Low-cost green hydrogen from renewable energy curtailment: A techno-economic analysis with a real-time AI-driven approach

The accelerating climate crisis, primarily resulting from fossil fuel consumption, compels countries to transition toward low-carbon energy systems to achieve their greenhouse gas reduction goals. In response to this challenge and in alignment with the Paris Agreement, numerous countries have implemented energy transition policies that prioritize increasing the proportion of renewable energy to attain carbon neutrality [1]. As a result, hydrogen production technologies using renewable energy without carbon emissions are receiving growing interest recently as a promising solution for decarbonization [2]. Hydrogen can be produced without carbon dioxide emissions, positioning it as an environmentally sustainable energy carrier. Its high gravimetric energy density and flexibility in storage and conversion make it suitable for supporting large-scale and long-duration energy applications [3]. Consequently, hydrogen can play an essential role not only in decarbonizing energy systems but also in establishing a stable and resilient energy supply network [4].

Since 2019, countries including as Japan, the United States, and various European nations—have introduced national hydrogen strategies to achieve net-zero emissions. These policy initiatives have accelerated the expansion of the global hydrogen economy and increased interest in large-scale hydrogen production systems capable of stable and reliable operation [5]. Despite these advancements, current hydrogen production remains predominantly dependent fossil-fuel-based. In 2023, global hydrogen output reached approximately 97 Mt, of which more than 99% was derived from natural gas and coal using conventional processes such as steam methane reforming (SMR) and coal gasification [5]. Although these methods remain cost-competitive due to mature infrastructure and low feedstock costs, they are fundamentally limited by their dependence on fossil fuels and the resulting CO₂ emissions [6]. The projected increase in global hydrogen demand highlights the urgent need for scalable, low-carbon production alternatives.

However, hydrogen production pathways present a fundamental trade-off between carbon intensity and production cost. As illustrated in Fig. 1, fossil-based methods such as SMR and autothermal reforming (ATR) remain economically competitive but are associated with high greenhouse gas emissions. In contrast, renewable-powered electrolysis achieves near-zero carbon intensity, though at considerably higher costs. To address this issue, research has increasingly focused on improving the viability of green hydrogen. Recent studies have explored various solar-driven hydrogen production methods, including concentrating solar power-based electrolysis concepts [[7], [8], [9], [10], [11], [12], [13]] and heliostat-driven solar–thermochemical cycles [[14], [15], [16]], as well as high-temperature electrolysis and techno-economic optimization frameworks hydrogen production methods [[17], [18], [19]]. These developments demonstrate growing academic and industrial efforts to overcome the economic barriers of green hydrogen while maximizing its environmental benefits. Nevertheless, the current production cost of green hydrogen remains considerably higher than that of fossil-based alternatives, typically ranging from $4.1 to $7 per kilogram—three to four times higher than by-product hydrogen priced at $1.46/kg [20]. This cost gap is primarily due to electricity prices and capital expenditures (CAPEX) associated with green hydrogen production [21]. When curtailment is used for hydrogen production, it can significantly reduce costs by replacing high-priced electricity with otherwise wasted renewable energy (RE). Additionally, this approach facilitates more efficient utilization of renewable resources and supports the increasing demand for clean hydrogen during the energy transition [22].

As renewable energy penetration increases, curtailment, resulting from transmission constraints or mismatches between supply and demand, is becoming more frequent in global power systems [37]. One promising strategy to mitigate this issue is to use curtailment to support green hydrogen production via electrolysis [38]. As shown in Table 1, curtailment has been documented in Europe, North America, and Asia, with many instances attributed to region-specific operational constraints. In addition, curtailment is influenced by system-level limitations such as weak sector coupling (e.g. limited deployment of power-to-hydrogen technologies such as water electrolysis), insufficient or slowly deployed large-scale energy storage systems (ESS), and inadequate grid flexibility and market design [[39], [40], [41]]. As renewable energy deployment continues to expand, these structural constraints are likely to intensify, leading to higher curtailment levels in the coming years. Therefore, hydrogen production using curtailment is emerging as a key approach to enhance system flexibility and support carbon-neutral energy transitions.

Table 2 shows the present study within existing literature by comparing recent contributions across four criteria: curtailment modelling, electricity price forecasting, use of weather data, and real-time hydrogen cost analysis. Shams et al. (2021) developed a machine-learning-based framework to forecast wind and solar curtailment under uncertainty and assessed the techno-economic feasibility of operating alkaline water electrolysis (AWE) and ESS using curtailment, while relying on fixed market prices rather than electricity price forecasting [56]. Biggins and Brown (2022) investigated system-level design by optimizing electrolyzer and battery capacities using wind-curtailment patterns generated via a Markov Chain model and historical wind-speed data, under a pre-determined electricity price assumption [57]. Meng et al. (2022) advanced the forecasting domain by improving day-ahead electricity price prediction using an attention-based long-short term memory (LSTM) model under high renewable penetration, but did not address curtailment modelling or real-time hydrogen cost analysis [58]. Park et al. (2023) incorporated curtailment by estimating curtailment volumes based on historical meteorological data and grid constraints, and evaluated hydrogen production performance using weather-driven renewable generation scenarios without developing a dedicated short-term curtailment forecasting model [59]. Balci and Erbay (2024) analyzed the economic feasibility of solar-based hydrogen production in Türkiye using solar potential atlas data and fixed feed-in tariffs or future cost targets, rather than real-time weather inputs and electricity price forecasting [60].