CFD-coupled multi-objective optimization of a solid oxide electrolysis cell stack

Author

• Erin WebbFollow

Date of Award

2026

Degree Name

Mechanical Engineering

College

College of Engineering and Computer Sciences

Type of Degree

M.S.M.E.

Document Type

Thesis

First Advisor

Dr. Asad Salem

Second Advisor

Dr. Yousef Sardahi

Third Advisor

Dr. Iyad Hijazi

Abstract

Solid oxide electrolysis cells (SOECs) can convert steam into hydrogen with lower electrical demand than lower-temperature electrolysis technologies. Their performance depends on coupled fluid flow, heat transfer, species transport, and electrochemical behavior, which makes optimization difficult because improving one response can negatively affect another. Detailed computational fluid dynamics (CFD) can capture these interactions, but the computational cost makes trial-and-error design impractical. This thesis develops a three-dimensional model of a two-cell planar SOEC stack in ANSYS Fluent and couples it directly to a customized NSGA-II in MATLAB. Each candidate design is evaluated using the CFD model rather than a surrogate. Eight variables were varied simultaneously: inlet velocity, pressure differential, inlet steam mass fraction, inlet temperature, anode reference current density, cathode reference current density, projected electrolyte area, and open circuit voltage. The two objectives were to maximize outlet hydrogen mass fraction and minimize electrical power. A population of 200 candidates over 100 generations produced nine representative non-dominated solutions. Outlet hydrogen mass fraction ranged from 0.02346 to 0.24847, while electrical power ranged from 0.93880 to 1.15877 W. The resulting Pareto front showed a nonlinear tradeoff between the two objectives. Hydrogen mass fraction increased from 0.02346 to 0.13824 while electrical power increased from 0.93880 to 0.95959 W, representing an approximately 489.3% increase in hydrogen response for about a 2.2% increase in power. At the high-hydrogen end of the front, additional increases in hydrogen mass fraction required proportionally larger increases in electrical demand, indicating diminishing returns. Across the non-dominated set, projected electrolyte area remained near the lower end of the optimization range, while anode reference current density increased from approximately 7.10×10³ to 8.73×10³ A/m². Inlet temperature, velocity, pressure differential, steam fraction, cathode reference current density, and the open circuit voltage parameter varied more broadly across the Pareto set, showing that the optimizer used combinations of thermal, transport, compositional, geometric, and electrochemical effects to improve hydrogen response. Comparison with published SOEC studies showed that the predicted outlet hydrogen mass fractions were consistent with previously reported numerical behavior. The results demonstrate the feasibility of the modeled SOEC as a hydrogen production pathway and establish a CFD-coupled optimization framework that can be extended to larger stacks, additional objectives, and future hydrogen systems.

Subject(s)

Mechanical engineering.

Chemical engineering.

Fluid mechanics.

Computational fluid dynamics.

Electrolysis -- Technology.

Hydrogen.

Renewable energy sources.

Solid oxide fuel cells.

Mathematical optimization.

Energy storage.

Recommended Citation

Webb, Erin, "CFD-coupled multi-objective optimization of a solid oxide electrolysis cell stack" (2026). Theses, Dissertations and Capstones. 2162.
https://mds.marshall.edu/etd/2162