Towards a rational design of gravel media water treatment filters: MRI investigation of the spatial heterogeneity in pollutant particle accumulation

Minto, James Martin (2014) Towards a rational design of gravel media water treatment filters: MRI investigation of the spatial heterogeneity in pollutant particle accumulation. PhD thesis, University of Glasgow.

Abstract

Gravel filters are potentially a low cost, low maintenance water treatment solution. They require no mechanical or electrical parts and can operate without the addition of chemicals or the need for close supervision. As such, they are an appropriate technology for treating road runoff as a component of Sustainable urban Drainage Systems (SuDS) and as an initial stage of drinking water treatment in rural areas. However, the processes by which pollutant particles are removed in gravel filters are poorly understood and practical experience shows that many filters fail long before their expected design life is reached. For this reason gravel filters are little used for drinking water treatment and, when they are incorporated into SuDS, their removal efficiency and maintenance requirements are unpredictable.

The aim of this thesis was to better understand particle removal processes and the implications for gravel filter design. This was achieved through a combination of lab-based experiments and numerical modelling.

• The change in conservative tracer transport characteristics with pollutant particle accumulation was assessed through column experiments.
• The spatial heterogeneity of particle accumulation was measured by collecting 3D data with magnetic resonance imaging (MRI). Multiple scans of filters allowed the temporal evolution of particle accumulation to be assessed. A method for processing the raw MRI data to yield the change in 3D pore geometry was developed, assessed and applied.
• A simple method for extracting and comparing pore network characteristics at different stages of particle accumulation was applied to the MRI derived geometry.
• Direct modelling of the 3D MRI pore geometry with the open source software OpenFOAM allowed correlation of flow velocities with particle accumulation at each point in the pore network. Lagrangian particle tracking was used to simulate the transport of a conservative tracer through the filter.

Key findings were that spatial heterogeneity in particle accumulation was influenced by both initial pore geometry and the temporal evolution of the pore network with accumulation. This was attributed to the formation of high velocity preferential flow paths that were evident in both the 3D MRI data and the numerical model of that data. Pore networks exhibited a decrease in connectivity with accumulation and this was mirrored by a decrease in the volume of the filter that was accessible to a conservative tracer.

Conclusions of this thesis are that MRI is a useful tool for non-invasively assessing the spatial variability of clogging in gravel filters and, when combined with numerical modelling of the pore geometry, for establishing the link between pore velocity and particle removal. The formation of preferential flow paths is detrimental to the pollutant removal efficiency of a filter and could explain why many filters fail to produce good quality effluent well before their physical pollutant storage capacity is reached.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: magnetic resonance imaging, pore network, gravel filter, numerical modelling, tracer residence time distribution; flow through porous media
• Subjects: T Technology > TA Engineering (General). Civil engineering (General); T Technology > TC Hydraulic engineering. Ocean engineering; T Technology > TD Environmental technology. Sanitary engineering; T Technology > TE Highway engineering. Roads and pavements
• Colleges/Schools: College of Science and Engineering > School of Engineering > Infrastructure and Environment
• Supervisor's Name: Phoenix, Dr. Vernon, Haynes, Dr. Heather , Dorea, Dr. Caetano C. and Sloan, Prof. William
• Date of Award: 2014
• Depositing User: Mr James M. Minto
• Unique ID: glathesis:2014-5711
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 03 Nov 2014 10:43
• Last Modified: 03 Nov 2014 10:46
• URI: https://theses.gla.ac.uk/id/eprint/5711