Prior to full-scale implementation of process modifications at the Brantford WTP, a pilot-scale treatability study was conducted to investigate intermediate ozonation/AOP and to determine the most suitable granular media (anthracite, GAC, and Filtralite®) for deep-bed biological filtration. The primary objectives of this research were to provide insight into the destruction of natural and synthetic organics and assess ozonated and halogenated DBP formation. Ozone alone was unable to achieve the 1-log removal target for geosmin or MCPA, unless disinfection-level dosages were applied. No improvement was observed when adding hydrogen peroxide. A major obstacle to the implementation of ozonation in bromide-laden source waters is the formation of bromate. There is a direct correlation between ozone dose and bromate formation and by applying ozone dosages at disinfection levels, bromate is likely to exceed regulatory limits. However, adding hydrogen peroxide reduced the amount of bromate formed, and in most cases levels fell below regulatory limits. A linear correlation was established between bromate inhibition and increasing H2O2/O3 ratio at constant ozone dose. Amongst the three filtration media investigated, only GAC achieved 1-log removal for geosmin and MCPA. The superiority of GAC over anthracite and Filtralite® was attributed to its adsorption affinity. Filtralite® and anthracite media were both ineffective for MCPA removal due to its non-biodegradable nature under conventional water treatment conditions. At a 1 mg/L-ozone dose, GAC and Filtralite® filters achieved a 1-log geosmin removal. In contrast, a 1.44 mg/L ozone dose was required to meet this target with anthracite. The tandem of ozone followed by biological filtration was very effective for the control of distribution system TTHM production regardless of filter media, with levels well below current and anticipated provincial regulatory limits. The combination of intermediate ozonation followed by deep-bed biological filtration is well suited for treating Grand River water. Scale-up considerations include pairing the proper filter media to the size of the ozone generator. The best two treatment scenarios were: Option 1: select GAC media and size the ozone generator to produce a 1 mg/L dose. Option 2: select anthracite media and size the ozone generator to deliver a 2 mg/L dose.

Biological filtration (Biofiltration) processes have been used first in Europe and then in North America for decades, however currently there is not a good overall parameter to guide biofiltration design and operation except adopting parameters from traditional particle- removal filtration process. On the basis of the biofilm model developed by Rittmann and McCarty (1980a) and the pseudo-analytical solution for the model, Zhang and Huck (1996a) obtained an analytical solution for PF (plug flow) reactors (which can be used for biofilters approximately) after demonstrating that axial dispersion could be reasonably ignored and developed a new parameter, X*, which incorporates considerations of physical contact time, filter media particle size, kinetics, etc. A small-scale application on peers’ engineering/research data by Huck (1999) demonstrated it was a better indicator than other parameters for biofiltration performance. By collecting, screening and investigating literature on AOC, BDOC and odorous compounds removal by biofiltration process, this thesis applied the X* concept to the collected investigations to assess process performances among different target parameters, different filters and different investigations. To the author’s knowledge, this is the first such attempted comprehensive comparison of literature studies, interpreted in terms of a common parameter (X*). The wide ranges of particle sizes, EBCTs, temperatures and high diversity of pre-treatment and operation conditions for the collected cases were considered to be able to well represent biofiltration practices for studied removal targets. No significant relationship between EBCTs and removal percentages were found, indicating that EBCT alone is not able to guide biofiltration design and operation. Based on kinetics parameter comparison, BDOC removal-X* relationship was established. A new parameter, θα, was developed in this thesis to refer to estimated X* values only considering EBCT and particle size. θα parameter values were estimated by comparison of ratios of θα products ((θα)’) based on the properly chosen calculation bases. Distribution of the θα values for temperature-favored (i.e. temperature ≥15°C) AOC and BDOC removal biofiltration processes matched the established removal-X* relationship reasonably. Given the exploratory nature of this research and the complexity of attempting quantitations, fits were assessed based on visual comparison. With the assistance of supporting information and by adopting available temperature activity coefficients, temperature-adjustment coefficients for θα values were determined for the different temperature ranges. Temperature-adjusted AOC and BDOC removal-θα relationships were developed and temperature-adjusted θα parameter values for AOC and BDOC removal were also estimated. Comparisons were conducted, showing fair matches based on visual examinations, for most of the temperature ranges. No relationships were found between ozone dosages and AOC/BDOC removal percentages and the statistical analysis indicated there was significant difference of removal efficiencies between ozonated and non-ozonated influents for biofilters, suggesting ozonation may not only increase the amount of BOM for following biofilter and increase the biodegradability of bulk water; it may also increase the biodegradability of AOC and BDOC themselves. It may not be realistic to obtain the estimated θα values for MIB and geosmin removal by biofiltration. However, plotting θα product vs. removal percentage for the collected MIB and geosmin removal cases shows more positive co-relationships than EBCT-removal percentage relationships visually. A utilization factor η was proposed to guide biofilter design and operation and to assess “over-design” and “under-operated”. Biofilter over-design or under-operated is common for the collected cases. In general, examining X* (or θα, a parameter incorporating the physical components of X*) provided useful information in terms of evaluation and prediction of biodegradable organic compounds removal by biofiltration, which confirms that X* is a better parameter for biofiltration design and operation than other parameters, such as EBCT.

Biological filtration (Biofiltration) processes have been used first in Europe and then in North America for decades, however currently there is not a good overall parameter to guide biofiltration design and operation except adopting parameters from traditional particle- removal filtration process. On the basis of the biofilm model developed by Rittmann and McCarty (1980a) and the pseudo-analytical solution for the model, Zhang and Huck (1996a) obtained an analytical solution for PF (plug flow) reactors (which can be used for biofilters approximately) after demonstrating that axial dispersion could be reasonably ignored and developed a new parameter, X*, which incorporates considerations of physical contact time, filter media particle size, kinetics, etc. A small-scale application on peers’ engineering/research data by Huck (1999) demonstrated it was a better indicator than other parameters for biofiltration performance. By collecting, screening and investigating literature on AOC, BDOC and odorous compounds removal by biofiltration process, this thesis applied the X* concept to the collected investigations to assess process performances among different target parameters, different filters and different investigations. To the author’s knowledge, this is the first such attempted comprehensive comparison of literature studies, interpreted in terms of a common parameter (X*). The wide ranges of particle sizes, EBCTs, temperatures and high diversity of pre-treatment and operation conditions for the collected cases were considered to be able to well represent biofiltration practices for studied removal targets. No significant relationship between EBCTs and removal percentages were found, indicating that EBCT alone is not able to guide biofiltration design and operation. Based on kinetics parameter comparison, BDOC removal-X* relationship was established. A new parameter, θα, was developed in this thesis to refer to estimated X* values only considering EBCT and particle size. θα parameter values were estimated by comparison of ratios of θα products ((θα)’) based on the properly chosen calculation bases. Distribution of the θα values for temperature-favored (i.e. temperature ≥15°C) AOC and BDOC removal biofiltration processes matched the established removal-X* relationship reasonably. Given the exploratory nature of this research and the complexity of attempting quantitations, fits were assessed based on visual comparison. With the assistance of supporting information and by adopting available temperature activity coefficients, temperature-adjustment coefficients for θα values were determined for the different temperature ranges. Temperature-adjusted AOC and BDOC removal-θα relationships were developed and temperature-adjusted θα parameter values for AOC and BDOC removal were also estimated. Comparisons were conducted, showing fair matches based on visual examinations, for most of the temperature ranges. No relationships were found between ozone dosages and AOC/BDOC removal percentages and the statistical analysis indicated there was significant difference of removal efficiencies between ozonated and non-ozonated influents for biofilters, suggesting ozonation may not only increase the amount of BOM for following biofilter and increase the biodegradability of bulk water; it may also increase the biodegradability of AOC and BDOC themselves. It may not be realistic to obtain the estimated θα values for MIB and geosmin removal by biofiltration. However, plotting θα product vs. removal percentage for the collected MIB and geosmin removal cases shows more positive co-relationships than EBCT-removal percentage relationships visually. A utilization factor η was proposed to guide biofilter design and operation and to assess “over-design” and “under-operated”. Biofilter over-design or under-operated is common for the collected cases. In general, examining X* (or θα, a parameter incorporating the physical components of X*) provided useful information in terms of evaluation and prediction of biodegradable organic compounds removal by biofiltration, which confirms that X* is a better parameter for biofiltration design and operation than other parameters, such as EBCT.