Surrogate Model for Heat Transfer Prediction in Impinging Jet Arrays using Dynamic Inlet/Outlet and Flow Rate Control

This study presents a surrogate model designed to predict the Nusselt number distribution in enclosed impinging jet arrays, where each jet functions independently and can be transformed from inlets to outlets. Although computational fluid dynamics (CFD) simulations can model heat transfer with high fidelity, their costs prohibit real-time applications such as model-based temperature control. To address this, we developed a CNN-based surrogate model that predicts the Nusselt distribution in real time, trained with data from implicit large-eddy CFD simulations with Reynolds numbers less than 2,000. Two distinct models were trained: one for a five by one array of jets (83 simulations) and another for a three by three array (100 simulations). We introduced a method to extrapolate predictions to higher Reynolds numbers (up to 10,000) using correlation-based scaling. The surrogate models achieved high accuracy, with a normalized mean average error below 2% for the five by one array and 0.6% for the three by three array on validation data, with experimental validation corroborating the predictions. This work lays the groundwork for model-based control strategies in advanced thermal management applications.