the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Apr 2026
Case study of a site-specific design and operation optimization of a wind farm co-located PEM electrolyzer and BESS including degradation
Abstract. The expansion of volatile renewable energy sources leads to increased electricity price volatility and cannibalization effects, intensifying the economic pressure on project developers. Decentralized hybrid energy systems for renewable hydrogen production offer a solution to exploit these price fluctuations and counteract curtailment during hours of low or negative electricity prices. However, the design and operation of those systems are inherently coupled and significantly influenced by external factors, such as electricity and hydrogen prices that can be achieved over the lifetime. To determine an economically optimal design, specifically the power of an electrolyzers and the capacity of a battery, both site-specific and plant-specific characteristics must be considered.
First, this paper presents a methodology for determining the optimal electrolyzer rated power and lithium-ion buffer battery capacity size for a 68 MW wind farm in north-western Germany. The approach extends an existing site-specific design method by introducing a battery storage system and enhancing the electrolyzer model with part-load efficiency and operating-mode-dependent degradation. Results indicate that neglecting degradation leads to an underestimation of the levelized cost of hydrogen (LCOH) by 1.2 € kg−1, corresponding 21 %, while neglecting both degradation and part-load efficiency increases this underestimation to 35 %. Concurrently, the inclusion of the battery energy storage system (BESS) can reduce electrolyzer degradation by more than one-third and increase the annual operational profit by 5 %, while it leads to a marginal LCOH increase of 1 % due to higher capital expenditures. For the design phase, a price-independent operational strategy aiming to maximizing renewable hydrogen yield was implemented, representing a necessary simplification.
This operational assumption within the design phase was validated in a second step through a mixed-integer linear (MIL) operational optimization. This assessment reveals two key findings: Firstly, the assumption of a constant archivable electricity price over the systems lifetime leads to a 23 % overestimation of annual operational profits when compared to the more realistic electricity sales at the German day-ahead market in 2024. Secondly, the operation heuristic of the design method demonstrates high economic competitiveness, deviating by only 4.5 % from the theoretical MIL optimum even under reduced hydrogen pricing. Nevertheless, this performance may be site-specific, as integrated optimization may yield significantly higher added value in markets characterized by greater price volatility or different meteorological profiles. Beyond these specific results, the model showcases the critical importance of integrating high-fidelity physical effects for electrolyzer models, alongside the strategic inclusion of battery storage. Furthermore, it demonstrates that a rigorous consideration of the operational strategy is necessary for a robust and reliable system assessment to account for volatile external factors. Overall, the proposed method provides wind farm developers with a tool to evaluate and optimize site-specific wind-hydrogen-battery systems to derive well-founded strategic investment decisions.
- Preprint
(2185 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 07 May 2026)
- RC1: 'Comment on wes-2026-63', Anonymous Referee #1, 25 Apr 2026 reply
-
RC2: 'Comment on wes-2026-63', Anonymous Referee #2, 26 Apr 2026
reply
The paper investigates the optimal operation of a wind farm co-located with BESS and an electrolysis plant, considering the efficiency and the degradation impact of the latter. The topic is well chosen, and the conclusion - that neglecting the degradation and efficiency variation leads to the profit overestimation - is both clear and compelling. My comments on the content are listed below:
1) The treatment of downstream hydrogen demand is insufficiently specified. In particular, it is unclear whether hydrogen demand is assumed to be fixed over time (e.g., annually, monthly, or daily). One would typically expect the hydrogen to be supplied to specific industrial clusters with relatively fixed consumption profiles. While the hydrogen storage tank and pipeline network may act as buffers enabling operational flexibility, the paper does not clarify whether total hydrogen demand must be fully met. Alternatively, it remains unclear whether hydrogen demand is assumed to be flexible or whether unmet demand can be supplied by other hydrogen production facilities.
2) Section 2.1: How fast do HRES (or WHP) require electrolyzers to operate? What ramp-rate or other characteristics would you expect from an electrolysis plant to follow the daily wind profile? Is PEM the only technology that fits these requirements?
3) Section 2.1.3: Eq. 10-14 are not clear. It would help the reader first to introduce the voltage and Faraday efficiencies (eq. 11 and 12), and then derive eq. 10 using the variables from 11 and 12. Equations 13 and 14 also use different variables, not connected to eq. 10-12, making it hard to link equations to each other.
4) Some of the abbreviations (like RV and MIL) are used before they are defined.
5). Section 2.2.3: You mentioned the voltage level of 2.5 V at which the stack should be replaced. What is the nominal voltage of one cell at nominal current, nominal temperature, and at the beginning of life? The comparison of the start point and the end point would be helpful to understand this value. Also, the choice of this voltage level is not clear: "the maximum cell voltage of PEM electrolyzers ranges between 1.65 and 2.5 V at
a nominal current density of 2 A cm⁻²." - It ranges depending on what and how it justifies the choice of the replacement level of 2.5 V?7) Formatting errors: reference gives message "Fehler!" and reference to figure as "Fig. Figure 7".
8) Typos: "Eighter", "Bevor", "pocc", etc
Citation: https://doi.org/10.5194/wes-2026-63-RC2 -
RC3: 'Comment on wes-2026-63', Anonymous Referee #3, 26 Apr 2026
reply
General Comment:
This manuscript addresses the design and operation optimization of wind-to-hydrogen systems including degradation. While the research topic is interesting, the authors should more clearly articulate the specific advancements and unique scientific contributions that distinguish this study from prior studies to justify its publication as a full research paper. Furthermore, the use of a local search algorithm (Nelder-Mead) in a non-convex space is not well justified, and treating degradation as a passive evaluation step rather than an active driver within the optimization loops limits the robustness and novelty of the study. Additionally, since the BESS is used specifically to mitigate electrolyzer shut-downs, the resulting impact on BESS degradation and the subsequent economic bias should be carefully addressed. Finally, although the authors identify a 21% LCOH underestimation due to degradation and a very significant 104%-110% profit gaps between the heuristic and MILP operation, the study fails to integrate these findings into the actual design loop. Relying on a sub-optimal heuristic for system sizing despite such a significant performance gap remains a major methodological concern.
Specific Comments:The abstract is overly long and functions more as a summary of specific numerical results rather than a clear statement of the paper's core contributions. While the findings are interesting, the inclusion of multiple secondary analyses distracts from the main message. I recommend refocusing the abstract on the primary research gap, the novel aspects of the proposed methodology, and the most significant high-level findings. Furthermore, it is recommended to move detailed comparisons and specific data points to the Results and Discussion sections to ensure the main contribution is clearly highlighted for the reader.
Line 65: The literature review frequently lists multiple references to support general statements without a critical synthesis. Since many of these studies are cited only once and not used for later comparison, I recommend streamlining the list or providing a more specific explanation of how they specifically relate to the current work.
Line 105: The Introduction provides a comprehensive background. However, it fails to explicitly state the main contributions and the novelty of this work compared to existing literature. While several studies are discussed, the research gap is somewhat buried in the text. I recommend that the authors clearly highlight their specific contributions, for example, by supplementing (or even replacing) the research questions with a dedicated paragraph that ensures the reader understands what distinguishes this methodology from previous works.
Line 120: The authors explicitly state that the methodology is an extension of the method by Reichartz et al. (2024b), with the primary contributions being a more detailed electrolyzer model and a shift in the objective function from LCOH to Annual Profit (AP). As currently presented, the work appears to be an incremental contribution that lacks the fundamental novelty required for a full research paper. Therefore, it is recommended that the authors elaborate further on the scientific novelty and the broader applicability of their findings beyond a specific model extension.
Figure 1: It is unclear why a two-stage approach was adopted (fixed-strategy design followed by MILP evaluation) rather than integrating the MILP model directly into the design phase. The authors should justify this methodological choice, especially considering that the design heuristic already shows significant deviations from the theoretical MILP optimum.
Figure 1: The authors should clarify the functional role of the "Degradation evaluation" and "Electrolyzer stack exchange" blocks within the design loop. As presented in the flowchart, these steps appear as passive post-processing elements rather than active drivers of the optimization. Since the power flows are pre-defined by a "fixed operation strategy" before degradation is even assessed, it remains unclear how degradation dynamically influences the Nelder-Mead decision-making process for the sizing variables. The authors should address this apparent discrepancy between the "degradation-aware" objective and the sequential, non-integrated flow shown in Figure 1.
Line 145: The choice of the Nelder-Mead algorithm for the design optimization requires further justification. It is also important to address its ability to navigate a potentially non-convex design space to find a reliable optimum, given its nature as a local search method.
Line 150: The design phase relies on a fixed operation strategy, which prevents the optimizer (Nelder-Mead) from exploring active trade-offs between dynamic operation and stack longevity. The authors should justify why a heuristic was preferred over an integrated sizing-operation approach and clarify if this "fixed" logic might lead to a sub-optimal system design in the presence of degradation. Relying on a simplified fixed strategy for the design phase, while recognizing its limitations regarding the paper's core topic (degradation) raises concerns about the validity of the reported optimal configurations.
Line 160: The manuscript explicitly states that the operational MILP model "neglects the impact of degradation on the electrolyzers efficiency", creating a discrepancy with the design model. To ensure a fair assessment of the reported deviation, the authors should clarify how this omission biases the comparison. It is recommended to either align the MILP model assumptions with the design model or provide a thorough discussion on the limitations of using an "idealized" benchmark to validate a degradation-aware study.
Line 270: The authors state that the BESS is operated specifically to minimize electrolyzer shut-downs and mitigate its degradation. However, BESS cycling significantly impacts battery lifetime and, consequently, the project's CAPEX and LCOH. It is recommended to explain how this issue is addressed, or otherwise provide a clear justification for this omission and acknowledge it as a limitation.
Lines 364-367: The manuscript explicitly states that the optimization 'neglects the impact of degradation on the electrolyzers efficiency' to avoid non-linearities. This represents a significant methodological gap: if the optimizer is blind to efficiency losses, the operational decisions cannot be considered truly optimal regarding the stack's state of health. Given the paper's focus on detailed degradation modeling, this lack of integration should be addressed.
Lines 505-510: The authors state that identifying a global optimum is "considerably more challenging" due to the non-convex nature of the problem. However, this is a weak justification in the context of current research. Established methodologies, such as convex optimization or metaheuristic global search algorithms, are well-known for addressing these complexities. It is recommended to improve the justification of relying on a local search method like Nelder-Mead, without even a multi-start validation or a discussion on why a convex reformulation was not explored.
Lines 555-565: The finding that neglecting degradation leads to a 21% underestimation of the LCOH is highly significant and underscores the importance of the topic. However, there is a fundamental disconnect in the study's contribution: while these impacts are quantified, they are not actively integrated into the optimization process. Both the design and MILP stages remain "blind" to degradation effects during the decision-making loops. Consequently, the manuscript succeeds in quantifying the error of ignoring degradation but falls short of providing a methodology that actually optimizes life-cycle performance by accounting for these dynamics within the optimization itself.
Lines 685-705: The authors report 104%-110% profit differences when switching from the heuristic strategy to MILP optimization. Despite this, the system sizing remains based on the heuristic approach. This demonstrates that the fixed operation strategy is unable to capture the real conditions of these systems, including variable market prices and degradation, leading to a design optimized for a sub-optimal operational logic. The authors should clarify how a design based on a "blind" heuristic can be considered robust given these identified performance gaps.
Technical corrections:
125 Typo in "Bevor starting"
Line 148: Broken cross-reference ('Fehler!...'). A thorough proofreading of the full manuscript is recommended.
Citation: https://doi.org/10.5194/wes-2026-63-RC3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 73 | 15 | 10 | 98 | 0 | 2 |
- HTML: 73
- PDF: 15
- XML: 10
- Total: 98
- BibTeX: 0
- EndNote: 2
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper studies the techno-economic design and operation of a wind-farm-coupled PEM electrolyzer with BESS, considering part-load efficiency and operation-dependent electrolyzer degradation. The topic is timely and relevant. However, the manuscript requires substantial revision before publication. My comments are as follows:
1. The presentation should be improved. Several abbreviations are used before being clearly defined. For example, “RV” and “AP” should be introduced with their full names when they first appear.
2. There is a cross-reference error on page 6, lines 148–149. The text contains the unresolved reference: “Fehler! Verweisquelle konnte nicht gefunden werden. (Lagarias et al. 1998).” This should be corrected.
3. The terminology should be made consistent. The manuscript uses both “HRES” and “HWF” to describe the co-located wind farm, electrolyzer, and battery system. This may confuse readers. In addition, this type of generation system is commonly referred to as a hybrid power plant ( HPP). The authors are encouraged to check the definition of HPP used by IEA TCP Wind Task 50 and align the terminology accordingly.
4. The operational strategy in Figure 2 should be clarified. The strategy appears to be rule-based, but the current figure is difficult to follow. The authors are encouraged to present the operational logic as a flowchart, which would make the decision process for power allocation among the wind farm, electrolyzer, BESS, and grid more transparent.
5. Battery degradation should be considered or discussed in more depth. The BESS is central to the conclusions: it reduces electrolyzer degradation and increases annual profit, while only slightly increasing LCOH. However, BESS degradation and replacement are neglected, although the manuscript acknowledges that cycling degradation could affect the optimal BESS capacity. The authors should either include battery degradation in the model or provide a more detailed discussion of how battery degradation may influence the main conclusions.
6. The design problem is nonlinear and likely non-convex, and the paper uses the Nelder–Mead algorithm after an exploratory design-space sweep. The authors should provide more information on the optimization bounds, initialization, stopping criteria, convergence tolerance, number of function evaluations, and whether multiple starting points were tested. In addition, terms such as “global optimum” should be avoided unless global optimality can be justified.
7. The robustness of the case-study conclusions should be strengthened. The case study is based on one specific wind farm, one wind time series, German 2024 day-ahead electricity prices, and a fixed hydrogen price. However, the optimal electrolyzer and BESS sizing, as well as the performance gap between the rule-based heuristic and the MIL dispatch optimization, are likely sensitive to wind-resource profiles, electricity price volatility, and hydrogen price assumptions. To strengthen the conclusions, the authors are encouraged to include sensitivity analyses using at least one additional wind year, electricity price year, or hydrogen price scenario.
8. Correct spelling and wording issues, for example: “Bevor” → “Before,” “eighter” → “either,” “prize cannibalization” → “price cannibalization,” “European Comission” → “European Commission,” “archivable electricity price” → probably “achievable electricity price,” “reasonably solution” → “reasonable solution,” and “the integrate of adaptive…” → “the integration of adaptive…”, etc.