Long-term efficiency of water networks with demand uncertainty

Abstract

Water distribution network (WDN) design problems generally consider minimising network costs while meeting minimum allowable pressure under specified demand patterns. However, extension and renovation of networks based on assessment of demand uncertainty causes an increase in efficiency of the system and monetary savings. Attention to uncertain changes in pattern and the value of nodal demands in an operational period can thus show the effect of uncertainties in the network's efficiency. Using Monte Carlo simulation and shuffled frog leaping algorithm optimisation, this paper analyses the efficiency of a designed network based on nodal demand uncertainty during the operational period. Assessment of a network efficiency index according to a probabilistic normal distribution for nodal demand in an operational period shows that a network design that does not consider nodal demand uncertainty decreases network efficiency, especially at the end of the operational period. Moreover, an increase in demand uncertainty decreases network efficiency. The results of the proposed model for a benchmark WDN show that a network designed deterministically for a 30-year operational period is unable to secure nodal demands with suitable pressure after 10 years. In addition, the minimum required pressure considered in network design is met in approximately the first 18 years of the operational period. Water distribution network (WDN) design problems generally consider minimising network costs while meeting minimum allowable pressure under specified demand patterns. However, extension and renovation of networks based on assessment of demand uncertainty causes an increase in efficiency of the system and monetary savings. Attention to uncertain changes in pattern and the value of nodal demands in an operational period can thus show the effect of uncertainties in the network's efficiency. Using Monte Carlo simulation and shuffled frog leaping algorithm optimisation, this paper analyses the efficiency of a designed network based on nodal demand uncertainty during the operational period. Assessment of a network efficiency index according to a probabilistic normal distribution for nodal demand in an operational period shows that a network design that does not consider nodal demand uncertainty decreases network efficiency, especially at the end of the operational period. Moreover, an increase in demand uncertainty decreases network efficiency. The results of the proposed model for a benchmark WDN show that a network designed deterministically for a 30-year operational period is unable to secure nodal demands with suitable pressure after 10 years. In addition, the minimum required pressure considered in network design is met in approximately the first 18 years of the operational period.