Intersection between engineering mechanics and supervised learning for modelling clogging dynamics in depth filters

Granular depth filtration technology is widely deployed in many water-related infrastructures, which include water and wastewater treatment plants, desalination facilities, natural and engineered aquifers, with the focus objectives of providing potable water to communities and/or to pre-treat contam...

全面介紹

Saved in:
書目詳細資料
主要作者: Chew, Alvin Wei Ze
其他作者: Law Wing-Keung, Adrian
格式: Thesis-Doctor of Philosophy
語言:English
出版: Nanyang Technological University 2020
主題:
在線閱讀:https://hdl.handle.net/10356/137078
標簽: 添加標簽
沒有標簽, 成為第一個標記此記錄!
機構: Nanyang Technological University
語言: English
實物特徵
總結:Granular depth filtration technology is widely deployed in many water-related infrastructures, which include water and wastewater treatment plants, desalination facilities, natural and engineered aquifers, with the focus objectives of providing potable water to communities and/or to pre-treat contaminated feedwater prior to further treatment processes in engineering plants. Notwithstanding the dominance and conventionality of the depth filtration technology, effective quantifications of the complex clogging dynamics continue to be a hindrance for predictive maintenance actions by field engineers today. To address this problem statement, this thesis study innovates on traditional engineering techniques, specifically engineering mechanics, dimensional analysis and supervised machine learning algorithms, to build novel predictive tools to model complex clogging conditions, of physical and chemical characteristics, in operating depth filters under varying flow conditions, i.e. microscale filtration Reynolds numbers. Measured data, as extracted from previous studies and experimental runs carried out in this study, are then leveraged to calibrate the proposed engineering models to encapsulate the real-world complexity of the clogging problem inside the filters. Following which, quantitative and qualitative details are provided to demonstrate the predictive capabilities of the calibrated models for the predictive maintenance objective, which can serve as useful results for field engineers in adopting the hybrid engineering approach as proposed in this thesis study for other important environmental and chemical engineering applications. Finally, details are also provided about the author’s present research study to build engineering tools, using the same modelling methodology from that of the physical and chemical clogging conditions, to quantify the complex biological clogging behavior inside depth filters for the same target objective. Succinct details about innovating the proposed engineering approach further are also put forth.