A proposal of topological sampling design

In the developed countries the majority of water distribution networks (WDNs) were built in the last century and today the principal task is to manage those hydraulic systems for different socio-economic and environmental purposes. Thus, there is a need for building WDN models to predict the system behaviour and support decision of water managers. For this reason once built a WDN model using field information, one problem is to estimate some boundary conditions (e.g. pipe hydraulic resistances, nodal demands) which cannot be directly measured. This procedure is also called model calibration and requires the availability of observations of the hydraulic system. These observations are generally pressure measurements because cheaply available. The related devices need to be properly located and distributed in the system. The selection of the location of observations and their number is named sampling design and plays an important and sometimes decisive role in order to achieve an accurate calibration, then a good predictive model. This paper proposes a sampling design method based on a topological analysis in order to reduce the size of the mathematical inverse problem related to model calibration and to achieve better pre-conditions for the observability of the hydraulic system