Holditch_Using Statistical and Artificial Neural Networks to predict the Permeability.pdf (1.08 MB)
Using statistical and artificial neural networks to predict the permeability of loosely packed granular materials
journal contribution
posted on 2017-01-10, 11:41 authored by Faiz M. Mahdi, Richard Holdich© 2016 Taylor & FrancisWell-known analytical equations for predicting permeability are generally reported to overestimate this important property of porous media. In this work, more robust models developed from statistical (multivariable regression) and Artificial Neural Network (ANN) methods utilised additional particle characteristics [‘fines ratio’ (x50/x10) and particle shape] that are not found in traditional analytical equations. Using data from experiments and literature, model performance analyses with average absolute error (AAE) showed error of ~40% for the analytical models (Kozeny–Carman and Happel–Brenner). This error reduces to 9% with ANN model. This work establishes superiority of the new models, using experiments and mathematical techniques.
History
School
- Aeronautical, Automotive, Chemical and Materials Engineering
Department
- Chemical Engineering
Published in
Separation Science and Technology (Philadelphia)Pages
1 - 12Citation
MAHDI, F.M. and HOLDICH, R.G., 2016. Using statistical and artificial neural networks to predict the permeability of loosely packed granular materials. Separation Science and Technology, 52(1), pp. 1-12.Publisher
© Taylor & FrancisVersion
- AM (Accepted Manuscript)
Publisher statement
This work is made available according to the conditions of the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) licence. Full details of this licence are available at: https://creativecommons.org/licenses/by-nc-nd/4.0/Acceptance date
2016-09-01Publication date
2016Notes
This is an Accepted Manuscript of an article published by Taylor & Francis in Separation Science and Technology on 20 September 2016, available online: http://www.tandfonline.com/10.1080/01496395.2016.1232735.ISSN
0149-6395eISSN
1520-5754Publisher version
Language
- en