Systematic Surge Protection for Worst-Case Transient Loadings in Water Distribution Systems

Abstract

Estimating appropriate water demands for the design of a distribution system is itself difficult, but the continuously fluctuating nature of these demands has the added potential of creating water hammer problems that might result in catastrophic pipeline or system failure. To first identify and then avoid these eventualities, this paper searches a predefined set of possible water hammer events in water distribution systems to identify the most severe transient loadings and then conducts a search for suitable surge protection strategies. Genetic algorithms and particle swarm optimization are combined with transient analysis first to identify a set of worst-case loads and then to seek an optimal protection strategy to cope with them. Case studies show that the worst case is not always obvious and cannot always be assumed a priori to correspond with high or low demand scenarios. Both the search for the worst-case loading and its associated optimal protection strategies are strongly sensitive to the characteristics of both the pipe system and the candidate transient events.