Electronic Thesis and Dissertation Repository


Doctor of Philosophy


Chemical and Biochemical Engineering


Professor Jesse Zhu, Professor George Nakhla


Biological nutrient removal (BNR) refers to removal of carbon (C), nitrogen (N) and phosphorus (P) from wastewater (WW) by the aid of various microorganisms. Because of the public concern for the environment C, N and P effluent standards have become stricter. Different BNR processes such as suspended growth and attached growth have been studied during the last three decades in order to meet the increasingly stringent discharge standards.

In this work, two novel processes called Twin Fluidized Bed Bioreactor (TFBBR) and Twin Circulating Fluidized Bed Bioreactor (TCFBBR) were developed and tested for BNR from municipal WW. Both TFBBR and TCFBBR comprise of an anoxic column and an aerobic column with particle recirculation between the two reactors achieved mechanically (TFBBR) and hydraulically (TCFBBR). Moreover, a newly developed system called Anaerobic Fluidized-Circulating Fluidized Bed Bioreactor (AF-CFBBR) was developed and tested to accomplish BNR from high strength industrial WW. AF-CFBBR comprises of an anaerobic, an anoxic and an aerobic columns. In all three aforementioned systems, fine carrier media are employed for biofilm attachment. After the development of biofilm, the particles are called biofilm-coated particles.

TFBBR and TCFBBR were operated at organic, nitrogen and phosphorus loading rates (OLR, NLR and PLR) of up to 2.8 kg COD/m3×d and 4.5 kg COD/m3×d, 0.3 kg N/m3×d and 0.5 kg N/m3×d and 0.041 kg P/m3×d respectively to study the performance of the system with respect to biological nutrient removal. The nitrification rates based on biofilm surface area in TFBBR and TCFBBR were 0.91 g N/m2×d and 1.26 g N/m2×d respectively and the denitrification rates based on biofilm surface area in TFBBR and TCFBBR were 0.65 g N/m2×d and 1.32 g N/m2×d respectively. Both systems removed >96% organic matter, 84%-88% nitrogen and 12%-50% phosphorus at overall hydraulic retention time of (HRT) 2h. TFBBR and TCFBBR achieved long SRTs of 72-108 d and 37-40 d respectively, which rationalized the very low observed yield of 0.06-0.07 g VSS/g COD and 0.09-0.1 g VSS/g COD. The AF-CFBBR demonstrated 99.7% COD removal, 84% nitrogen removal, with a very low sludge yield of 0.017 g VSS/g COD while treating a wastewater containing 10700 mg COD/L and 250-300 mg NH3-N/L. The system was operated at an organic loading rate (OLR) of 35 kg COD/m3·d based on the AF volume and 1.1 kg N/m3·d based on the CFBBR at an overall HRT of less than 12 h in the AF-CFBBR. The nitrification, denitrification and organic removal rates based on aerobic, anoxic and anaerobic biofilm surface area in AF-CFBBR respectively were 2.6 g N/m2×d, 9.03 g N/m2×d and 12.1 g COD/m2×d. Additionally, the inhibitory effect of nitrate on methanogenic activities in a high rate anaerobic fluidized bed with organic loading rate of above 35 kg COD/m3·d was studied in order to evaluate the feasibility of simultaneous denitrification and methanogenic activities (SDM) in a high rate anaerobic system.

Terminal settling velocity and bed expansion of biofilm-coated particles as the two main hydrodynamic criteria in a fluidized bed, were studied. Archimedes was superior to Reynolds number for drag coefficient and bed expansion definitions. A new equation for determining drag force on fluidized bed bio-film coated particle (Fd) as an explicit function of terminal settling velocity was generated based on Archimedes numbers (Ar) of the biofilm coated particle. The proposed equation adequately predicted the terminal settling velocity of other literature data with an accuracy of >90%. A new equation based on Archimedes number was proposed to calculate bed expansion index of biofilm-coated particles, which predicted the existing experimental data with less standard error than all other literature equations that related bed expansion to Reynolds number. A two-phase and three-phase predictive fluidization model based on the characteristics of a system such as media type and size, flow rates, and reactor cross sectional area was proposed to calculate bed expansion, solid, liquid and gas hold up, specific surface area of the biofilm particles. The model was subsequently linked to 1d AQUIFAS APP software (Aquaregen) to model two and three phase fluidized bed bioreactors. The model was validated for biological nutrient removal using the experimental data from a Twin Circulating Fluidized Bed Bioreactors (TCFBBR) treating synthetic and municipal wastewater. Two-sided t-tests showed that there were no statistically significant difference between the experimental and the modeled TCOD, SCOD, NH3-N, NOx-N.