The specific research work described in this thesis forms part of a much larger research project that was funded by the Technology Programme of the UK Government. This larger project considered improving the design and efficiency of membrane bioreactor (MBR) plant by using modelling, simulation and laboratory methods. This research work uses phenomenological mechanistic models based on MBR filtration and biochemical processes to measure the effectiveness of alternative behavioural models based upon input-output system identification methods. Both model types are calibrated and validated using similar plant layouts and data sets derived for this purpose. Results prove that although both approaches have their advantages, they also have specific disadvantages as well. In conclusion, the MBR plant designer and/or operator who wishes to use good quality, calibrated models to gain a better understanding of their process, should carefully consider which model type is selected based upon on what their initial modelling objectives are (e.g. using either a physically mechanistic model or an input-output behaviourial model). Each situation usually proves unique. In this regard, this research work creates a "Model Conceptualisation Procedure" for a typical MBR which can be used by future researchers as a theoretical framework which underpins any newly created model type. There has been insufficient work completed to date on using a times series input-output approach in the model development of a wastewater treatment plant, so only general conclusions can be made from this research work. However, it can be stated that this novel approach seems to be applicable for a membrane filtration model if care it taken to select appropriate input-output model structures, such as those suggested in the "Model Conceptualisation Procedure". In the case of the development of a MBR biological model, it is thought that a conventional Activated Sludge model produced by the IWA could be coupled to a input-output model structure as suggested by this report to give a hybrid model structure that may have the advantages of both model types. Further research work is needed in this area. Future work that should follow on from this research study should focus on whether these input-output models could be used for predictive control purposes, whether an integrated model could be created, and whether a benchmark could be created for the three main MBR configurations.

Advanced wastewater treatment plants must meet permit requirements for organics, solids, nutrients and indicator bacteria, while striving to do so in a cost effective manner. This requires meeting day-to-day fluctuations in climate, influent flows and pollutant loads as well as equipment availability with appropriate and effective process control measures. A study was carried out to assess performance and process control strategies at the Falkenburg Road Advanced Wastewater Treatment Plant in Hillsborough County, Florida. Three main areas for control of the wastewater treatment process are aeration, return and waste sludge flows, and addition of chemicals. The Falkenburg AWWTP uses oxidation ditches where both nitrification and denitrification take place simultaneously in a low dissolved oxygen, extended aeration environment. Anaerobic selectors before the oxidation ditches help control the growth of filamentous organisms and may also initiate biological phosphorus removal. The addition of aluminum sulfate for chemical phosphorus removal ensures phosphorus permit limits are met. Wasting is conducted by maintaining a desired mixed liquor suspended solids (MLSS) concentration in the oxidation ditches. For this study, activated sludge modeling was used to construct and calibrate a model of the plant. This required historical data to be collected and compiled, and supplemental sampling to be carried out. Kinetic parameters were adjusted in the model to achieve simultaneous nitrification-denitrification. A sensitivity analysis found maximum specific growth rates of nitrifying organisms and several half saturation constants to be influential to the model. Simulations were run with the calibrated model to observe relationships between sludge age, MLSS concentrations, influent loading, and effluent nitrogen concentrations. Although the case-study treatment plant is meeting discharge permit limits, there are several recommendations for improving operation performance and efficiency. Controlling wasting based on a target MLSS concentration causes wide swings in the sludge age of the system. Mixed liquor suspended solids concentration is a response variable to changes in sludge age and influent substrate. Chemical addition for phosphorus removal should also be optimized for cost savings. Finally, automation of aeration control using online analyzers will tighten control and reduce energy usage.

Advanced wastewater treatment plants have complex biological kinetics, time variant influent rates and long processing times. The modeling and operation control of wastewater treatment plant gets complicated due to these characteristics. However, a robust operational system for a wastewater treatment plant is necessary to increase the efficiency of the plant, reduce energy cost and achieve environmental discharge limits. These discharge limits are set by the National Pollutant Discharge Elimination System (NPDES) for municipal and industrial wastewater treatment plants to limit the amount of nutrients being discharged into the aquatic systems. This document summarizes the research to develop a supervisory operational and control system for the Valrico Advanced Wastewater Treatment Plant (AWWTP) in the Hillsborough County, Florida. The Valrico AWWTP uses biological treatment process and has four oxidation ditches with extended aeration where simultaneous nitrification and denitrification (SND) takes place. Each oxidation ditch has its own anaerobic basin where in the absence of oxygen, the growth of microorganisms is controlled and which in return also helps in biological phosphorus removal. The principle objective of this research was to develop a working model for the Valrico AWWTP using BioWin which mimics the current performance of the plant, predicts the future effluent behavior and allows the operators to take control actions based on the effluent results to maintain the discharge permit limits. Influent and experimental data from online and offline sources were used to tune the BioWin model for the Valrico Plant. The validation and optimization of the BioWin model with plant data was done by running a series of simulations and carrying out sensitivity analysis on the model which also allowed the development of operation policies and control strategies. The control strategies were developed for the key variables such as aeration requirements in the oxidation ditch, recycle rates and wastage flow rates. A controller that manipulates the wasting flow rate based on the amount of mixed liquor suspended solids (MLSS) was incorporated in the model. The objective of this controller was to retain about 4500-4600 mg/L of MLSS in the oxidation ditch as it is maintained by the Valrico Plant. The Valrico AWWTP recycles around 80% of their effluent and hence, the split ratios were adjusted accordingly in the model to recycle the desired amount. The effluent concentrations from the BioWin model for the parameters such as Total Nitrogen (TN), Ammonia, Nitrate, Nitrite, Total Kjeldahl Nitrogen (TKN) complied with the discharge limits which is usually less than 2 mg/L for all the parameters.