The sketch on the right in figure 2 schematically shows the allocation of demands (in this case building areas) to nodes using a Voronoi diagram. In this case the actual demand is calculated by using a coefficient relating the population to a building area unit and a consumption per capita value. Alternatively, building areas can be allocated to the nearest nodes based on the Voronoi diagrams if the correspondent data is available.

If demands and their location are not known, they can be determined based on the size of the Voronoi regions, the population distribution and density and a consumption per capita value. This approach is also commonly applied to determine, for example, the rainfall catchment areas of inlets of sewer networks. Thus, known demands (value and location) can be allocated to the closest nodes according to the Euclidean distance. The borders of the Voronoi regions constitute of lines called bi-sectors. They represent the catchment areas of the nodes. In the following these areas are referred to as Voronoi regions.

The Voronoi diagram is based on a set of nodes and defines one area for each node with every point within the area being closer to the originating node of the area than to any other node. The Voronoi diagram is also known as Thiessen polygon and dual to the Delaunay triangulation (c.f. To allocate demands to the nodes Voronoi diagrams might be used. Besides, location and demand of consumers are often not known, especially in developing countries. demand at a house connection) is modelled as a node. To reduce the model size typically not every known demand (e.g. Illustration of a water distribution system and the corresponding model graph ( Klingel, 2010)

Changing pipe length watergems