Pressurized fluid-distribution networks are strategic elements of infrastructure. In the case of fresh-water distribution networks, advanced sensor-based diagnostic methodologies have the potential to provide enhanced management support. Since a significant percentage of fresh water is lost globally due to leaks in these networks, the challenge to improve performance is compatible with goals of sustainable development. The scope of this research includes the diagnosis of water-distribution networks and more generally, pressurized fluid-distribution networks through development of model-based data-interpretation methods to assess performance. The strategy of model falsification is combined with network reduction techniques to obtain reliable and computationally efficient diagnoses. A case study involves the detection of leaks from an initial set of 263 leak scenarios. Preliminary results show that this methodology has the potential to detect leak regions, even with a small number of sensors.