand-alone wind farm may recover its investment faster during high-price scenarios, its overall profitability, as indicated by the NPV, lags behind the hydrogenintegrated system.

Sensitivity analysis
The sensitivity analysis showcases the system's resilience and adaptability to key parameter changes, specifically conducted for wind farm 4 in the 2030 scenarios.Fig. 11 emphasizes the relationship between fluctuating hydrogen market prices and the system's economic viability.
The anticipated decline in hydrogen market prices by 2030, from €4/ kgH 2 to a projected €3/kgH 2 , introduces a new dimension to the economic dynamics of wind-electrolytic hydrogen production systems.The data show that as hydrogen prices decrease, both NPV and IRR decline, especially under conditions of high electricity prices.This indicates that while the system remains profitable, the profit margins could be compressed in a lower hydrogen price environment.However, even with this anticipated decline, the system remains profitable at the lowest hydrogen price in all electricity price scenarios, suggesting that it can still yield a net gain over its operational lifetime.Although the IRR is reduced, it remains positive, indicating that the system can continue to provide a positive annual return on investment.Fig. 12 illustrates the sensitivity analysis of varying PEMEC sizes under three different electricity price assumptions.
Fig. 12 reveals that LCOH sees a slight rise with increased PEMEC sizes during average and high electricity price periods.Yet, this increase becomes more pronounced when electricity prices are low.NPV, on the other hand, peaks for PEMEC sizes between 32 and 40 MW across all price scenarios, dropping off for sizes larger than this range.Smaller  electrolyzers (around 24 MW) appear to be the most profitable under high electricity prices.However, for low and average electricity prices, an electrolyzer size of around 32-40 MW is optimal, indicating a balance between equipment scale and energy costs.Regardless of electricity price scenarios, smaller electrolyzers consistently offer a higher IRR.This suggests that while larger electrolyzers may be appealing due to their scale, they may not provide superior annual returns.Overall, the sensitivity analysis suggests that smaller electrolyzers are more financially viable under high electricity prices, while larger electrolyzers are more financially viable under low electricity prices.It is important to note that electrolyzer sizes that exceed the wind farm's capacity may introduce inefficiencies.If the capacity of an electrolyzer exceeds the wind farm's output, it risks underutilization, which could compromise the system's economic balance.This analysis suggests that diminishing returns occur as the electrolyzer size extends beyond the wind farm's capacity.The integrated system is evaluated at different electricity prices in different years.To evaluate how electricity prices affect the economic viability of the system, a sensitivity analysis of the system to varying electricity prices is depicted in Fig. 13.Fig. 13 reveals that for electricity prices higher than approximately €60-70/MWh, the behavior of the system changes due to a strategic shift in operations.When electricity prices exceed this threshold, it becomes more economically viable to sell wind power directly to the electricity market.In contrast, for prices below this threshold, the system finds it beneficial to produce hydrogen using both wind power and imported grid electricity.This is illustrated in the top-left figure, which depicts the amount of wind power sold to the market and used to produce hydrogen, as also illustrated in Fig. 6 during the hourly optimal operation of the system.LCOH initially rises with increasing electricity prices, reaches a peak, and then experiences a sharp decline.The non-linear trend is a direct reflection of the system's adaptive operational strategy.NPV, IRR, and payback period diminish as electricity prices approach the €60/ MWh boundary.After the threshold, there is a subtle reversal in the declining trends.In essence, the economic performance of the windelectrolytic hydrogen production system is deeply linked with electricity market dynamics.The €60-70/MWh price range acts as a critical point, determining whether the system prioritizes hydrogen production or electricity sales.Stakeholders must be acutely aware of this interplay to make informed decisions, ensuring profitability while navigating the complexities of fluctuating electricity prices.

Discussion
The economic analysis and operational strategies outlined in this study, though situated within the Finnish context, carry implications for the broader application of wind-hydrogen integration systems.The adaptability of these systems to market fluctuations, as demonstrated, offers insights for regions contemplating or currently managing wind projects under market-based conditions, without subsidies.This relates especially to areas with comparable wind resource potentials and electricity market dynamics, suggesting global relevance for the findings.The implications for industry and planning are considerable.For industry stakeholders, the findings indicate the potential for diversifying revenue streams through hydrogen production, thereby enhancing financial resilience against electricity price fluctuations.From a planning perspective, the integration of wind energy with hydrogen production supports the strategic alignment with decarbonization goals and energy security enhancements, applicable not only within Finland but also in other regions pursuing similar objectives.
The exploration of future scenarios within this research offers a forward-looking lens through which the progress of wind-hydrogen integration can be evaluated, and anticipated market conditions can be analyzed.These projections not only aid in understanding the evolving dynamics of renewable energy markets but also in preparing stakeholders for strategic decision-making in an uncertain future.The inclusion of potential technological advancements and market shifts up to 2030 provides a roadmap for the sustainable expansion of windhydrogen systems globally, ensuring their economic and operational resilience in the face of future energy landscapes.

Limitations and future studies
In discussing the economic viability and cost estimations, it is crucial to recognize the variability and location-specific nature of both CAPEX and OPEX costs.Despite grounding our cost assumptions in a review of recent publications and data from ongoing projects, we acknowledge that these assumptions may not universally apply across all geographical contexts.This variability underscores a limitation within the current study, emphasizing the need for project-specific assessments to fully understand the economic implications.The study's approach to electrolyzer cost assumptions was adopted for simplicity and model clarity.However, this method may not fully reflect the cost dynamics due to technological advancements and learning.Therefore, future research should explore the impacts of technological learning curves on electrolyzer costs to provide more refined economic projections.Another source of error arises from the deterministic assumptions made regarding electricity and hydrogen market prices.Moreover, the study's reliance on historical weather data to project future wind farm outputs may overlook potential effects of climate change on wind patterns, potentially affecting the operational efficiency and economic viability of wind-hydrogen systems.However, to mitigate these limitations and provide a more nuanced understanding of potential variability and volatility, this study evaluates different years with distinctive energy market conditions from 2019 to 2022, across four different locations in Finland.Furthermore, a sensitivity analysis has been conducted to account for variations in electricity and hydrogen market prices.This approach underscores our effort to encompass a broad spectrum of scenarios, thereby enhancing the robustness of our findings and providing a diversified perspective on the economic viability and operational efficiency of wind-hydrogen systems under varying market and environmental conditions.
Future research endeavors should incorporate dynamic models capable of adjusting to real-time market conditions and consider the impacts of climate change on renewable energy generation.There is also a compelling case for expanding the scope to include environmental implications of wind-hydrogen systems through life cycle assessments.This would ascertain the overall carbon footprint reduction achievable through integration.Additionally, investigating alternative water sources for electrolysis, particularly in regions facing freshwater scarcity, could enhance the applicability and sustainability of wind-hydrogen systems.By addressing these areas, future studies can not only refine economic projections but also significantly contribute to the sustainability and adaptability of wind-hydrogen integration across diverse contexts.

Conclusions
In response to the challenges posed by volatile day-ahead electricity prices in Finland and the increasing role of green hydrogen in decarbonizing various sectors, this study explores the economic viability of integrating hydrogen production with wind farms.Amid fluctuations ranging from extremely low to exceptionally high day-ahead electricity prices in recent years, our analysis evaluates the system's optimal operation and economic feasibility under various market conditions from 2019 to 2022, while projecting into 2030 with potential technological advancements and market shifts.Focusing on four hypothetical wind farms in Western Finland, the area with the highest wind power generation potential in Finland, the study assessed optimal operation and economic metrics, including the internal rate of return (IRR), net present value (NPV), payback period, and levelized cost of hydrogen (LCOH).A sensitivity analysis on different parameters, including proton exchange membrane electrolyzer (PEMEC) size, hydrogen market price, and electricity prices, was also conducted to assess the sensitivity of the results to the input parameters.Our findings underscore the economic resilience and strategic advantage of combining hydrogen production with wind energy, revealing that such integration can substantially enhance profitability and reduce the LCOH to as low as €0.6/kgH 2 in scenarios of high electricity prices by 2030.In regular electricity price scenarios in 2030, LCOH was achieved as €2.0/kgH 2 .The analysis indicates a distinct advantage in diversifying revenue streams, which enables the system to maintain profitability even during periods of low electricity prices, contrasting with standalone wind farms that struggle economically under similar conditions.Key insights include the critical role of site selection, with optimal locations yielding up to a 47% reduction in LCOH, and the significance of operational flexibility, particularly the part-load efficiency of PEMEC electrolyzers.The study also highlights that smaller electrolyzers (compared to the size of the wind turbine) are more profitable in high-price scenarios, while an electrolyzer size of 80-100% of the wind turbine size is optimal for lower prices.Moreover, operational shifts driven by electricity prices emphasize the need for strategic decision-making based on market dynamics.In contrast to existing literature, our study demonstrates a lower LCOH for an integrated wind-hydrogen system, benefiting from both wind energy and grid electricity.This approach not only promises future economic scalability but also shows resilience during economically challenging times, suggesting a sustainable pathway for wind farm operations in the face of evolving energy landscapes.

Fig. 2 Fig. 1 .
Fig.1.The schematic of the studied system in the electricity and hydrogen markets.

N
. Javanshir et al. the most favorable wind conditions.The right figure shows the upcoming onshore wind farm projects in Finland, which are scheduled to start operating between 2023 and 2025 [2].To accommodate variations in wind speeds across different regions, this study focuses on four hypothetical wind farms located in different parts of western Finland, labeled 1 through 4 in the right panel of Fig. 2.

Fig. 2 .
Fig.2.Map of annual wind energy resource of Finland (left)[30] and map of under construction onshore wind projects (orange dots) in Finland in 2023 and selected case study wind farms (numbers 1 to 4 in the right picture)[2].(For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4 .
Fig. 4. Annual wind production and the capacity factors of the studied wind farms in different years.

Fig. 5 .
Fig. 5. Annual electricity import/export and hydrogen production of each wind farm location in different scenarios.

Fig. 6 .
Fig.6.Hourly power flows in the optimal operation of the system.

N
.Javanshir et al.

N
.Javanshir et al.

Fig. 11 .
Fig. 11.NPV and LCOH of 2030 scenarios with varying hydrogen market price, for high, regular, and low electricity prices.

Fig. 12 .
Fig. 12. Economic metrics of 2030 scenarios with varying PEMEC size at a hydrogen market price of €3/kgH 2 , for high, regular, and low electricity prices.
Annual capital cost (€) C DA PEM Cost of electricity import from the grid (€) C OM wind Variable O&M cost of wind turbines (€)

Table 7
Economic indicators of different scenarios.