Optimal Upgrading of Water Distribution Network Redundancy

Abstract

A new two-stage iterative methodology is presented for upgrading of single-source water distribution networks to achieve redundancy. In the first stage, the existing network is strengthened using linear programming (LP) for a fixed flow distribution obtained using a model that maximizes flow uniformity by minimizing variance of flows in the distribution network. In the second stage, the performance of the network is assessed under single-pipe failure conditions using a node-flow analysis, and the most critical pipe-failure condition having the maximum deficiency in meeting nodal demands is noted. To also make the network satisfactory during the failure of a critical pipe, a new flow distribution is obtained by removing the critical pipe. Constraints for this new flow distribution along with previous flow distribution are then simultaneously considered in a new LP problem. An improved design of the network is obtained by repeating the first stage. This two-stage methodology is continued until the network becomes capable of meeting demands at all nodes during any single link failure. The proposed methodology is illustrated by upgrading a water main system for redundancy. The results of the application demonstrate that the new methodology is more efficient and less computationally demanding as compared to other methodologies used for upgrading of the illustrative water main system.