The complete process for the base case of the integrated system developed in Aspen Plus is presented in Figure 1. In summary, the system is divided into three sections: the first corresponds to the sulphur removal zone from biogas, the second to CO₂ removal, and the third to the reforming-SOFC zone. Simulation assumptions were extracted from specific literature and can be found in Table 1. In the first zone, the biogas stream is fed into a scrubber, where a sodium hydroxide solution is pumped in counter current. In this process, the gaseous H2S in the stream reacts with NaOH to form NaHS. The treated gas stream, now with low contents of H2S, exits at the top and continues through the process, while the bottom stream is sent to a bioreactor. In this aerated bioreactor, NaHS is converted into N_a2SO₄ and elemental sulphur. The mixture is then sent to a settler for the removal of solid particles, and the liquid phase is recycled back to the scrubber, reducing sodium hydroxide solution feed. This process, known as THIOPAQ, is widely used on an industrial scale [14]. In the simulation environment, stoichiometric reactors (RStoic) were used, with predefined conversion parameters seen in Table 1. The non-random two liquid (NRTL) and PR-BM are adopted for properties estimation in the unit models [15].

Figure 1.

Process flow diagram of the integrated process.

In the CO₂ removal section, the CO₂-rich stream is pressurized and fed into a washing tower, where it flows counter currently with water. The treated biomethane stream exits at the top and is directed to the reforming unit, while the effluent stream is sent to a flash vessel and subsequently to a stripping tank. In this tank, a heated steam stream is introduced to desorb the residual CO₂ from the liquid phase, which is then recirculated within the process.

In the reforming-SOFC unit, ethanol, water, and biomethane streams are preheated before being fed into the reformer. Several scenarios were evaluated, including different ethanol-to-methane compositions, as well as cases using either natural gas from the grid or biomethane. The reformer outlet stream, which is rich in hydrogen, is then directed to the SOFC module, which supplies electrical energy to the plant. After passing through the SOFC, the residual gas stream is sent to a combustion reactor, which provides heat for the plant’s energy integration. In this system, the reformer was simulated using a tubular reactor (RPlug) with implemented kinetic reactions. The SOFC module was modelled by simulating the anode and cathode separately: the anode as an equilibrium reactor (RGibbs) and the cathode as a separator block (Sep) capable of enriching the stream with O2 before being fed into the anode.

Hydrogen is a key element in the transition to a low-emission and energy-efficient economy. Among various production methods, steam reforming (SR) is the most widely used due to its high efficiency and hydrogen yield. Ethanol, derived from renewable biomass, has emerged as a promising feedstock for ethanol steam reforming (ESR) [25]. The ESR process is endothermic and constrained primarily by equilibrium rather than reaction kinetics, making in-situ hydrogen separation a potential optimization strategy [33].

ESR involves multiple competing reactions, including ethanol steam reforming, the water-gas shift reaction (WGSR), ethanol decomposition (ED), and steam methane reforming (SMR), represented in equations (1) to (4). Both ESR and SMR require high temperatures for effective hydrogen production. The process generates H₂, CO, CO₂, and CH₄. with kinetics playing an important role in the reformer operation [34], [35]. These kinetics fall into two categories: general reforming reactions (SMR and WGSR) and ethanol-specific reactions (ESR and ED). Heterogeneous Kinetic Reactor

A key challenge in steam reforming is coke formation, which deactivates catalysts over time. Since this phenomenon is highly catalyst-specific and difficult to model, it has been excluded in this study to simplify the modelling approach. As a result, the simulated outcomes may be optimistic, with actual performance depending on catalyst behaviour. In this project, the kinetic model was implemented in a simulation environment using Aspen Plus, specifically in a plug flow reactor (RPlug). The system was divided into two reaction sets due to its hybrid nature. The first set primarily corresponds to methane reforming reactions, whose kinetics are well-studied and extensively documented in the literature. The second set is related to ethanol reactions, which exhibit specific characteristics due to their more complex molecular structure and the influence of intermediate products. This approach enables a more accurate modelling process by accounting for the differences in reaction mechanisms and kinetic limitations of each fuel. For the WGSR, the reaction kinetics follow the Langmuir-Hinshelwood-Hougen-Watson (LHHW) model based on literature [36], and its corresponding parameters can be found at Table 2. The reaction rate expressions are as follows:

Where and, r_SMR and r_WGS denote the reaction rates of SMR and WGS, respectively. P H 2 , P CH 4 , P H 2 O , P CO and P C O 2 , represent the partial pressures of hydrogen, methane, water, carbon monoxide, and carbon dioxide in bar, respectively. k_SMR and K_wgs are chemical equilibrium constants of SRM and WGS; and K_co, K_H2, K_CH4, and K_H2O are the adsorption constants for carbon monoxide, hydrogen, methane, and water.

The equilibrium constants K_SMR and K_WGS can be found in literature and are typically either constant or temperature dependent. In this work, these constants are calculated based on the equations from Rahimpour et al. [37]:

To accurately model ethanol reforming, the selected reaction set was designed to capture not only the primary reaction pathways but also the formation of by-products and intermediates that significantly influence process behaviour. By distinguishing the kinetics of ethanol from those of methane, the model was tailored to account for the specific characteristics of thermal decomposition and the interactions between reactants and the catalyst — both fundamental for hydrogen production [26]. The inclusion of key reactions such as ethanol dehydrogenation (EH, (eq.10), direct ethanol decomposition (ED (eq. 11)), and acetaldehyde reforming (AR (eq.12)) enables better control over intermediates like carbon monoxide (CO), methane (CH₄), and acetaldehyde (CH₃CHO). Although the WGSR reaction is incorporated in the model, it was deactivated in the simulation environment to prevent redundancy and interference with other implemented models.

These reactions were formulated based on the law of mass action and treated as direct functions of the reactant concentrations, suitable for gas-phase systems like this one. Applying the modified Arrhenius equation to the kinetic constants allows the model to consider temperature variations accurately, as the reaction rates are highly sensitive to thermal changes. All the parameters used are shown in Table 3 and the rates follow as:

Here, k_∞j is the pre-exponential factor, k_j the kinetic constant for each reaction, E_aj the activation energy, T the system temperature, T_ref the reference temperature, and PC2H5OH,PCO,PH2O,PCH3CHO refer to the partial pressures of each component in bar.

For the development of reaction modelling within the SOFC, it was necessary to implement phenomenological models that encompass the half-reaction of hydrogen combustion. Although it is well known that reforming reactions occur at high temperatures, it is common practice to assume that only hydrogen reacts in the medium to simplify the analysis.

For computational simulation, an equilibrium reactor (RGibbs) was implemented, where only the global reaction (19) was considered. The cathode reaction corresponds to the reduction of oxygen to form the ionic species (17). Hydrogen is adsorbed at the anode, while the oxide ion crosses the electrode and reacts with hydrogen at the anode-electrode interface (18), ultimately leading to global reaction (19). These steps are represented in equations (17) to (19).

For this study, a steady-state model was developed within a computational module to analyse and assess the system's performance, considering different input conditions. To achieve this, the cell operational potential must be defined, accounting for the electrochemical, thermodynamic, and transport phenomena that govern the system's behaviour, V_cell ((eq. 20)) which is determined by the difference between the open-circuit electrochemical potential (V_Nernst, (eq. 21) and (eq. 22)) and the losses due to polarizations (V_loss, (eq. 23)). The Nernst equation calculates V_Nernst, considering the Gibbs free energy and the partial pressures of the reactant gases [9]. However, the actual cell potential is reduced by polarization losses, categorized as ohmic (V_ohm), activation (V_act), and concentration (V_Conc). These losses stem from resistance to ion and electron flow, energy barriers at the electrodes, and mass transport limitations, respectively. Understanding these losses enables optimizing system efficiency by minimizing them to enhance net power output. All the parameters required for the modelling can be found in Table 4.

This model, based on literature equations, provides a framework for predicting SOFC performance and optimizing its operation to maximize efficiency and power output.

In these equations, E⁰ represents the standard potential of the cell at a reference temperature, T_SOFC is the system temperature in Kelvin, R is the gas constant (8.314 J/(mol∙K)), n is the number of electrons transferred in the electrochemical reaction (for an SOFC, n = 2), F is Faraday's constant (96,485 C/mol), and PH2,PH2O, and PO2 represent the partial pressures of hydrogen, water vapor, and oxygen gases, respectively.

Additionally, determining the cell's power output requires consideration of several factors, including the cell area, the fuel utilization factor, and the molar flow rate of hydrogen in the feed stream. The cell power (W) and the current density (i) can be expressed as:

Where i is the current density, A is the cell area, N is the number of cells, V is the cell voltage, η_inv is the inverter efficiency, and nH2cons, is the molar flow rate of hydrogen consumed. The fuel utilization factor (U_f) is defined in equation (27) as the ratio of the consumed hydrogen nH2,cons to the available hydrogen in the feed nH2.in.

In this model, maximizing power output requires optimizing each of these variables, as they directly impact the efficiency and overall performance of the fuel cell system. To accurately account for polarization losses, the equations were categorized into three distinct types: ohmic, activation, and concentration polarizations. Each of these losses originates from different physical and electrochemical phenomena, influencing the system's voltage drop and energy conversion efficiency. For the ohmic, it follows as:

Where r_k represents the resistances associated with the anode, cathode, electrolyte, and interconnections; p_k is the specific resistivity of each material; δ_k is the material thickness; and A is the cell exchange area. A_κ and B_κ are empirical factors, and T_sofc is the system temperature.

Activation polarization arises from electrochemical reactions and can be described by the following Butler-Volmer equation:

Where β is the transfer coefficient and i₀ is the exchange current density. Theoretically, it represents the fraction of the activation polarization that influences the energy barriers of the electrochemical reaction. In fuel cell applications, a typical value for β is 0.5 [6]. The previous equation then can be simplified by:

The current exchange densities for the anode and cathode are as follows:

Where γ is the pre-exponential factor, P is the system pressure, p_i is the partial pressure of each component, and E is the activation energy.

Concentration polarization occurs when the input is consumed at the electrode surface faster than it can be supplied by diffusion, creating a concentration gradient. The total concentration polarization V_conc in SOFC is given by the equation (35) below:

Where Vconcan and Vconcca represent concentration polarizations at the anode and cathode, respectively. iL,H2iL,H2O,iL,O2 represent the limiting current densities for hydrogen, water vapor, and oxygen, respectively.

To calculate the limiting current density, it is necessary to determine both the Knudsen diffusivity and the binary diffusivity. Some of these diffusivities are obtained through empirical models based on particle collision parameters, which influence mass transport within the system. For this study, these parameters were derived from values reported in the literature [15], [17] based on Leonnard-Jones potential [35], ensuring consistency with established experimental and theoretical data. To calculate the effective diffusivity of hydrogen in a porous medium the equation follows as:

Where Deff,H2, is the effective diffusivity of hydrogen, ε is the electrode porosity, and τ is the tortuosity factor. Here, D_i,_k is the Knudsen diffusivity and D_i,j is the binary diffusivity. The equations for calculating Knudsen diffusivity D_i,k and binary diffusivity D_i,j are:

Where r_por is the pore radius, M_i and M_j are the molar masses of species i and j, Ω_Di,j is the collision integral, and σi,j2 is the collision diameter, a parameter that reflects the "width" of the molecules and affects collision frequency.

For the process configuration, multiple scenarios were implemented to investigate the impact of the methane-to-ethanol ratio on operational feasibility. Five distinct scenarios were evaluated for biogas, varying the ethanol-biomethane ratio in the reformer feed. The analysed proportions were 0% biogas (100% ethanol), 25% biogas (75% ethanol), 50% biogas (50% ethanol), 75% biogas (25% ethanol), and 100% biogas (0% ethanol). The feed stream dilution was adjusted so that the ethanol flow rate served as the reference parameter. The molar water-to-carbon ratio was maintained at 3:1 across all scenarios. The chosen baseline case was 50% biogas:50% ethanol, in which the hydrogen flow rate at the reformer outlet was 9,000 Nm³/h. For the remaining scenarios, the feed flow rate was adjusted to ensure that hydrogen production remained constant. This adjustment was performed in the simulation environment by a design spec, ensuring the expected values converged. Four natural gas scenarios were also tested replacing biogas stream feed. For these scenarios, the same adjustment strategies were applied. Since natural gas is acquired with the compositional specifications needed, the biogas pre-treatment area was omitted. Table 5 presents the flow rate specifications for each scenario, considering ethanol, biogas, and natural gas streams. To emphasize, all scenarios were simulated using the same molecule (methane). In the scenario involving biogas, pre-treatment units must be acquired, whereas in the natural gas scenarios, the gas is already purchased with ideal technical specifications.

The economic assessment was conducted in several stages. Initially, after the mass and energy balance estimation based on Aspen Plus results, the equipment was sized and quoted using the APEA methodology. Costs were estimated not only for the equipment but also for the overall infrastructure, including installation, freight, piping, electrification, administrative expenses, land acquisition, project contingency, and other associated costs, thereby consolidating the capital expenditures (CAPEX). Based on calculated CAPEX, the total investment costs (TIC) were estimated by incorporating working capital and start-up costs. Subsequently, following the principles of economic engineering outlined by Turton [13], operational expenditure (OPEX) was estimated, divided into variable costs—associated with raw materials—as well as direct and indirect field costs and additional expenses. These costs also include auxiliary process expenses necessary for system maintenance and operation, such as administrative costs, research and development, payroll for workers, product distribution and selling, among others. In addition, the degradation of the SOFC stack was assumed to be synchronized with the project investment horizon; a 10-year lifetime was therefore adopted, avoiding the need to assume stack replacement or additional CAPEX during the project lifetime. Finally, a cumulative and discounted cash flow was constructed, linked to a minimum attractiveness rate of return (MARR), allowing for the evaluation of key economic indicators, such as net present value (NPV), payback time, and electricity minimum selling price ($/MWh). Additionally, the Levelized Cost of Electricity (LCOE, $/MWh) was used to compare electricity production via SOFC with other typical back-up systems, such as conventional fuel cells, solar energy coupled with batteries, and diesel engines. Table A.1 of this work presents typical ranges for OPEX composition, with median values adopted for calculations. Furthermore, Table 5 and Table 6 provide the prices of the raw materials used and the assumptions considered in calculating the electricity selling price.

A portion of the equipment acquisition costs were estimated using APEA methodology, as previously described in this work, while specific CAPEX was applied to components, most notably the SOFC module — due to the limited information available in databases due to its technological maturity. SOFCs, being relatively new and still evolving, exhibit considerable variability in cost estimates, making them one of the main sources of uncertainty in the overall project assessment. The cost of SOFC systems generally depends on the power output, and for the purposes of this study, a specific cost of $2,000/kW was adopted, based on literature data [41]. Given the high level of uncertainty associated with this parameter, a sensitivity analysis was conducted in subsequent sections to evaluate the impact of cost variations on the final electricity generation cost.

Table 8 summarizes the breakdown of cost for two scenarios: one based on biogas and the other on natural gas. In both cases, the SOFC unit represents the highest share of total plant costs, ranging between 60% and 70%. This proportion is consistent with figures reported in recent studies for this technology [42], reinforcing the validity of the adopted estimates.

For the remaining process units, cost estimates were derived from external validated sources. The steam reformer, for instance, was assigned a specific capital cost of approximately $237/kW. The H₂S treatment and removal unit had an estimated installed cost of MM$3.347 for the base case, representing an 11% deviation from values reported in the literature [43]. This estimate was calculated using a scaling factor of 0.6 and included a monetary adjustment to align with the reported cost of $1,099.33/(Nm³/h).

Regarding the CO₂ removal system, the installed cost for the base case was estimated at MM$5.717, showing a deviation of about 12% from literature values [43], [44]. According to the references, the typical cost for this technology is around £2,000/(Nm³/h), which, when converted and scaled for the process flow rate of this study, results in approximately MM$5.104. Considering the relatively small deviations across the process units, the overall cost estimates can be deemed adequately validated for the techno-economic analysis presented.

Figure 4 illustrates the cost distribution across the evaluated scenarios. In all cases, SOFC modules represent the primary cost, accounting for around 70–80% of total plant investment—approximately MM$35. The reformer section is the second most significant contributor, representing 10–20% of costs. A noticeable trend is that increasing the biogas share in the feed leads to higher costs in pre-treatment units due to the need for larger vessels and columns. Conversely, when natural gas is predominantly used, total costs decrease. This reduction is mainly attributed to lower ethanol consumption, which in turn reduces the demand for pumps and heat exchanges related to energy integration.

Figure 4.

Equipment cost for the different scenarios and its impact on the LCOE.

The graph also correlates capital investment with LCOE. This relationship reflects not only CAPEX but also operational costs. Scenarios with higher ethanol usage exhibit elevated energy costs due to the OPEX variable costs contribution related mainly to raw material acquisition. Moreover, within the same biogas or natural gas composition, configurations with higher CAPEX also show higher LCOE values. This is explained by the extended payback time required to amortize the investment, which directly increases energy costs over the evaluated time horizon.

To estimate the Total Investment Cost (TIC), expenses are categorized into: (i) equipment costs—including spare parts, installation, and contingencies; (ii) direct field costs—piping, structural components, instrumentation, etc.; (iii) indirect field costs—civil works, services, and project management; and (iv) non-field costs—regulatory fees, logistics, contracts, and administrative expenses. After applying correction factors using APEA results and including a 10% contingency, CAPEX was estimated to range from MM$50 to MM$65 (Appendix). Including working capital and start-up costs—covering liquidity and initial testing, the TIC varied between MM$50 and MM$70 across the ten analyzed scenarios defined previously.

To evaluate the operational expenses, it is essential to estimate the plant’s variable costs, which are directly linked to the consumption of raw materials such as chemical reagents (alkalis and acids), electricity, steam, and other process utilities. These are termed variable costs because they fluctuate over time, mainly due to plant production capacity and to the volatility of commodity prices. In this study, the estimation of these costs was based on the technical coefficients detailed in Appendix C and the market prices listed in Table 6.

As illustrated in Figure 5, scenarios with lower ethanol content present significantly lower variable operational costs. This is primarily due to the high flow rate and elevated market price of ethanol, making it a major contributor to overall operating expenses. Consequently, ethanol consumption is identified as one of key factors for financial viability of the project.

Figure 5.

Operational costs breakdown for different scenarios and their corresponding impact on LCOE.

Additionally, the potential revenue from the sale of solid sulphur, a byproduct of the biogas pre-treatment stage, was also accounted for. While its contribution is relatively minor, in all analyzed scenarios, the sale of this byproduct helps to partially offset external utility costs, such as processing water and base electricity consumption.

Beyond variable costs, accurate estimation of direct operating costs is also important for a comprehensive assessment of operational expenditure. These include labor (e.g., operator wages), maintenance, and related administrative charges. In contrast, indirect operating costs are associated with broader administrative and strategic functions, including management salaries, marketing, distribution, and sales. The detailed breakdown of all OPEX components is provided in Table A.1. In addition to the cost components, long-term operational aspects related to catalyst stability should also be considered. Carbon deposition was not explicitly included in the reformer kinetic model. Although operating conditions were selected to mitigate carbon formation, catalyst deactivation may still occur over long-term operation, potentially increasing maintenance requirements and OPEX. To address this effect, the investment horizon was aligned with the expected lifetime of the SOFC stack and reformer catalyst. A 10-year horizon was therefore adopted, avoiding the need to assume additional CAPEX for mid-life replacement while capturing degradation effects within a unified economic framework

As expected, raw material expenses are the dominant component of OPEX (Table 9), accounting for up to 80% in the scenario utilizing 100% ethanol. The remaining 10–30% correspond mainly to indirect and auxiliary operational costs. Notably, scenarios with elevated ethanol usage exhibit the highest energy costs—up to 466 USD/MWh—substantially exceeding typical market prices for fossil-based alternatives.

Conversely, the most cost-effective scenarios (3 to 5 and 8 to 9), which rely on lower ethanol concentrations, achieved more competitive energy prices. A nearly linear relationship was observed between ethanol consumption and energy costs:

• A 30% reduction in ethanol usage leads to an approximate 30% reduction in energy costs.

• A 50% reduction in ethanol results in a 40–50% decrease in both operating and energy costs.

These findings underscore the strong influence of ethanol on the process’s economic performance and highlight the importance of optimizing its use and exploring more cost-effective alternatives for the plant’s energy matrix.

To evaluate the cost of electricity produced by the SOFC unit, this study adopts the Levelized Cost of Electricity (LCOE) as the primary economic indicator. LCOE represents the average cost of generating electricity over the entire lifetime of the system, accounting for capital expenditures as well as operation and maintenance costs. It is calculated by dividing the discounted total costs by the discounted total electricity generated, providing a standardized metric for comparing different power generation technologies.

Unlike other economic approaches that incorporate revenues from byproducts or additional cash-flow components, LCOE is a purely cost-based measure. It does not include potential income from coproduct sales, carbon credits, or other financial mechanisms. This makes LCOE particularly useful for isolating the intrinsic cost of electricity generation and assessing the economic competitiveness of different feedstock and configuration scenarios.

The calculated LCOE values for all scenarios are presented in Table 10, ranging from US$ 166.93/MWh to US$ 463.39/MWh. Scenarios with lower ethanol content (Scenarios 4, 5, 8, and 9) achieve the lowest levelized costs. This trend reflects the high price of ethanol, which drives up operating expenses as its share in the fuel mixture increases.

Regarding capital investment, scenarios relying on natural gas tend to be more favorable because they do not require biogas pre-treatment units, resulting in reduced upfront costs. It is also important to note that this assessment does not include carbon capture systems or credit mechanisms; if considered, biogas-based scenarios would likely exhibit improved competitiveness. Overall, the LCOE analysis highlights how feedstock cost and system configuration directly influence the economic performance of the SOFC unit, offering a consistent and technology-agnostic basis for comparing the electricity generation cost across all scenarios.

Subsequently, an additional analysis was conducted to compare the electricity costs obtained with other off-grid energy sources, including hybrid diesel-photovoltaic systems, diesel-only systems, and solar systems with battery storage (Figure 6). These systems were selected because they are solutions for intermittent off-grid power generation, like SOFC technologies. This approach provides a fairer basis for comparison, as evaluating energy from an SOFC against fossil-based on-grid energy sources would not be an equitable comparison.

Figure 6.

Energy cost comparison with other off-grid sources.

In this context, the literature indicates that the LCOE for photovoltaic cells, hybrid systems, and standalone solar systems with storage is approximately $289.00/MWh, $320.00/MWh, and $352.56/MWh, respectively [46], [47]. The analysis shows that scenarios without ethanol or with a 25% ethanol fraction in the process exhibit energy costs lower than those reported in the literature. This highlights the competitive potential of these scenarios within the current off-grid energy landscape. The results are promising and suggest that SOFC technology is a viable alternative to existing off-grid solutions.

Additionally, a second analysis was performed by varying the specific cost of SOFC technology, as shown in Figure 7. In this case, only the baseline scenario with 50% ethanol content was evaluated. The specific cost variation ranged from $500/kW to $2,000/kW. All scenarios remained below the electricity price of solar panels, with natural gas-based configurations proving more competitive than diesel-based ones. This result is also encouraging, as it underscores the importance of SOFC technology costs in determining the feasibility of the process. The insights from these two analyses also shed light on both operational costs—particularly ethanol consumption—and the capital cost of SOFC technology.

Figure 7.

Energy cost with different SOFC specific costs compared with other off-grid sources.

The technology studied is subject to a series of uncertainties and is highly dependent on factors such as production capacity, feedstock costs (particularly ethanol), capital investment, and others. To assess the sensitivity of these variables, a tornado analysis (Figure 8) was conducted by applying positive and negative fluctuations to key parameters to evaluate their impact on project feasibility and financial performance indicators.

Figure 8.

Tornado chart of different impacts on financial key parameters.

A set of key variables was selected for this analysis, including CAPEX, ethanol cost, biogas cost, production capacity, rate of return, taxes, and electricity selling price. The applied fluctuations ranged from ±25% for certain variables to ±15% for others. The base scenario (50% ethanol – 50% biogas) was chosen as the reference case, with its financial indicators as the baseline for comparison.

The main financial responses analysed included the electricity selling price, payback time, internal rate of return (IRR), and net present value (NPV). Across scenarios, a consistent pattern emerged, where the electricity price, CAPEX, and ethanol cost were the variables with the most significant influence on financial outcomes. Regarding payback time, a 25% increase in costs can render the project unfeasible within the investment horizon. Conversely, a 25% decrease in CAPEX can reduce the payback period by approximately four years. For the internal rate of return, the electricity selling price was the most impactful variable—a predictable result given its direct relationship with project revenue. Even small increases in the electricity price significantly improve the IRR. The CAPEX proved to be the most influential factor, followed by ethanol price and process capacity. This reinforces earlier findings: the technology’s cost plays a major role in economic viability. Ethanol's high price and consumption rate substantially increase operating costs. Additionally, scaling up the system demonstrates movement along the cost curve, suggesting that the current plant size has not yet reached an asymptotic cost minimum. Therefore, there is clear potential for capacity expansion to further reduce the final cost of electricity.

Finally, since CAPEX and ethanol cost were identified as the most influential variables throughout the analysis, an additional study was carried out to investigate their specific effects on the LCOE. In this assessment, the specific cost of the SOFC system was varied between $500/MWh to $5000/MWh, while the ethanol purchase price ranged from $272.72/t to $872.72/t (Figure 9).

Figure 9.

Sensitivity surface of the effect of SOFC-specific technology cost and ethanol price on the LCOE.

As expected, higher ethanol prices and higher SOFC technology costs result in increased LCOE, while reductions in these costs lead to lower LCOE values. However, a deeper analysis reveals an important asymmetry: a 25% increase in ethanol price leads to approximately a 10% rise in LCOE, whereas a 25% increase in the specific SOFC capital cost results in only a 2% increase in LCOE. This clearly indicates that operating costs, particularly ethanol, have a much stronger impact on project viability compared to capital expenditures. When the ethanol price is doubled (a 50% increase), the LCOE can rise by around 50%, while doubling the specific capital cost leads to only an 8% increase in LCOE. These findings highlight the nonlinear and non-parallel sensitivities of each variable on the system’s cost structure.

The heat map produced from this analysis further illustrates that, for every 10% increase in ethanol cost, a corresponding 2% reduction in SOFC technology cost would be required to maintain competitiveness. This is especially relevant when benchmarking against other intermittent renewable technologies such as photovoltaic systems with battery storage, which currently achieve LCOE values around $352.56/MWh. Therefore, optimizing ethanol