A New Weighted Modularity-Based Approach for DMA Demarcation in Large and Complex Water Distribution Networks

Water distribution networks (WDNs) need restructuring/sectorization into District Metered Areas (DMAs) depicting smaller communities for ease in ensuring equitable water distribution and pressure management. DMA demarcation also helps in systems operation and management, apart from leak and contaminant localization. Many different strategies for DMA demarcation are in use, as none is established to be universally superior. Hence, there is ambiguity in the choice of a strategy for DMA demarcation. A WDN can be viewed as a complex network due to strong interconnections among its components and imposed limitations (being a 2D network). Against this backdrop, the recent decade witnessed the use of community detection approaches from complex network theory (CNT) for DMA demarcation. Community detection is a very basic yet pivotal task in the field of CNT. Modularity maximization is the most widely used approach for community detection. The modularity index defines the quality of the subgraphs or communities delineated from a network. In the case of a WDN, some nodes may be shared by more than one DMA, in which case the conventional and existing variants of the modularity index cannot be used for assessing the quality of the delineated DMAs. A more comprehensive community (DMA) detection procedure must incorporate such nodes with multiple associations among communities. Facilitating this, the present study proposes a comprehensive approach for DMA demarcation in large and complex WDNs considering a weighted modularity index. Edge weights are assigned to incorporate the hydraulic behaviour of a network and the association of nodes among various communities. The efficacy of the proposed approach vis-à-vis existing methods is demonstrated through a case study on a benchmark WDN. Effective demarcation of DMAs helps in their prioritization (based on the existing network level measures) to devise mitigation strategies for improving their performance.