Integrating Uncertainty in Performance Assessment of Water Distribution Networks by Scenario Building

Abstract

This paper presents and demonstrates a novel scenario-building methodology that integrates contextual and future time uncertainty into the performance assessment of water distribution networks (WDNs). A three-step approach is proposed: (i) System context analysis, identifying the main key factors that impact the WDN performance; (ii) Scenario definition, identifying the implicated WDN variables, describing its possible evolution, and conjugating them to further establish the reference scenario and the two most relevant and opposite ones; and (iii) Scenario modelling, simulating the WDN behaviour for those scenarios. The obtained spatial and temporal hydraulic results are further used to calculate performance metrics. The methodology is applied to a real WDN to assess resilience performance considering infrastructure asset robustness (real water loss performance indicator), service reliability (minimum pressure index), and service flexibility (network resilience index). A new formulation to assess the metric evolution over time is proposed, deducting the further-away performance results by using an uncertainty weight. The results demonstrate that the increase in metric amplitude for the opposite scenarios over time highlights future uncertainty, reflecting context uncertainty, and the comparison of metric spatial distribution (i.e., at the pipe/node levels) highlights critical areas with higher associated uncertainty.

Keywords

• drinking water networks
• aleatory uncertainty
• scenario planning
• scenario modelling
• resilience metrics