In establishing the system boundary for this research, as illustrated in Figure 1, we focused on the operational dimensions of fuel consumption associated with machinery employed in crop farming. This encompasses a range of activities, starting from field preparation (cultivation) through to the post-harvest processes (processing), specifically for crop farms. The choice to centre on cereal crops for this study is driven by their significant contribution, accounting for more than 70% of global crop production (Fantin et al., 2017).

To undertake a meaningful study with so many variables (location, type of crop, farm structure, machinery fleet etc), four case scenarios were developed.

The categorization of farm structures in this study adheres to the EU ³ criteria for farm ownership, which is predominantly based on the size of the land. Under this classification, farms are categorized as small-holder if they utilize less than 15 hectares (ha), medium-scale if they range between 20 and 50 ha, medium-intensity if they span 50–150 ha, and intensively mechanized if they exceed 150 ha (EU, 2020). ⁴ This classification also considers the diversity in farming practices that arise due to regional climatic conditions, agricultural legislation, and mandates, even though the focus is on the same types of crops.

The scale of agricultural operations, ranging from small family-run farms to large commercial enterprises, can impact fuel consumption patterns. Smaller farms, often limited by resources such as land size, typically rely on manual or traditional farming methods. However, for this study, we assumed data on farm machinery intensity per scenario. More importantly, it is evident that larger farms with more land are likely to have a greater number of machinery pieces than smaller farms. Therefore, it is crucial to consider this assumption. To accurately reflect the differences in farm operations, the scenarios developed in this study are based on farmland distribution data from Eurostat. This allows us to capture the true demographics of the national agricultural industry in the selected regions.

As detailed in Table 1, the study includes scenarios of small-holder farms in Italy, medium-scale farms in Germany and Sweden, and large, intensively mechanized farms in the United States. The crops selected for focus in this study are wheat, maize, and barley, which are significant contributors to global arable farming. This methodological approach enables a detailed understanding of how fuel consumption varies across different types and sizes of farms, within the context of diverse agricultural practices and policies.

Under this section of the research, we conducted a detailed investigation into the fuel consumption patterns associated with a variety of crop farming practices, structured across different scenarios to capture the nuances of crop seasons, farming techniques, and regional variances. This analysis was grounded in an evaluation of the fuel footprint linked to the operation of agricultural machinery, particularly tractors, through various stages of farming activity (refer to Figure 2). Drawing upon the foundational work of Grisso et al. (2003), Šarauskis et al. (2014), and Tihanov and Ivanov, (2021), we categorized farming operations into five key areas based on their energy demands and operational phases: land preparation, planting and seedbed preparation, crop management, harvesting, and post-harvest processes.

For land preparation, including tasks such as ploughing, harrowing, and levelling, our methodology pinpointed the necessity for machinery within the 90 kW–130 kW power class, a decision informed by literature that emphasizes the significant influence of soil properties and machinery efficiency on fuel consumption (Klöckner, 2020; Paris et al., 2022; The U.S.Department of Agriculture (USDA), 2022). The subsequent phase of planting and seedbed preparation was noted for its shift towards precision, requiring machinery on the lighter end of the spectrum (<50 KW) and underscoring the differential fuel use driven by operational precision and duration.

In addressing crop management practices—fertilizing, irrigation, and pest control—our approach leveraged insights from the European Union’s Common Agricultural Policy (CAP) as a framework for integrating eco-friendly practices through precision agriculture (Balogh, 2023). This included the deployment of 2WD tractors equipped with variable-rate technology, optimizing fuel consumption by adjusting to soil-specific needs, even if with a marginal increase in fuel use due to the technology itself.

Harvesting was identified as a particularly fuel-intensive stage, with combine harvesters at the forefront of operations in large-scale farming contexts, such as those prevalent in the United States (The U.S.Department of Agriculture (USDA), 2022). Our analysis here was enriched by examining how the volume of crops harvested, field terrain, and crop moisture levels collectively impact fuel efficiency, reflecting on the complex dynamics highlighted by Klöckner (2020) and Paris et al. (2022).

Lastly, the post-harvest phase, which includes drying, storage, and transportation, was scrutinized for its fuel consumption implications. Drying and storage does not belong to the category of offroad (mobile) machinery, but hydrogen could be used for those operations and hence (positively) impact the hydrogen economy of the farm. Emphasizing the critical role of strategic planning in reducing carbon emissions during transportation—a key concern under the EU’s Green Deal—our study assumed for the utilization of heavy-duty 4WD tractors for these energy-demanding tasks.

By integrating these diverse strands of literature and analysis, our methodology offers a nuanced understanding of fuel consumption across the agricultural cycle, positioning our study as a significant contribution to the discourse on agricultural sustainability and efficiency. Through this comprehensive examination we were able to trace fuel usage in accordance with agricultural machinery and crop stages across all scenarios under our scope of investigation.

A mathematical model is developed to estimate annual fuel consumption for various crop practices (refer to Equation 1). To create an accurate model, it is essential to identify and consider the significant factors that influence fuel usage. The model incorporates variables such as land area allocated for each crop, the number of operations per crop type, fuel consumption per operation, and machinery efficiency. For this study these are the factors that were selected; the type of operation being carried out, the type of crop being planted, the size of the farm, the type of machinery used per distinct operation, the frequency and duration of operations, as well as machinery fuel consumption patterns under similar operations. By considering these variables, we can model fuel consumed by each farm under each scenario. F = A   .   ∑ i = 1 n f i   .   r i   .   M i (1)

F: Annual farm fuel consumption

The fuel estimation model presented here is a simplified version that aims to provide an estimate of fuel consumption for tractors used in farming operations. However, it is important to note that this model does not consider certain agronomic factors such as weather, soil type, and machine age/technicalities, which can have a significant impact on fuel consumption.

The data used for tractor fuel consumption per field activity was collected from a variety of sources including secondary sources, case studies, and standard field performance tests from DLG ⁵ (Deutsche Landwirtschafts-Gesellschaft – German Agricultural Society). These sources have already accounted for the agronomic characteristics while providing performance metrics during field operations in various locations. As a result, this model can be considered a linear model for fuel estimation, applicable to farms growing the same crops in different geographical locations. It is important to keep in mind that this model is not an all-encompassing solution for fuel estimation. Rather, it provides a basic estimate that can be used as a starting point for further analysis and optimization of fuel usage in farming operations.

A hydrogen refuelling station comprises several key components, including an electrolyser, compressors, hydrogen tanks, a dispenser, and a pre-cooling unit (PCU), each playing a crucial role in the fuelling process as described briefly in 5. Compressors, especially piston compressors as noted in this study, are essential for elevating hydrogen pressure from low levels in storage to the higher pressures necessary for fuelling operations, a process highlighted by Micena et al. (2020). Maintaining the purity of hydrogen during this compression is critical, as emphasized by Grüger et al. (2018), given its application as a pure fuel source in vehicles. The selected piston compressor in this study is assumed to increases the pressure from 1 MPa to a discharge pressure of 90 MPa, aligning with the electrolyser’s output and meeting the SAE J2601 standard, which sets the maximum pressure for Class 70 refuelling stations at 87.5 MPa as greatly detailed by studies (Grüger et al., 2018; Micena et al., 2020; Šimunović et al., 2022).

This research, therefore, adopts a two-tiered approach to hydrogen storage, mirroring the design suggested by Grüger et al. (2018). Initially, the hydrogen is stored at low pressure, serving as a buffer, and ensuring a consistent flow to the compressor (Grüger et al., 2018). This low-pressure tank, directly linked to the electrolyser’s output, maintains the initial pressure level (Grüger et al., 2018). Post-compression, the hydrogen is stored in a high-pressure tank, ready for use during peak demands and operating at the same 90 MPa pressure as the compressor’s output (Micena et al., 2020) (refer to Figure 3). The dispenser and pre-cooling unit (PCU) are governed by the industry-standard SAE J2601 protocol, as detailed by Janke et al. (2020) and Šimunović et al. (2022). This protocol stipulates that hydrogen’s temperature must not exceed 85°C during refuelling. Consequently, when compression increases the hydrogen’s temperature, cooling is required to prevent overheating. The system targets a pre-cooling temperature of −40°C for refuelling at 70 MPa (Micena et al., 2020). Equipped with a pre-cooling unit, hose, and nozzle, the dispensing system facilitates the transfer of cooled hydrogen into the vehicle’s tank, ensuring efficient and safe refuelling.

Our methodology involves optimizing the sizes of electrolysers and hydrogen storage tanks to meet the energy demands of agricultural operations. We employ a heuristic and iterative approach, tailoring hydrogen production to the fluctuating demand schedules across different farming cycles and geographic regions as shown in the Figures 4, 5. The goal is to balance production efficiency with storage capacity, ensuring a continuous hydrogen supply, particularly during peak agricultural periods.

Operation Cycles (OC): Assumptions for operation cycles were based on regional weather patterns and agricultural activities. In SC2-G and SC3-S, we assumed a split usage of hydrogen, 10% of the annual fuel consumption in the first + fourth quarter due to winter seasons while 30%, 60% is consumed in the second, third quarter respectively (refer to Figure 5). In addition, for SC1-I and the SC4-U we assumed that they consume hydrogen quarterly, with the consumption being distributed as 30%, 20%, 35%, and 15% across the four quarters as depicted in Figure 4. Similarly, to SC2-G and SC3-S, these countries also have 40 non-operational days per year. Based on the assumptions detailed above, Equation 2 was used to model daily hydrogen demand of these different scenarios as a first step towards hydrogen storage tank size modelling. D a i l y   d e m a n d   O C = % a g e ∗ T o t a l   H 2 T o t a l   d a y s   i n   a   y e a r − n o n   A g r i c u l t u r a l   O p e r a t i o n a l   d a y s (2)

Daily Hydrogen Demand and Component Sizing: This involved calculating the operational days for a hydrogen-powered tractor (8 h a day, 5 days a week) and determining the minimum required production capacity of the electrolyser and storage capacity of the tanks. Maximum daily hydrogen demand was computed to ascertain the minimum electrolyser production capacity and storage tank size (Equations 2–5). Max   D a i l y   H 2   D e m a n d = max   D a i l y   D e m a n d s   f o r   e a c h   o p e r a t i o n   p e r i o d (3) E l e c t r o l y z e r   S i z e   k W = Max   D a i l y   H 2   D e m a n d E l e c t r o l y s e r   D a i l y   h y d r o g e n   d e m a n d   (4) T a n k   S i z e   k g = Max   D a i l y   H 2   D e m a n d × D a y s   o f   B u f f e r (5)

The models used in this analysis integrate financial projections and economic assessments. The computational steps taken to arrive at the results are as follows:

i. When calculating energy requirements and costs, this model assesses the economic feasibility of adopting hydrogen as an alternative fuel, by computing for capital requirements, both variable and Fixed operational costs associated with producing the targeted hydrogen demanded/needed by each farm annually to fully cover its fuel needs on site using the average spot prices of electricity in each locations/scenario.

To calculate the NPV, we need to estimate the annual savings or revenue and subtract the OPEX for each year, then discount these annual net savings back to the present value using the discount rate (refer to Equation 6). N P V = − C A P E X + Σ t = 1 n   A n n u a l   S a v i n g s − O P E X 1 + r t (6) where:

• n is the number of years (15 years in this case).

Annual Savings would be the value of the hydrogen produced each year if it were to be sold or the savings from not purchasing diesel. OPEX is the annual operational cost.

The LCOH is calculated by dividing the total discounted cost (CAPEX + OPEX) by the total discounted hydrogen production over the system’s lifetime (refer to Equation 7). L C O H = Σ t = 1 n I t + M t + F t 1 + r t Σ t = 1 n   E t 1 + r t (7)

• I_t, Investment cost

The economic parameters outlined below are applied to analyse the Net Present Value (NPV) and Levelized Cost of Hydrogen (LCOH) for decentralized hydrogen production facilities on farms in Italy, Germany, Sweden, and the United States. These parameters include essential financial and technical metrics that underpin the evaluation of the profitability and viability of such systems in different national contexts.

General Parameters.

• Lifetime (n): The assumed operational lifespan of the hydrogen production system is consistently set at 15 years across all countries.

Under economic modelling of hydrogen production, a detailed assessment of capital expenditures (CAPEX) was performed using the data in Table 3. This assessment included costs associated with key components necessary for the setup and operation of hydrogen production facilities. The economic cost data used was collected from secondary literature sources including feasibility reports, and articles that have already extensively researched the implications of developing hydrogen refuelling stations in the countries understudy to reflect national market prices of these components however, some parameters were applied uniformly across all scenarios as follows.

• Electrolyser Costs: The electrolyser, a crucial component for hydrogen production, exhibited country-specific costs per kW. The costs ranged from 970 €/kW in Sweden to 1,450 €/kW in the United States, significantly impacting the initial investment required for each location.

Operational expenditures (OPEX) were analysed to understand the ongoing costs associated with the hydrogen production facilities refer to Table 4. These costs encompass electricity prices, maintenance, and operational overheads, among other factors.

• Electricity Prices (Average Spot Prices 2023): There was significant variability in electricity prices, ranging from 40 €/MWh in the United States to 126.766 €/MWh in Italy. These prices directly influence the operational costs, as electricity is a primary input for hydrogen production via electrolysis.

By standardizing these assumptions across different countries, the study ensures a consistent framework for evaluating the economic costs associated with operating these systems. This approach allows for comparability and reliable assessment of the potential financial outcomes associated with transitioning to hydrogen-based systems in agricultural settings.

Proost, (2019) underscored the critical role of annual operating hours (load hours) in determining the cost of electrolytic hydrogen. The study revealed that low operational times (<2000 h per year), even with zero electricity costs, would result in hydrogen prices exceeding 10 €/kg, a finding that is both surprising and counterintuitive. Given that the electrolyser capacities across various scenarios, it is evident that the target annual production capacities do not achieve a utilization rate of even 40% except for United States.

The European Hydrogen Roadmap anticipates that the hydrogen economy will be supported by an abundance of cheaper renewable electricity, driven primarily by the increased adoption of wind and solar energy. These renewable sources are projected to contribute more than 15% of grid electricity by 2030 and 65% by 2050 (EU Hydrogen Road Map, 2019). Janke et al. (2020) highlighted the potential for farmers to use renewable electricity from sources like wind to produce hydrogen for farm machinery, benefiting from lower electricity costs. Additionally, the authors proposed that such hydrogen systems could generate extra revenue through waste heat recovery and selling oxygen to industrial consumers, thereby enhancing the profitability of these projects users (Janke et al., 2020).

These perspectives formed the basis for creating a best-case scenario in this study, which accounts for several factors: increasing the utilization rate (UR) of each electrolyser to 70%, reducing electricity prices based on future projections of cheaper renewable energy due to increased investments, and incorporating a subsistence production model where excess hydrogen is sold to other end. This comprehensive approach aims to provide a more accurate and optimistic assessment of the economic viability of decentralized hydrogen production systems on farms (refer to Table 5).

The economic analysis of the hydrogen system was investigated based on two critical financial metrics: Net Present Value (NPV) and Levelized Cost of Hydrogen (LCOH). These metrics were evaluated under different operational scenarios to understand the economic feasibility and competitiveness of hydrogen production for machinery operations across Italy, Germany, Sweden, and the United States. The results are divided into two sets: one based on hydrogen demand for machinery operations (HDP) under assumptions that these farms produce fuel for their own needs on the farm and nothing further. However, the results of the first economic model revealed that the projects are not feasible across all scenarios refer to Table 10.

The negative NPVs across all countries indicate that, under current conditions, the hydrogen production systems are not financially viable Table 10: Results of the first economic model. The LCOH values vary significantly, with the United States achieving the lowest LCOH at 10.9 €/kg H2, reflecting its larger scale and potentially lower operational costs. Conversely, Italy has the highest LCOH at 57.4 €/kg H2, suggesting higher costs per kilogram of hydrogen produced due to smaller scale operations and higher electricity price. Thus, the LCOH where quite higher for all farms with lower volumes of production therefore confirming the trend of LCOH being directly proportional to the utility rate of the electrolyser in use due to the existence of economies of scale refer to Figure 7.

The second results of the economic model explored the impact of increasing the electrolyser utilization rate to 70% (refer to Figure 8). Additionally, a sub-scenario includes the effect of offsetting extra revenue from by-products. The results revealed that Increasing the electrolyser utilization rate to 70% significantly improves the NPV and reduces the LCOH in all countries (refer to Figures 7, 8). These improvements highlight the importance of optimizing electrolyser utilization and leveraging additional revenue streams to enhance the economic feasibility of hydrogen production systems as viable option that can increase feasibility of these systems however, issue of how these products will reach the market is still under study. The United States still maintains a relatively low LCOH at 10.8 €/kg H2, with a further increment to 14.6 €/kg if the utility rate is raised to 70% while it reduces to 12.9 €/kg H2 when considering by-product revenue offsets. Germany achieves the most substantial reduction in LCOH, from 21.8 €/kg H2 to 11.6 €/kg H2, and further down to 7.5 €/kg H2 with by-product revenue refer to Figure 8. Significant reductions are seen under the scenario of Italy whereby increasing its utility rate would reduce the LCOH from 57.4 €/kg H2 to 18.7 €/kg under the best-case scenario of Italian farms.