Novel Approaches in Optimization of

Wastewater Systems Rehabilitation

Chiang, Po-Kuan

April 2006

Abstract

In recent years, there has been increasing concern for wastewater system failure and identification of optimal set of remedial works requirements. So far, several methodologies have been developed and applied in asset management activities by various water companies worldwide, but often with limited success. In order to fill the gap, there are several research projects that have been undertaken in exploring various algorithms to optimize remedial works requirements but mostly for drinking water supply systems and very limited work has been done for wastewater assets. Some of the major deficiencies of commonly used methods can be found in either one or more of the following aspects: inadequate representation of systems complexity, incorporation of a dynamic model into the decision-making loop, the choice of an appropriate optimization techniques and experience in applying the technique. This study is oriented towards resolving these issues, discusses the system performance simulated by the urban drainage and sewer system model MOUSE (DHI, Denmark) and four optimization techniques performed by a global optimization tool GLOBE (Solomatine 1995, 1998). Four urban drainage sub-systems are considered as four case studies. These urban drainage systems respectively have 5, 12, 63, and 185 pipes and their design events based simulation of runoff process are simulated in MOUSE model. The problem of rehabilitation is based on considering two conflicting criteria: minimization of flooding damage, and minimization of rehabilitation costs. In this study, the multi-criterial problem is solved by combining the two criteria following the two options: (a) compromise between the two criteria, and (b) ensuring no flood damage. These two options will provide decision makers with a possibility to determine the most suitable rehabilitation strategy.

The final results show that the minimal composite cost and minimal total surcharge are realized by means of using optimization tool in all (four) cases. In addition, Adaptive Cluster Covering (ACCO), Adaptive Cluster Covering with Local search (ACCOL), Genetic Algorithm (GA), and Multi-start algorithm (M-Simplex) also demonstrate their good search ability in the optimization process to design optimal pipe network of this wastewater system. In this research, it was demonstrated that the global optimisation techniques present an efficient technology to achieve objectives posed.

Keywords: Remedial works requirements; Optimization techniques; Urban drainage system; MOUSE model; GLOBE tool; Adaptive Cluster Cover (ACCO); Adaptive Cluster Cover with Local Search (ACCOL); Genetic Algorithm (GA), Multi-start Algorithm (M-Simplex)

Back to the list of MSc abstracts