High pressure versus low pressure? How GA optimization was used for the cost-effective design of the rehabilitation of loxton irrigation district

Abstract

A genetic algorithm optimisation study has been conducted for the proposed pressurised pipe distribution system for the rehabilitation of the Loxton Irrigation District. The original off-farm water distribution system was constructed soon after World War II and although more than half of it was piped it operated as an open channel system. Project managers -SA Water intended to use a genetic algorithm study to optimise design of the new distribution system, but expanded the study to assist in determining the optimum pumping and residual pressures. The minimum residual pressures considered in the study were 3m, 20m, 25m, 30m, 35m and 40m. The study investigated only costs that were likely to vary between the various residual pressure options. The results of the study were then used by the growers of the Loxton Irrigation District as a basis for their decision as to which pressure option to adopt for the new system. Total costs for the 3m minimum residual pressure option are close to $34.7 million, and range up to $45.1 million for the 40m minimum residual pressure option. As a result of the genetic algorithm optimisation study, the growers of the Loxton Irrigation District adopted a residual pressure of 35m, even though they must pay 100% of the incremental cost above a system with a residual pressure of 3m. The decision was unanimous based on the fact that the 35m residual pressure system would effectively not cost any more than the 3m residual pressure system when on-farm expenses and the like were added into the equation.