Design of complex heat exchanger geometries using AI and simulation

Complex geometries in heat exchangers can enhance turbulence and mixing but are computationally intensive to analyze. Helically corrugated double-tube heat exchangers offer improved thermal performance over smooth tubes, yet optimizing their design remains challenging. This study combines numerical simulations with artificial intelligence to explore and enhance the performance of such exchangers. Using the k-ω SST turbulence model, the effects of varying annulus Reynolds numbers (10,000–25,000), corrugation pitch ratios (1–4), and corrugation height ratios (0.2–0.5) were examined. Simulation results were used to train artificial neural networks (ANNs) to predict key performance indicators, including the relative Nusselt number (Nu_r) and the performance evaluation criterion (PEC). Genetic algorithms (GA) were then applied to optimize these ANN models. The optimized designs achieved a maximum Nu_r of 2.89 and a PEC of 1.016. Parametric analysis showed that the highest Nu_r occurs at moderate Reynolds numbers (15,000), the lowest pitch ratios (1), and the highest corrugation height ratios (0.5). Flow visualizations highlighted the importance of secondary flow structures in enhancing turbulence and disrupting thermal boundary layers. This integrated approach demonstrated an effective strategy for balancing heat transfer and pressure drop, providing a promising direction for the design and optimization of heat exchanger systems.