Development of ANN prediction model for estimation of heat transfer utilizing rectangular-toothed v-cut twisted tape

Abstract

This work explores the heat transfer performance and friction characteristics of toothed v-cut twisted tapes, while employing an artificial neural network (ANN) as a predictive model. The novelty of this study lies in the innovative use of toothed v-cut twisted tapes to enhance heat transfer performance, coupled with the application of ANN for precise prediction and optimization. Focusing on a specific geometric range by adjusting the depth ratio of rectangular teeth and the width-to-depth ratio of the v-cut, the study investigates turbulent flows with Reynolds numbers spanning from 6000 to 13,000, mirroring real-world applications. The investigations unveil that the introduction of teeth to the v-cut generates a secondary vortex flow, contributing significantly to improved heat transfer by enhancing the Nusselt number (Nu) and mitigating the reduction in heat transfer rate with increasing depth of cut at higher Reynolds numbers (Re). The nuanced behavior of the friction factor is revealed, showcasing its inverse proportionality to Re and e/c, and direct proportionality to b/c, offering valuable practical insights. Remarkably, the analysis of heat transfer rate variations underscores the ANN model's predictive accuracy. Key findings include the most substantial increase in heat transfer rate for b/c = 0.67 and e/c = 0.14, with the ANN model predictions closely aligning with these results. The ANN model, trained on extensive datasets derived from experiments, emerges as a robust predictive tool, demonstrating mean relative errors constrained to less than 3.3% for Nusselt numbers and 0.08% for friction factors. Validation against previously unseen datasets further substantiates its efficacy, with an average percentage error of 3.32% for friction and 0.96% for Nusselt numbers. These results, along with the 97% and 99% accuracy for friction and Nusselt numbers, respectively, position the ANN model as a reliable tool for precision in predicting and optimizing heat transfer dynamics across varied engineering scenarios.