Techno-economic uncertainty quantification and robust design optimization of a directly coupled photovoltaic-electrolyzer system. (February 2019)

• Record Type: Journal Article
• Authors: Coppitters, Diederik; Paepe, Ward De; Contino, Francesco
• Is Part Of: Energy procedia. Volume 158(2019)
• Page Start: 1750
• Page End: 1756
• Publication Date: 2019-02
• Subjects: Levelized cost of hydrogen -- Photovoltaic-electrolyzer system -- Robust design optimization -- Uncertainty quantification; Power resources -- Congresses; Power resources -- Periodicals; Power resources; Conference proceedings; Periodicals; 333.7905
• DOI: 10.1016/j.egypro.2019.01.405
• Languages: English
• ISSNs: 1876-6102

Abstract

Abstract: To solve the problem of large time shifts between renewable energy supply and user demand, power-to-H2 is a well-known option. In this framework, previous studies have shown that the direct coupling of a photovoltaic array with an electrolyzer stack is a viable solution. However, these studies assumed perfectly known operating parameters to optimize the setup. Moreover, they focused on maximizing hydrogen and minimizing the energy loss, while the cost was not addressed. We have performed an optimization including uncertainty quantification (i.e. surrogate-assisted robust design optimization) for several locations with the Levelized Cost Of Hydrogen (LCOH) as objective. This paper provides the least sensitive design to uncertainties and shows which parameters are most affecting the variability of the LCOH for that design. The robust design optimization illustrates that the mean and standard deviation of the LCOH are non-conflicting objectives for the robust designs of all considered locations. The optimal robust design is established at the considered location with the highest average yearly solar irradiance, achieving a mean LCOH of 6.6 €/kg and a standard deviation of 0.72 €/kg. The discount rate uncertainty is the main contributor to the LCOH variation. Therefore, installing a PV-electrolyzer system in locations with a high average yearly solar irradiation is favorable for both the LCOH mean and standard deviation, while de-risking the technology has the highest impact on further reducing the LCOH variation. Future works will focus on including accurate probability distributions and adding batteries to the system.