PERFORMANCE PREDICTIONS OF AIR-COOLED CONDENSERS HAVING CIRCULAR AND ELLIPTIC CROSS-SECTIONS WITH ARTIFICIAL NEURAL NETWORKS

Abstract

In this study, mathematical models of air cooled condensers with circular and elliptic cross-sections were developed and performances were evaluated with artificial neural networks. Air velocity, orientation angle and ambient temperature were used as the input to the neural network structure while heat transfer rate to the air was used as the output. The data sets were generated from high fidelity, computationally inefficient expensive three dimensional computational fluid dynamics simulations. It was observed that artificial neural network model replaces computational fluid dynamics model and based on the mathematical model with artificial neural network, elliptic condensers perform better in terms of heat transfer compared to circular ones.