An open-source moving-boundary approach for shell-and-tube heat exchanger sizing optimization

Conventional single-zone heat exchanger models fail to resolve local temperature gradients, property variations, and phase transitions, while high-fidelity distributed models are often too computationally demanding to be embedded in integrated design optimization. To bridge this gap, this work presents a shell-and-tube heat exchanger sizing framework that couples a novel tube-pass-aware one-dimensional moving-boundary model with a particle swarm optimization algorithm. The modeling framework includes a user-defined discretization level, allowing a tunable balance between accuracy and computational cost. The optimization objective is the minimization of total heat exchanger mass, thereby reducing thermal inertia while lowering material use, handling requirements, and overall cost. Comparative validation against published reference cases under single-phase and two-phase operating conditions demonstrates heat exchanger mass reductions of 22 to 24%, while increasing modeling fidelity. The predictive accuracy was comparatively validated with the reference studies, with heat transfer deviations of approximately 1% and pressure-drop deviations below 10% for low discretization modeling. Achieving these improvements within a reasonable computational time, the optimization results show that the factors most strongly affecting heat exchanger mass are, in order of importance, the tube-thickness assumptions (−28 to −46%) as tubes represent 60 to 80% of the total mass, the discretization level (−10 to ＋56%), and the choice of objective function (−10%).