Optimal Design of Wastewater Treatment Plants Subjected to Time-Varying Inputs

The objective of this thesis is to develop a computational framework for the optimal design of wastewater treatment plants that takes into account dynamic variations in both flow rate and wastewater compositions. These variations can be very large over the course of a year (e.g. +/-100%) and it is therefore important to explicitly take them into account in the design rather than relying on averaged values. Our approach is to minimize an economic objective function using a heuristic optimization algorithm (particle swarm). Since we use the detailed industry standard activated sludge model no.3 (ASM3), this problem is computationally very challenging. However, the presented work shows that this method is a very effective one. We demonstrate a cost saving of 67% when our novel method is compared with existing methods based on averaged inputs.