Application of an artificial neural network to predict the entrance length of three-dimensional magnetohydrodynamics channel flow
Department of Mechanical Engineering, Sari Branch, Islamic Azad University, Sari, Iran
* e-mail: email@example.com
Accepted: 16 May 2019
Published online: 27 September 2019
In this paper, a feed-forward, back-propagation neural network was employed for modeling the three-dimension magnetohydrodynamics (MHD) developing fluid flow. The aim of the study is to obtain a correlation for calculating the entrance length by applying an artificial neural network (ANN). To collect the data for training ANN, the numerical finite volume method (FVM) was conducted. The data were collected including Reynolds number (Re) ranging from 500 to 1000 and Hartmann number (Ha) ranging from 4 to 14 for a three-dimensional rectangular channel with four different aspect ratios (AR in a three-layer ANN for modeling the flow and computing the entrance length as a function of AR , Re and Ha. The results obtained from the ANN, FVM and the proposed correlations were compared and it was observed that the variation of the entrance length was different for each AR . Thus, two correlations were proposed with the different range of the AR and Ha. The contours and vectors of the velocity along the channel direction and for different cross-sections were illustrated. In addition, the effect of AR and Ha on the entrance length and pressure loss was presented.
© Società Italiana di Fisica / Springer-Verlag GmbH Germany, part of Springer Nature, 2019