Author(s): Armando Di Nardo; Michele Di Natale; Michele Iervolino
Linked Author(s): Michele Iervolino, Armando Di Nardo
Keywords: No Keywords
Abstract: The modern paradigm of water supply networks management asks for reduction of excessive leakages. Even if water losses may derive from several different processes (junction obsolescence, pipe creeping, etc..), their intensity is usually assumed to scale with a suitable power of the pressure head. The same law is commonly employed in the context of mathematical modelling to compute the quota of nodal discharge due to water losses. In such a case, the pressure/discharge relationship may also depend on the degree of skeletonization of the network model. Literature reports several empirical relationships based on field data. A huge fluctuation in the values of the parameters of pressure/discharge relationship is anyway observed. So, the choice of a suitable leakage law represents a major uncertainty in water network modelling, while being nevertheless an unavoidable step for both the analysis of existing scenarios and the design of leakage control techniques based on pressure management. In the present paper, the simultaneous identification of hourly demand and leakage pattern is performed using time histories of pressure and flow at few sites on the network, with the aid of genetic algorithms. The proposed procedure is applied to a field test case.
Year: 2009