Comparison of perceptron and radial basis function neural networks in modeling heat exchangers with rectangular helical channels

Document Type : Research Article

Author

Department of Chemical Engineering, Faculty of Engineering, University of Kurdistan, Sanandaj, Iran

Abstract

In this research, computational fluid dynamics method was used to investigate the effect of geometrical parameters of rectangular spiral channels on heat transfer coefficient. Two artificial neural networks including perceptron (MLP) and radial basis function (RBF) models were used to model the heat transfer in helical channels. The model inputs included the Reynolds number and geometric parameters of the channel, and output was the Nusselt number. 135 data were generated by Computational Fluid Dynamics (CFD) simulation and after validation were used for training and evaluation of neural network models. The results of the research showed that the accuracy of MLP was slightly higher than RBF, however, both models were acceptable. Due to the high and acceptable accuracy of these two models, they can be well used in future research and applications. In this research, the main innovation is comparing two different methods for modeling the heat exchanger with a rectangular helical channel. This research shows that the use of perceptron neural network and radial basis function can both be effective in improving the performance and efficiency of the heat exchanger. This research can be used as a guide to choose the appropriate method for modeling heat exchangers and help to improve technologies related to this field.

Keywords

Main Subjects

Amirkabir Journal of Mechanical Engineering
Volume 56, Issue 5
2024
Pages 699-716

Files

Share

How to cite

Statistics