The search for the worst case transient loadings in water distribution systems

Estimating appropriate water demands to design a distribution system is difficult and their continuously time-varying characteristic has the potential both to cause and to modify water hammer problems that might result, in the worst case, in catastrophic pipeline failure. To identify and protect against the worst case, this paper searches the set of the possible water hammer loadings in water distribution systems to find the most severe transient loading. Evolutionary approaches, in particular genetic algorithms and particle swarm optimization, are combined with transient analysis to represent a variety of loading conditions and to identify the worst-case scenario. This approach shows that not only are the loading conditions important when searching for the worst case in a pipeline network, but also the selection of the system characteristics such as system topography, pipe size, material and thickness are crucially important to prevent or mitigate the worst case events in the system.