Water distribution network disruptive events. Generation and exploration of an incident hub to increase the network preparedness

Abstract

[EN] This paper seeks to develop increased knowledge about disruptive events in a water distribution network (WDN) through the experience acquired by previous anomalous events in the system. This work explores the various relationships between several parameters in an incident hub (specifically water loss events) in a Spanish real, small WDN. The incident hub consists of basic elements recorded during an incident (e.g. breakdown, maintenance activity, among others) and the corresponding causes that generated the incident (e.g. breakage due to excess pressure, breakage due to tree roots, etc.), as well as the management times of the incident (e.g. awareness time, isolation, and repair time). The utility collected and stored these data, which were completed with direct interviews with the system operator. Measurements were performed at pressure and flow sensors, which allowed evaluating the effects of both the incident itself and the actions taken to solve it. The records of the incidents are categorised depending on the nature of the data they contain to facilitate mapping their causes and effects. To characterise the disruptive events, a feature extraction process has been proposed using a temporal-spatial approach combined with a migration proposed that describes parameters’ behaviour in the spatial dimension for a certain period of time. The characteristics obtained in the previous lessons of the incidents contained in the incident hub are compared with potential causes obtained with different control parameters. The objective is to determine the potential causal relationship of the incident that allows its characterisation. The results of this characterisation are presented and analysed in this contribution. The outcomes are promising in the sense of a clear ability to provide WDNs with key parameters that foster prediction and classification processes.