On the pillow-plate heat exchanger with wavy walls and filled with a hybrid nanofluid: Numerical simulation and data-assisted prediction

Pillow-plate heat exchangers have garnered significant attention due to their enhanced heat transfer capabilities compared to traditional flat-plate designs. Considering previous studies, the effects of incorporating wavy plates and hybrid nanofluids have not been widely investigated. This study aims to address this gap by numerically simulating the flow and heat transfer characteristics. After validation by the experiments, the numerical investigation examines the influence of varying wave amplitude and wavelength as well as the Reynolds number and the volume fraction of the nanofluid on the hydrodynamic and thermal field. Furthermore, the irreversibility analysis due to heat transfer and friction is also performed. An artificial neural network model is applied to provide a data-assisted prediction of the heat exchanger performance, reducing the need for costly simulations. Findings: The results show that using wavy plates and the hybrid nanofluid boosted the Nusselt number in the range of 12–42 %, while it has a modest impact on the pressure drop by a value of almost 10 %. The proposed geometries demonstrate consistent thermo-hydrodynamic advantages, with performance evaluation criteria (PEC) values exceeding unity. Decreasing the wavelength and increasing the wave amplitude can further improve the PEC, with a maximum value of nearly 1.8.