An investigation into control strategies for activated sludge wastewater treatment plants
The activated sludge process is widely used throughout the world for the treatment of wastewater from domestic and industrial users. This process is not normally efficiently controlled and hence increasingly important financial incentives and environmental considerations exist for improving the efficiency and quality of the treatment before releasing the treated water into the environment. This thesis presents the development of MATLAB computer simulation models for activated sludge wastewater treatment plants. A comparison of control systems has been made using these models for typical operating conditions of wastewater treatment processes, such as influent flow pattern and temperature. The investigation identified the control of dissolved oxygen as an important area to study because insufficient levels of dissolved oxygen in the wastewater prevent the successful degradation of organic matter present, whereas too high a level causes settling problems and inefficiencies. Three dissolved oxygen control methods, namely PID, Fuzzy logic and self-tuning control have been investigated, applied and their performances compared. As in most other processes, the number and location of sensors and actuators within a water treatment plant can have large implications for successful process control. Therefore, the model developed was used with real plant data to test different designs and investigate the best location of sensors and actuators for a specific North West Water plant to improve control of the process. Optimisation of process operation has also been investigated with the objective of improving effluent quality and reducing operation costs. Simulations suggest that all three dissolved oxygen control methods investigated are able to control the process satisfactorily with relatively little deviation from the setpoint. The PID and fuzzy logic controllers needed retuning for changing process conditions, but the adaptive nature of the self-tuner makes it more robust. Optimal sensor and actuator placements have been identified and a cost/quality benefit analysis performed. Significant cost reductions and effluent quality improvements may be achieved by applying optimisation techniques to regulating the concentration of the solids within the aeration stage. These objectives are conflicting and therefore simultaneous improvement is not always achievable. The project has demonstrated the potential benefits of employing models to simulate the process, subject to availability of data to parameterise them. Process operation can be significantly improved with the application of well-tuned controllers and optimisation techniques.