A review of pipe network optimisation techniques

Abstract

Millions of dollars are spent each year on water distribution supply infrastructure. Computer analysis of pipe networks is now common place, however, the use of pipe optimisation techniques has been limited. The use of optimisation techniques provides an opportunity for potential savings in costs for water supply authorities. This paper reviews four optimisation techniques including partial enumeration, non-linear programming, linear programming and genetic algorithms. The first 3 techniques have been applied to pipe network optimisation in the research literature over the last 15 years or so. Genetic algorithms provide a new approach to pipe network optimisation pioneered over the last 3 years at the University of Adelaide. Details of each method are presented in this paper. Each method is then applied to a case study and the ease of use of each method is assessed. Genetic algorithms and non-linear programming provide the least-cost solutions.