Adaptive Greedy-Heuristic Algorithm for Redundancy Augmentation by Loop Addition in Branched Water Distribution Systems

Abstract

Many people in developing countries do not have access to sufficient water supply according to the World Health Organization. In addition, most water distribution systems are branched due to their lower pipe costs and, therefore, do not have redundancy in case of a pipe break. As funds do become available, the question arises as to which loop to connect first to provide the most redundancy at the least cost. A loop selection algorithm suitable for the educational background of development workers, employing an adaptive greedy heuristic algorithm, is proposed and demonstrated on two real branched systems in Central and South America in which a sufficiently large pipe was added to each possible loop with adjacent dead-end junctions and the redundancy and cost quantified. These possible loops were then ranked in terms of their ability to provide water service to the most users for the least cost. It was found that first priority must be given to the possible loop that connects the water source to an additional system junction since, if the sole pipe that connects the source to the system fails, then no users have access to water. After that, connecting upstream loops first provided the most redundancy if all users were considered of equal importance, but not always at the least cost, since upstream loops may cost more to connect due to (1) long lengths, or (2) neighboring pipes having to be enlarged to handle the extra flow caused by the connection of the loop.