Title: Prediction of pressure drop in multi-size particulate pipe flow using correlation and neural network techniques
Authors: Krishnan V. Pagalthivarthi, A. Mittal, J.S. Ravichandra, S. Sanghi
Addresses: Department of Applied Mechanics, Indian Institute of Technology, Delhi, New Delhi 110016, India. ' Department of Electronics and Comp. Engg., Indian Institute of Technology, Roorkee-247667, India. ' Department of Applied Mechanics, Indian Institute of Technology, Delhi, New Delhi 110016, India. ' Department of Applied Mechanics, Indian Institute of Technology, Delhi, New Delhi 110016, India
Abstract: In this paper, we demonstrate how ANN based frameworks can be designed for prediction of pressure drop. First, Galerkin finite element method is used to simulate fully developed near-homogeneous multi-size slurry flow in horizontal pipe flow. After validating the finite element results with published data, two models (ANN) for the pressure drop are developed. The first model uses individual diameters and concentrations of each class size as inputs, and the second model uses the concentration-weighted mean diameter and standard deviation of particle size distribution as inputs. In comparison to the ANN models, the accuracy of predictions using correlations is found to be significantly inferior.
Keywords: finite element method; Galerkin FEM; multi-size particulate flow; pipe flow; pressure drop; feedforward neural networks; correlation; CFD; computational fluid dynamics; slurry flow.
Progress in Computational Fluid Dynamics, An International Journal, 2007 Vol.7 No.7, pp.414 - 426
Published online: 28 Aug 2007 *Full-text access for editors Access for subscribers Purchase this article Comment on this article