Use this URL to cite or link to this record in EThOS:
Title: Development of a virtual permeameter
Author: Islam, Mishal
ISNI:       0000 0004 2718 9448
Awarding Body: University of Leeds
Current Institution: University of Leeds
Date of Award: 2010
Availability of Full Text:
Access from EThOS:
The thesis concerns the development of a virtual permeameter that will enable the structure-flow relationships for bulk porous media to be assessed using numerical methods. Ultimately, the intention is to link a digital packing algorithm with lattice Boltzmann modelling (LBM) to predict the permeability of porous media and filtration performance using simulated microstructures, with potential applications to filter design and the oil industry applications. In order to prove the approach being pursued, the micro-structural details of porous media samples are gathered using X-ray tomography (XMT), rather than from packing algorithms, using small samples of such media to permit the subsequent prediction of their bulk permeability and related properties. The thesis presents a systematic sensitivity analysis of using LBM for predicting the permeability of packed beds. Individual packed beds have been created using spherical glass beads with a mean diameter of 116.4 urn, sand particles with a mean diameter of 223 urn, and polymorphous particles with a diameter between 115 to 375 IJm with XMT used to image samples of the beds and to permit the reconstruction of a three-dimensional image of the sample for use as the basis of the LBM simulations. LBM predictions are also compared to permeability measurements obtained from a packed bed formed by passing a low concentration saline solution through the various ) particles in a filtration rig. The saline solution allows the packed bed, once dried or encased in wax, to form a more solid structure from which samples can be more easily removed for analysis. In order to assess the possible effects of uncertainty in the accuracy of representation of the packing on the simulation results, sensitivity studies and XMT were employed to investigate the degree of representation of the sample required to reduce digitisation errors of the packed bed to a minimum. It concluded that any section of the sample scanned using XMT, as long as the sample is homogenous throughout, will give reproducible simulation results, irrespective of the section from which it is cut, with an average accuracy of 18% between simulation and experimental data for glass beads and 22% for sand particles. Polymorphous particles required averaging of several layers of the sample giving an acceptable error between simulation and experimental of 42%. Also, it is demonstrated that, regardless of wall effects, a digitised sphere must be greater than 16 voxels in diameter for relatively accurate results; furthermore, it is found that there is less than a 10% error between predictions and measurements if a 32 voxel diameter is used. The work also provides recommendations for the minimum dimensions necessary in sampling packed beds to allow the accurate simulation of filter performance using LBM. Lastly, the work 3 indicates a deviation from experimental results inherent in using the Carman-Kozeny formula frequently used for predicting the permeability of a packed bed, with the present LBM approach giving superior predictions. Overall, the work demonstrates that, with sufficient care, it is possible to use small representative samples of porous media for use as input to LBM to permit the accurate prediction of its permeability and filtration performance under laminar conditions regardless of particle geometry. The highest pressure used was 1.5 bar through the porous media giving a Reynold's number (Re) of under Re=2. Higher pressures where laminar flow tends towards transitional or turbulent flow have not been covered. The thesis also explores the accuracy of representing the micro-structure of such media using a digital packing algorithm, rather than obtaining such information from XMT and concludes that there is a noticeable reduction in error when using a packing algorithm compared to XMT scanned beds. 4.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available