Thesis (MScEng) -- Stellenbosch University, 2014. / ENGLISH ABSTRACT: Urine contains up to 80% of nitrogen, 50 % of phosphates and 90 % of potassium of the total load in domestic wastewater but makes up less than 1% of the total volume (Larsen et al., 1996). The source separation and separate treatment of this concentrated waste stream can have various downstream advantages on wastewater infrastructure and treated effluent quality. The handling of undiluted source separated urine however poses various challenges from the origin onward. The urine has to be transported to a point of discharge and ultimately has to be treated in order to remove the high loads of organics and nutrients. Wilsenach (2006) proposed onsite treatment of source separated urine in a sequencing batch reactor before discharging it into the sewer system. This study focused on the treatment of urine in a sequencing batch reactor (SBR) primarily for removal of nitrogen through biological nitrification-denitrification. The aim of the study was to determine nitrification and denitrification kinetics of undiluted urine as well as quantification of the stoichiometric reactions. A further objective was to develop a mathematical model for nitrification and denitrification of urine using experimental data from the SBR. The SBR was operated in 24 hour cycles consisting of an anoxic denitrification phase and an aerobic nitrification phase. The sludge age and hydraulic retention time was maintained at 20 days. pH was controlled through influent urine during volume exchanges. Undiluted urine for the study was obtained from a source separation system at an office at the CSIR campus in Stellenbosch. Conditions in the reactor were monitored by online temperature, pH and ORP probes. The OUR of the system was also measured online. One of the main challenges in the biological treatment of undiluted urine was the inhibiting effect thereof on nitrification rate. The anoxic mass fraction was therefore limited to 17 % in order to allow longer aerobic phases and compensate for the slow nitrification rates. Volume exchanges were also limited to 5% of the reactor volume in order to maintain pH within optimal range. Samples from the reactor were analysed for TKN, FSA-N, nitrite-N, nitrate-N and COD. From the analytical results it was concluded that ammonia oxidising organisms and nitrite oxidising organism were inhibited as significant concentrations of ammonia-N and nitrite-N were present in the effluent. It was also concluded that nitrite oxidising organisms were more severely inhibited than ammonia oxidising organisms as nitrate-N was present in very low concentrations in the effluent and in some instances not present at all. Ultimately the experimental system was capable of converting 66% of FSA-N to nitrite- N/nitrate-N of which 44% was converted to nitrogen gas. On average 48% of COD was removed. A mathematical model was developed in spreadsheet form using a time step integration method. The model was calibrated with measured online data from the SBR and evaluated by comparing the output with analytical results. Biomass in the model was devised into three groups, namely heterotrophic organisms, autotrophic ammonia oxidisers (AAO) and autotrophic nitrite oxidisers (ANO). It was found that biomass fractionation into these three groups of 40% heterotrophs, 30% AAO and 30% ANO produced best results. The model was capable of reproducing the general trends of changes in substrate for the various organism groups as well as OUR. The accuracy of the results however varies and nearexact results were not always achievable. The model has some imperfections and limitations but provides a basis for future work.