OPTIMIZATION OF WATER SUPPLY NETWORKS CONSIDERING RELIABILITY BY USING THE LEAST VARIANCE DISCHARGES

Abstract

A water supply network should be capable to meet the consumers demand, in normal and abnormal flow situations, with proper pressure in all parts of the network. This research aims to find a method in which a network is designed to meet the customer need with adequate pressure in addition to the lowest cost, in fact the aim of network cost optimization with regard to the reability. In order to apply reliability, uniformity of flow distribution network pipes is considered and to create this scale, the concept of Discharge variance of pipes is used. Optimization using genetic algorithms has been done, and its cost function, in addition to the investment and annual energy costs per year, includes the expected downtime and repair costs per year. Reliability is checked in two modes; in the first mode, only operational state of the network pipes is investigated; while in the latter mode, the removal of individual pipes as well as the overall operational state was considered. Since in the state of removed pipes, EPANET software is not able to do correct analysis, its modified version is used which is called EPANET + Water Net Gen. The comparison of results shows that, in general, use of this definition provides high levels of reliability.