project

Digital twin demonstrator

Expert(s):
Ina Vertommen MSc, Mario Castro-Gama MSc, Frederic Béen PhD, Dimitrios Bouziotas MEng. MSc

  • Start date
    01 Jul 2019
  • End date
    01 May 2020
  • Principal
    Platform hydroinformatica
  • collaborating partners
    De Watergroep, Dunea, Brabant Water en Evides

This project will create a so-called ‘digital twin’ of a Dutch city’s drinking water distribution system, and use it to demonstrate the added value of digital twins in the case of anomaly detection.

A digital twin is a digital replica of a living or non-living physical entity. In the context of technical systems, a digital twin is often seen as a digital representation of the physical system, which mirrors the real twin as perfectly as possible, in terms of its behaviour and outcomes. Data from various sources can be incorporated into the model, in order to make the mirroring of the real system sufficiently accurate for application purposes. Making a connection with other models is also common.

Application to drinking water

When it comes to drinking water, there are two obvious areas of application. The first is the detection of anomalies in the system’s operation: whenever a digital twin shows a different behaviour (a different pressure reading at a node, for instance) compared with the real twin, it is reasonable to assume that either the digital twin is not entirely representative, or that an anomaly exists in the real network’s condition (such as an unknown leak) or operation (such as a wrong valve position). For this application it is sensible, if not essential, to have the digital twin reflect the current situation, which is why live measurement data are often assimilated into the model. The second area of application is scenario studies, in which the system’s operation is examined under all kinds of (extreme) conditions for the purpose of improving it. In this case the assimilation of measurement data is less important.

Although hydraulic models can in a sense be considered (limited) digital twins, they are not as a rule fed with real-time information, nor are they connected to other models. The objective of this project is to demonstrate a digital twin of a distribution system that assimilates real-time data and, moreover, is connected to another model of a relevant connected system (in this case, water demand forecasts). In this way, a contemporary element is added to the concept of a digital twin as applied to a drinking water distribution system.

The project’s underlying objective is to demonstrate the added value of the concept’s application. This will be done for a specific case, namely, the detection of network anomalies. Based on this demonstration, a panorama will also be sketched showing other promising and valuable applications for digital twins, along with the associated courses of action open to water utilities.