A Multistep Optimization Approach to Rehabilitate a Deteriorated and Intermittent Water Distribution Network

Intermittent water supply increases operational and maintenance costs for the water utilities and leads to a negative impact on the customers. In the context of the Battle of Intermittent Water Supply (BIWS), a challenge is posed to restore the operation of a water distribution network (WDN) by the application of a list of possible interventions, to be optimally planned based on a five-year annual budget. The methodology proposed is based on a multistep optimization procedure considering nine given hydraulic key performance indicators (KPIs). First, a preanalysis step is performed where graph theory is applied to identify critical components in the system. Subsequently, a clustering method is carried out to highlight the areas of the system characterized by worse KPI values (i.e., below an acceptable level). A Multiobjective optimization (MOO) is performed for year zero and used as a weighting factor for the prioritization of the investment in terms of leakage reduction. The combination of clustering analysis and MOO allows for the selection of the best interventions during each year.