The team, led by Dustin Frings at the Chair for Wind Power Drives, focused on the Rhede (Ems) wind farm, an existing site with 21 turbines. Their goal was to answer a deceptively simple question: what rated power should the electrolyzer have, and what battery capacity should accompany it, to maximize the annual profit of the combined plant? The answer depends on a tangle of site-specific factors, including the distance to the planned German hydrogen grid, water access, cable routes, and the wind profile captured by the New European Wind Atlas. Unlike many earlier studies that minimized the levelized cost of hydrogen (LCOH), the Aachen group chose annual profit as their objective, reflecting what an actual operator cares about: the bottom line of the whole hybrid power plant, not just the hydrogen subsystem.

What sets the study apart is the fidelity of its electrolyzer model. Rather than assuming a constant efficiency, the researchers modeled the cell voltage through a polarization curve built from the Nernst voltage, activation overpotentials at cathode and anode, and ohmic losses dominated by the membrane resistance. Efficiency then emerges as the product of voltage efficiency and Faraday efficiency, and it varies with operating point. Crucially, they also modeled degradation as an additional voltage that must be overcome over time, drawing on experimentally validated degradation rates for five distinct operating modes: steady operation at rated power, steady part-load operation, off-state, fluctuating operation, and start-stop cycling. The model classifies any hourly load swing exceeding 3 percent of rated power as fluctuating operation, a regime known to accelerate membrane and catalyst aging.

The consequences of ignoring these effects are striking. Comparing their detailed model against a simplified approach that neglects degradation, the researchers found that the LCOH is underestimated by about 21 percent, equivalent to EUR 1.2 per kilogram of hydrogen. When part-load efficiency is also ignored, the underestimation grows to 35 percent. For a technology already struggling to compete on cost, a gap of that magnitude can flip a promising business case into an unbankable one. The team also introduced a physically grounded stack replacement criterion: instead of assuming a fixed lifetime, they tracked the rising cell voltage until it reached 2.5 volts, the literature-reported end-of-life point where cell efficiency falls to roughly 50 percent, and then charged the replacement cost, assumed to be about 40 percent of total electrolyzer investment, to the system.

The battery turned out to be more than a passive buffer. Under the design operation strategy, which prioritizes feeding as much wind power as possible to the electrolyzer, the BESS smooths out shortfalls in wind output and prevents the electrolyzer from shutting down whenever the wind dips below the minimum part-load threshold. The numbers are compelling: adding a 0.75 MWh battery cut the total annual electrolyzer degradation by more than one-fifth, largely because start-stop degradation increased by roughly 60 percent when the battery was removed. With the battery in place, annual operational profit rose by 7 percent while the LCOH held constant, because the extra hydrogen yield offset the battery’s capital cost. The battery itself, however, is not immune to wear, and the study conservatively assumed a ten-year calendar life without explicitly modeling cycle-induced degradation.

Applying the design optimization to the Rhede case produced an optimal electrolyzer rated power of 6.55 MW and a battery capacity of 0.75 MWh at a hydrogen price of EUR 7.8 per kilogram, the 2024 average on the HYDRIX exchange. That electrolyzer size sits near 10 percent of the wind farm’s rated power, consistent with the range identified in other recent studies. The hybrid configuration lifted the annual profit to EUR 3.53 million, about 10 percent above the EUR 3.21 million achievable by the stand-alone wind farm. Notably, roughly two-thirds of that profit came from electricity sales and only one-third from hydrogen, and electricity procurement dominated the hydrogen system’s total cost of ownership, accounting for about 69 percent of electrolyzer-related expenditure. The economics of green hydrogen, in other words, remain hostage to the power market.

Price assumptions proved to be the study’s most consequential sensitivity. The design phase assumed a constant achievable electricity price, a common simplification, but when the researchers compared this against actual hourly day-ahead market prices in Germany for 2024, the constant-price assumption inflated annual operational profits by 23 percent. As renewables proliferate, price cannibalization and volatility erode the value of raw electricity sales, and any design tool that ignores this systematically flatters the project. The optimal plant size also shifted dramatically with the hydrogen price: at EUR 5.0 per kilogram, the optimizer preferred a bare wind farm with no electrolyzer or battery at all, while at EUR 10 per kilogram the optimum grew to 16.58 MW of electrolysis and 4.38 MWh of storage.

To test whether their conservative, rule-based design strategy left money on the table, the team ran a mixed-integer linear (MIL) operation optimization on the finalized design, allowing the electrolyzer and battery to respond intelligently to hourly market prices. The result was reassuring for the design method: after correcting for the linearized efficiency and degradation effects that the MIL model initially neglected, the optimized operation delivered only about 2 percent more profit than the simple heuristic. The two approaches even produced similar electrolyzer utilization, with 6375 full-load hours for the heuristic versus 5892 for the optimizer. But this near-parity is fragile. When the hydrogen price dropped to EUR 5.0 per kilogram, the optimized operation cut electrolyzer use to 3157 full-load hours, strategically selling electricity during high-price hours instead, and outperformed the price-blind heuristic by more than 9 percent. The heuristic’s competitiveness, the authors caution, is specific to this site and price regime.

The broader lesson reaches beyond one German wind farm. The study demonstrates that high-fidelity physical modeling, especially of degradation and part-load behavior, is not academic luxury but a prerequisite for credible investment decisions in wind-to-hydrogen projects. It also shows that battery storage and electrolyzer sizing cannot be treated independently, since the battery’s buffering role directly modulates the electrolyzer’s aging. The researchers point toward several refinements: rolling-horizon optimization that adapts dispatch to stack aging, nonlinear models that embed degradation directly in operational decisions, and rotating or daisy-chain allocation strategies across multiple electrolyzer stacks. They also note that markets with greater price volatility or different wind regimes could widen the gap between static rules and adaptive optimization considerably. For developers contemplating the next wave of hybrid power plants, the message is clear: model the aging, respect the market, and only then commit the capital.

Subject of Research: Site-specific techno-economic optimization of a wind farm co-located with a PEM electrolyzer and battery storage, including degradation effects

Article Title: Case study of a site-specific design and operation optimization of a wind farm co-located PEM electrolyzer and BESS including degradation

Article References: Frings, D., Jacobs, G., Reichartz, T., Potthoff, T., Blickwedel, L., & Knops, M. (2026). Case study of a site-specific design and operation optimization of a wind farm co-located PEM electrolyzer and BESS including degradation. Wind Energy Science, 11(9), 3509-3529. https://doi.org/10.5194/wes-11-3509-2026

Image Credits: AI Generated

DOI: 10.5194/wes-11-3509-2026

Keywords: green hydrogen, PEM electrolyzer, wind farm, battery energy storage, degradation, levelized cost of hydrogen, hybrid power plant, optimization, electricity market, Germany, techno-economic analysis, energy transition