ARTIFICIAL NEURAL NETWORKS IN ESTIMATION OF HYDROCYCLONE PARAMETER D50(C) WITH UNUSUAL INPUT VARIABLES

Authors
EREN H FUNG CC WONG KW GUPTA A
Citation
H. Eren et al., ARTIFICIAL NEURAL NETWORKS IN ESTIMATION OF HYDROCYCLONE PARAMETER D50(C) WITH UNUSUAL INPUT VARIABLES, IEEE transactions on instrumentation and measurement, 46(4), 1997, pp. 908-912
Citations number
8
Categorie Soggetti
Engineering, Eletrical & Electronic","Instument & Instrumentation
Journal title
IEEE transactions on instrumentation and measurement → ACNP
ISSN journal
00189456
Volume
46
Issue
4
Year of publication
1997
Pages
908 - 912
Database
ISI
SICI code
0018-9456(1997)46:4<908:ANNIEO>2.0.ZU;2-J
Abstract
The accuracy in the estimation of hydrocyclone parameter, d50(c), can substantially be improved by application of artificial neural networks (ANN), With ANN, many nonconventional operational variables such as w ater and solid split ratios, overflow and underflow densities, apex an d spigot flowrates can easily be incorporated as the input parameters in the prediction of d50(c), The ANN yields high correlation of data, hence it can be used in automatic control and multiphase operations of hydrocyclones.