Steam methane reforming (SMR) is the most widely employed method for industrial hydrogen production owing to its 

cost-eff ectiveness. Existing studies have primarily focused on operational conditions, with relatively less attention given 

to the structural confi guration of the reformer. In this study, a computational framework integrating computational fl uid 

dynamics (CFD) modeling with Bayesian optimization (BO) is proposed to simultaneously optimize the design and operational

variables of an SMR reactor. A CFD model was developed by coupling and iteratively solving the furnace and tube 

domains to accurately simulate the heat transfer characteristics. A sensitivity analysis was conducted to identify the key 

design variables, followed by BO, to effi ciently investigate the design space. Consequently, methane (CH 4 ) conversion 

improved 3.0% based on the optimization scope. When design variables were optimized, CH₄ conversion increased by 

2.3%, while operating variables resulted in an improvement of 0.2%; the simultaneous optimization of both resulted in a 

total enhancement of 3.0%. The optimal steam-to-carbon ratio increased from 3.4 to 4.75 when design parameters such as 

tube spacing and diameter were also optimized, thereby highlighting the interdependence between reactor geometry and 

operating conditions. This study demonstrates the eff ectiveness of BO in optimizing high-fi delity CFD reactor models and 

highlights its applicability to other thermally driven systems.