Comparison of demand pattern calibration in water distribution networks with geographic and non-geographic parameterization

Abstract

Demands are one of the most uncertain parameters in water distribution network models. A good calibration of the model demands leads to better results when using the model for any purpose. A demand pattern calibration methodology that uses a priori information has been developed for calibrating the behavior of demand groups. In cities, similar demand behaviors are distributed all over the network, contrary to smaller villages where demands are clearly sectorised in residential neighborhoods, commercial zones and industrial sectors. In this work, demand pattern calibration has a final use for leakage detection and isolation. Detecting a leakage in a pattern that covers nodes spread all over the network makes the isolation unfeasible. Besides, demands in the same zone may be more similar due to the common pressure of the area rather than for the type of contract. A demand pattern calibration methodology is applied to a real network with synthetic non-geographic demands for calibrating geographic demand patterns. The results are compared with a previous work where the calibrated patterns were the original non-geographic ones