An hybrid approach for considering topography in graph-based optimization of water distribution networks

Vital components of the urban water infrastructures are water distribution networks (WDNs), which transport water from production (sources) to spatially distributed consumers (sinks). Multi-objective optimization procedures are performed to minimize the construction costs and at the same time maximize the resilience of such systems which is usually a very computational expensive task. Recently, highly efficient approaches based on complex network analysis (CNA) have been developed to solve this task more efficiently. With CNA very large WDNs can be optimized, taking into account network topology and demand distribution (demand edge betweenness centrality EBCQ). However, up to now these approaches are not able to consider the topography (i.e., the height differences between sources and sinks). Comparing CNA designs with design solutions from evolutionary algorithms (EA), shows that least cost design from CNA cannot compete with those from EA.
This work aims to enhance CNA based optimization of WDNs by considering the topography. This is achieved by a hybrid approach, where low cost design solutions from CNA are evaluated with an hydraulic solver (Epanet2), and subsequent the EBCQ distribution is iteratively altered. By giving demand nodes, which have low pressure a higher weight in the EBCQ calculations, the reported short-coming is eliminated.
The enhanced CNA based optimization is tested with 4 different large case studies and reported least cost designs from literature and shows promising results ( < 2% costs increase) but with tremendously less computational efforts, enabling tackling very large WDNs (>150,000 decision variables).