Modelling of an axial flow compact separator using neural network
| dc.contributor.advisor | Yeung, Hoi | |
| dc.contributor.author | Qazi, Nadeem | |
| dc.date.accessioned | 2017-04-24T14:33:10Z | |
| dc.date.available | 2017-04-24T14:33:10Z | |
| dc.date.issued | 2009-05 | |
| dc.description.abstract | A novel design axial flow cyclonic separator called I-SEP was tested with an extensive set of experiments using air-water two phase flow mixture at atmospheric pressure. These experiments provided valuable data on the separation efficiency and pressure drop under different inlet conditions. The performance parameters i.e. Gas Carry Under (GCU) and Liquid Carry Over (LCO) were found to be non-linearly related to the inlet operating conditions. However it was found that resistance on the tangential outlet of the I-SEP affects the GCU and that manipulating the pressure difference between the two outlets and the inlet of the I-SEP through manual control valves, the GCU could be controlled. The separator was also extensively tested and compared with a gravity separator, when they were placed at the exit of a riser, in severe slugging condition frequently encountered in the production pipe work from some oil fields. The tests revealed that the I-SEP has better tendency to suppress severe slugging as compared to the gravity separator. A framework for neural network based on multiple types of input was also developed to model the separation performance of the I-SEP. Mutual Information (one of the key elements of the information theory) was applied to select the appropriate candidate input variables to the neural network framework. This framework was then used to develop a neural network model based on dimensionless input parameters such as pressure coefficient. This neural network model produced satisfactory prediction on unseen experimental data. The inverse function of a trained neural network was combined with a PID controller in a closed loop to control the GCU and LCO at a given set point by predicting the manipulating variable i.e. pressure at the I-SEP outlets. This control scheme was simulated using the test data. Such controller could be used to assist the operator in maintaining and controlling the GCU or LCO at the I-SEP outlets.The work performed during this study also includes the development of a data repository system to store and query the experimental result. An internet based framework is also developed that allows remote access of the experimental data using internet or wireless mobile devices. | en_UK |
| dc.identifier.uri | http://dspace.lib.cranfield.ac.uk/handle/1826/11807 | |
| dc.language.iso | en | en_UK |
| dc.publisher | Cranfield University | en_UK |
| dc.rights | © Cranfield University, 2009. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder. | en_UK |
| dc.title | Modelling of an axial flow compact separator using neural network | en_UK |
| dc.type | Thesis or dissertation | en_UK |
| dc.type.qualificationlevel | Doctoral | en_UK |
| dc.type.qualificationname | PhD | en_UK |