Problem: degradation is not a small correction to green hydrogen TEA
The paper’s central point is simple but important: if techno-economic analysis of renewable-powered green hydrogen ignores electrolyzer degradation, it can structurally underestimate the levelized cost of hydrogen. This is especially relevant for alkaline water electrolysis. Alkaline electrolyzers are attractive for large-scale deployment because of cost and maturity, but renewable power profiles are not smooth. Wind and photovoltaic generation can force frequent start-up and shutdown events, and those on/off operations can accelerate stack degradation.
The paper therefore treats degradation as part of actual operation: renewable variability is connected to on/off cycles, on/off cycles to degradation, degradation to efficiency loss and stack replacement, and replacement to LCOH. The causal path is:
renewable variability
-> on/off operation
-> stack degradation
-> efficiency loss and shorter replacement interval
-> lower productivity and higher replacement cost
-> higher LCOH
This changes the interpretation of green-hydrogen economics. The question is not only whether renewable electricity is cheap enough. It is also whether the temporal pattern of that electricity damages the electrolyzer often enough to change the economic optimum.
Architecture: experiment-calibrated process-system TEA
The modeling flow is:
meteorological data
-> WT and PV generation models
-> hourly renewable electricity profile
-> electrolyzer on/off decision under minimum part-load
-> alkaline electrolyzer I-V and hydrogen production model
-> normal degradation plus on/off degradation
-> efficiency trajectory and stack replacement schedule
-> CAPEX, OPEX, replacement cost, hydrogen production
-> LCOH
The key modeling move is that the degradation trajectory is generated from the operation profile rather than assigned as a fixed lifetime number. Wind and solar generation are first converted into hourly power availability. If renewable power is above the minimum part-load, the electrolyzer operates; if not, it shuts down. An ESS can smooth part of this fluctuation, but it also adds cost and round-trip losses.
The alkaline electrolyzer model maps power input to current, voltage, Faraday efficiency, and hydrogen production. Degradation then modifies the efficiency trajectory over time. Once the efficiency reaches a replacement threshold, the stack is replaced and efficiency is reset.
Degradation model and economic mechanism
The paper separates degradation into normal operating degradation and on/off degradation:
In the paper’s experimental setting, the normal degradation rate is 0.00013% per hour and the on/off degradation rate is 0.00109% per operation. The normal component is based on a target lifetime assumption, while the on/off component is calibrated from a single-cell accelerated stress test with 500 cycles between 0.6 A/cm2 and 0 A/cm2 at one-minute intervals.
This is valuable because it links a materials-level degradation experiment to a system-level cost metric. The model does not merely say that lower efficiency reduces hydrogen output. It also captures the second economic channel: faster degradation shortens the replacement interval, and replacement cost can dominate the economic penalty.
The LCOH mechanism can be summarized as:
Main results: replacement cost matters more than the obvious efficiency loss
For the base design used in the paper’s experiment, the paper compares a 100 MW alkaline electrolyzer with 200 MW wind and 100 MW photovoltaic capacity. The reported cases are:
No degradation: LCOH 7.6 USD/kg, hydrogen production 1404 kg/h
Normal degradation: LCOH 8.8 USD/kg, hydrogen production 1342 kg/h
Normal + on/off degradation: LCOH 9.8 USD/kg, hydrogen production 1342 kg/h
The interesting part is that adding on/off degradation does not mainly change average hydrogen production relative to the normal-degradation case. Its larger effect is to shorten the stack replacement interval, from the supplied-note figure of about 6.5 years under normal degradation to about 4.9 years when on/off degradation is included.
That makes the economic message sharper. In renewable electrolysis TEA, degradation is not only an efficiency-loss issue. It is also a maintenance and replacement scheduling issue. A model that only updates the production denominator can miss the cost-side effect that actually drives LCOH.
Renewable mix: PV, wind, and hybrid profiles do not degrade stacks the same way
The paper’s application insight is strongest when it compares renewable profiles. PV-only systems can experience frequent daily shutdowns because solar generation follows a deterministic day-night pattern. The paper reports that PV-only operation can exceed 700 on/off operations per year. That makes PV-only designs particularly exposed to on/off degradation.
Wind-only operation is not automatically smooth, but its variability is different. Wind may continue at night and does not necessarily impose the same daily zero-generation structure. Hybrid wind-PV systems can therefore reduce low-power gaps by temporal complementarity:
PV-only:
daily zero-generation pattern
-> frequent shutdown
-> high on/off degradation exposure
WT-only:
stochastic variability
-> possible night-time generation
-> different on/off structure
Hybrid WT/PV:
temporal complementarity
-> fewer low-power gaps
-> reduced degradation cost
This is more subtle than saying “hybrid renewables have a higher capacity factor.” The degradation-aware view asks how the shape of the power profile changes electrolyzer switching frequency and stack life.
ESS and replacement threshold: design variables, not afterthoughts
The battery result is also a useful warning. ESS can reduce fluctuations and on/off cycles, but the economic benefit is not automatic. A battery helps only when the avoided degradation cost plus additional utilization benefit exceeds battery CAPEX, OPEX, and round-trip loss:
The paper indicates that larger batteries often do not improve LCOH, especially when renewable capacity is not large enough to create useful surplus energy. In large PV-heavy cases, such as the cited 700 MW PV-only setting, battery integration can become more defensible because it both increases utilization and reduces on/off cycling.
The replacement threshold creates another trade-off. Replacing the stack too early raises replacement cost. Replacing it too late keeps a degraded stack in service and reduces productivity. The paper reports an interior optimum around an efficiency threshold of roughly 0.60 for normal degradation and about 0.55-0.60 when on/off degradation is included. This is not a universal threshold for all plants. It is a scenario result that shows replacement policy is part of the design problem.
Limitations
The first limitation is that degradation is simplified into normal degradation and on/off degradation. Normal degradation is treated as proportional to operating time, and on/off degradation is treated as proportional to the number of on/off events. This is useful for system-level TEA, but real degradation can be much more complex because it depends on current, temperature, pressure, ramping, off duration, stack state, and balance-of-plant operation.
The second limitation is that the operation rule is deliberately simple. Capacity choices and replacement thresholds are explored through scenario tables rather than solved as an optimization problem. As the system becomes more complex, this table-based approach may not scale directly to realistic design and control decisions.
Takeaway
The paper is most useful because it makes electrolyzer lifetime endogenous to renewable operation. It shows that green hydrogen economics can be distorted if efficiency and lifetime are treated as fixed parameters, especially when PV-driven daily shutdowns or renewable intermittency create frequent stack stress.
Its contribution is not that degradation always makes one technology or design universally worse. The contribution is the modeling connection: renewable profile -> on/off operation -> degradation trajectory -> replacement schedule -> LCOH. That connection makes hybrid renewable design, ESS sizing, and replacement threshold part of the same economic question.
For follow-up research, the natural next step is degradation-aware optimal operation and design under uncertainty. The present paper provides a strong motivation for that direction, while also leaving open the harder questions of stack-scale degradation validity, nonlinear degradation laws, and stochastic operational control.
References
Park, J., Kang, S., Kim, S., Kim, H., Cho, H. S., Lee, C., … & Lee, J. H. (2025). The impact of degradation on the economics of green hydrogen. Renewable and Sustainable Energy Reviews, 213, 115472.