Ozonation is currently considered as one of the most effective microbial disinfection technologies due to its powerful disinfection capacity and reduction in levels of chlorinated disinfection by-products (DBP). However, ozonation of waters containing bromide can produce bromate ion above regulated levels, leading to tradeoffs between microbial and chemical risks. In efforts to meet increasingly stringent drinking water regulations and to be cost-effective, water suppliers are required to optimize ozone dosage. Therefore, there is a need to develop a robust and flexible tool to accurately describe ozone disinfection processes and contribute to their design and operation. Computational fluid dynamics (CFD) has come into use recently for evaluating disinfection systems. However, the focus of its application has been largely on modelling the hydraulic behaviour of contactors, which is only one component of system design.
The significance of this dissertation is that a fully comprehensive three dimensional (3D) multiphase CFD model has been developed to address all the major components of ozone disinfection processes: contactor hydraulics, ozone mass transfer, ozone decay, and microbial inactivation. The model was validated using full-scale experimental data, including tracer test results and ozone profiles from full-scale ozone contactors in two Canadian drinking water treatment plants (WTPs): the DesBaillets WTP in Montréal, Quebec and the Mannheim WTP in Kitchener, Ontario. Good agreement was observed between the numerical simulation and experimental data.
The CFD model was applied to investigate ozone contactor performance at the DesBaillets WTP. The CFD-predicted flow fields showed that recirculation zones and short circuiting existed in the DesBaillets contactors. The simulation results suggested that additional baffles could be installed to increase the residence time and improve disinfection efficiency. The CFD model was also used to simulate ozone contactor performance at the Mannheim Water Treatment Plant before and after installing new liquid oxygen (LOX) ozone generators and removing some diffusers from the system. The modelling results indicated that such changes led to an increase in effective residence time, and therefore an adjustment to operational parameters was required after system modification.
Another significant contribution is that, for the first time, the Eulerian and Lagrangian (or particle tracking) approaches, two commonly utilized methods for predicting microbial inactivation efficiency have been compared for the study of ozone disinfection processes. The modelling results of two hypothetical ozone reactors and a full scale contactor suggested that the effective CT values predicted by the Lagriangian approach were slightly lower than those obtained from the Eulerian approach but their differences were within 10%. Therefore, both approaches can be used to predict ozone disinfection efficiency. For the full scale contactor investigated, the tracer residence time distribution predicted by the Euerlian approach provided a better fit to the experimental results, which indicated that the Eulerian approach might be more suitable for the simulation of chemical tracer performance.
The results of this part of work provided important insight in understanding the complex performance of multiphase ozonation systems and played an important role in further improving CFD modelling approaches for full-scale ozone disinfection systems.
The third significant contribution of this work is that a CFD model was applied to illustrate the importance of ozone residual monitoring locations and suggest an improved strategy for ozone residual monitoring. For the DesBaillets ozone contactors, the CFD modelling results showed that ozone residuals in the cross section of the outlets of some contactor chambers differed by an order of magnitude. The “optimal” area of monitoring locations however varied at different operational conditions. Therefore, it was suggested that multiple ozone residual sampling points should be installed based on CFD analysis and experimental studies, to provide more accurate indicators to system operators. The CFD model was also used to study the factors affecting the residence time distribution (RTD). The results suggested that the selection of the tracer injection locations as well as tracer sampling locations might affect the RTD prediction or measurement. The CFD-predicted T10 values at different outlet locations varied by more than 10% variation. It is therefore recommended that CFD modelling be used to determine tracer test strategy before conducting a full-scale tracer test, and multiple sampling points should be employed during tracer tests, if possible.
In addition, a research based on full-scale investigation has also been done to compare the three different CT prediction approaches: CT10, integrated disinfection design framework (IDDF), and CFD, to determine the most appropriate method for design and operation of ozone systems. The CFD approach yielded more accurate predictions of inactivation efficacy than the other two approaches. The current results also suggested that the differences in the three approaches in CT predictions became smaller at higher contactor T10/T ratios conditions as the contactors performed more closely to ideal plug flow reactors.
This study has demonstrated that the computational fluid dynamics (CFD) approach is an efficient tool for improving ozone disinfection performance of existing water treatment plants and designing new ozonation systems. The model developed in this study can be used for ozone contactor design, evaluation, and troubleshooting. It can also be used as a virtual experimental tool to optimize ozone contactor behaviour under varying water quality and operational conditions.
Identifer | oai:union.ndltd.org:WATERLOO/oai:uwspace.uwaterloo.ca:10012/2669 |
Date | 05 December 2006 |
Creators | Zhang, Jianping |
Source Sets | University of Waterloo Electronic Theses Repository |
Language | English |
Detected Language | English |
Type | Thesis or Dissertation |
Format | 2967955 bytes, application/pdf |
Page generated in 0.0017 seconds