| Abstract: | Wastewater treatment plants play a crucial role in our modern lives processing sewage and reinserting this vital fluid into the natural water cycle. Estimates show that the United States consumes about 1% of their electricity in this process, and about 80% of residential solid waste is handled using these plants. These plants play a fundamental role in the continued sanitation of towns and cities. They do, however, present the electrical grid and its energy suppliers with a constant, baseline demand that is predictable. The research presented here attempts to enhance the control of these plants to provide cost savings through consumer-driven load shifting and an overall reduction in site peak demand. Because of their abundance, even a slight demand reduction percentage could also have substantial implications at the grid level. Model predictive control is used in simulation across multiple plant models and configurations to control the sewage pumping process and optimize for plant demand while taking into consideration a set of regulatory and physical constraints. Using model predictive control as the basis for both power management and plant operation is a novel approach that works to some extent. Not all elements of a plant contribute to the deferrable demand that can be utilized in a plant. In fact, some elements work against the proposed system. On average and over the time windows needed to change a month's worth of demand charges, a tuned controller might be able to reduce peak demand of its sewage-pumping motors by anywhere from 20-30%. That being said, model predictive control is highly sensitive to initial conditions, tuning parameters and other changing plant variables, so further work is needed to validate any level of robustness from this solution. Wastewater treatment plants come in a variety of configurations, sizes, and type, so not all are covered in this manuscript. Equally as important, the work presented here can be easily applied to many other municipal and industrial plant types with similar results. The advantage of systems like wastewater treatment and water distribution plants is precisely how ubiquitous they are. The electronic version of this dissertation is accessible from https://hdl.handle.net/1969.1/198503 |
|---|