Techno-economic optimization of renewable hydrogen infrastructure via AI-based dynamic pricing

Abstract

Abstract This study presents a techno-economic optimization of hydrogen production using hybrid wind-solar systems across six Australian cities, highlighting Australia’s green hydrogen potential. A hybrid PV-wind-electrolyzer-hydrogen tank (PV-WT-EL-HT) system demonstrated superior performance, with Perth achieving the lowest Levelized Cost of Hydrogen (LCOH) at $0.582/kg, Net Present Cost (NPC) of $27.5k, and Levelized Cost of Electricity (LCOE) of $0.0166/kWh. Perth also showed the highest return on investment, present worth, and annual worth, making it the preferred project site. All locations maintained a 100% renewable fraction, proving the viability of fully decarbonized hydrogen production. Metaheuristic validation using nine algorithms showed the Mayfly Algorithm improved techno-economic metrics by 3–8% over HOMER Pro models. The Gray Wolf and Whale Optimization Algorithms enhanced system stability under wind-dominant conditions. Sensitivity analysis revealed that blockchain-based dynamic pricing and reinforcement learning-driven demand response yielded 8–10% cost savings under ± 15% demand variability. Nevertheless, regional disparities persist; southern cities such as Hobart and Melbourne exhibited 20–30% higher LCOH due to reduced renewable resource availability, while densely urbanized cities like Sydney presented optimization ceilings, with minimal LCOH improvements despite algorithmic refinements. Investment in advanced materials (e.g., perovskite-VAWTs) and offshore platforms targeting hydrogen export markets is essential. Perth emerged as the optimal hub, with hybrid PV/WT/B systems producing 200–250 MWh/month of electricity and 200–250 kg/month of hydrogen, supported by policy incentives. This work offers a blueprint for region-specific, AI-augmented hydrogen systems to drive Australia’s hydrogen economy toward $2.10/kg by 2030.