Effects of spatial correlation of demands on the cost and resilience of water distribution networks: A new design optimization approach

Design optimization procedures for water distribution systems generally assume that the nodal demands follow the same diurnal pattern, which implies that the demands are fully synchronised and perfectly correlated. However, it is well known that in practice the spatial correlation of the demands is not perfect; i.e. the cross-correlation is less than unity. This study investigates the effects of the widespread assumption that demands are perfectly correlated on the overall performance and capital cost of water distribution systems. A new optimal design approach that considers the cross-correlation of the nodal demands was developed herein. Essentially, the novelty of the research is that the optimization procedure incorporates the spatial properties of the demands in terms of the cross-correlation while preserving the diurnal demand patterns. Thus, in addition to the cross-correlation effects, the nodal demand patterns are assumed to be independent and unique and are addressed explicitly using extended period simulation.