Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
491352 | Procedia Technology | 2013 | 9 Pages |
Abstract
Flow through piping components are more complex than that of straight pipe and the hydrodynamic parameters are important for design it. Artificial Neural Network (ANN) modeling is useful for the prediction when the solution from first principle equations is not tractable. Experimental data on air-water flow through U-bends are collected from our earlier published paper and ANN modeling is used for the prediction of frictional pressure drop across the U-bends using three different algorithms of Multilayer Perceptrons (MLPs), i.e., Backpropagation, Levenberg-Marquardt and Scaled Conjugate gradient having a single hidden layer.
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science (General)