Artificial neural network model to predict compression-permeability characteristics of solid-liquid systems

• Main Authors: Iwata, Masashi, Jami, Mohammed Saedi, Shiojiri, Susumu
• Format: Article
• Language: English
• Published: The Filtration Society and the American Filtration & Separations Society 2007
• Subjects: TP155 Chemical engineering
• Online Access: http://irep.iium.edu.my/40934/3/Published.pdf; http://irep.iium.edu.my/40934/; http://www.filtsoc.org/journal/2007/volume-7-issue-4/

A statistical modeling tool called artificial neural network (ANN) is used in this work to predict the compression-permeability (C-P) characteristics of a solid-liquid system. Extensive cake properties database containing experimental data spanning various material types, particle size distribution, flocculated and unflocculated particles has been developed and used to train two ANNs. The input parameters for the first ANN were the applied pressure and the particle size distribution whereas the output parameter was the porosity of the compressed cake. In the case of the second ANN, the input parameters were the porosity and the particle size distribution. The logarithm of the specific cake resistance multiplied by the particle true density of the cake (log (s)) was chosen as the output parameter of this network. The use of porosity obtained from gravitational and centrifugal sedimentation experiments as one of the input parameters to the ANNs gave excellent results in predicting both the cake porosity and the specific cake resistance. With the help of this method, after the ANN is thoroughly trained with various slurries, the C-P characteristics of another slurry can be predicted with much less amount of the target slurry compared with conventional C-P test.