Machine learning assisted modeling of thermohydraulic correlations for heat exchangers with twisted tape inserts

This article presents the application of machine learning (ML) algorithms in modeling heat transfer correlations (e.g., Nusselt number and friction factor) for a exchanger with twisted tape inserts. The experimental data at different Reynolds numbers (Re), twist ratio (t), percentage perforation (p), varied tapes (n) were used correlation modeling. Three ML algorithms: polynomial regression (PR), random forest (RF), artificial neural network (ANN) data-driven surrogate hyperparameters models carefully optimized to ensure generalizability. performance parameters R2 mean absolute error (MAE)) analyzed. It was observed that ANN predictions coefficients outperform PR RF across test datasets. Based on our analysis we make recommendations future efforts similar studies.