Optimum Rehabilitation of a Water Distribution System Considering Cost and Reliability

Abstract

This paper describes the development of a multi-objective genetic algorithm that can be used to identify the optimum tradeoff between cost and reliability for a set of pipe replacement decisions. Reliability is defined in terms of the expected number of customer interruptions per year. It may be estimated using a break model and a hydraulic model for the pipes in a region. The tradeoff curve can be used to identify the reliability that can be achieved for a specified expenditure on pipes and the pipes that should be replaced to achieve this. The optimum schedule of pipe replacement over time can also be identified. The approach is demonstrated using data from part of the water supply system in Adelaide, Australia.