Written by Chris Jones (Northumbrian Water Group), Professor Chris Kilsby (Newcastle University) and Brett Cherry (Newcastle University)
When responding to an incident that has potential to adversely affect people and their homes or businesses, it is essential to have a plan that sets out roles and responsibilities, standard procedures and lines of communication. Every incident is unique. As the incident evolves, the nature and extent of its impact as well as options available to responders, depend on the location, time of day, antecedent conditions and a host of other dynamic factors.
It is possible to develop detailed plans for a few scenarios with potentially damaging consequences, although good practice would dictate that the potential impact is reduced through mitigating actions as far as possible. Even so, planning around a few scenarios that might take place in the future, at least helps design mitigation and highlights the possibilities of unexpected consequences. A much better option would be to use real-time modelling and simulation to produce bespoke response plans and dynamic decision support for any incident, taking into account the location and conditions at the time, and updating as the incident evolves.
This idea of combining real-time data with digital modelling and visualisation to create actionable insight is what lies at the heart of the digital twin. At first glance this may sound like something out of a science fiction novel: an accurate, intelligent real-time simulation that has oversight over multiple systems and could communicate through an AI interface. But as with other technologies of the digital age that have surpassed expectations, the digital twin is very much science fact.
Digital twin has been used for many different kinds of systems, the automotive and aircraft industries use digital twin for car and jet engines. NASA has used digital twin since the 1970s and has developed a digital twin for the Langley Research Center. Digital twin has potential applications for almost any process for business and industry in any sector. It can be used for redesigning, adapting and improving systems, and systems of systems: digital twins could be applied to a single pump, a pumping station (with multiple pumps), a system of underground pipes, or potentially the entire water network, creating opportunities for innovation and customer engagement never before possible. The ultimate goal of the digital twin is to enhance the ways we currently manage the water network and to improve outcomes for customers and the environment.
Water infrastructure is extensive and complex; its many parts and operations make possible essential water and sanitation services to millions of customers, both residential and commercial. But there are many things about the water network, the surrounding built and natural environment and the communities served, that may not be possible to know in advance or at least not within enough time to make a decision to act. To make more confident decisions for responding to incidents requires tools that provide insights based on all the available and up to date information, which is what makes the digital twin of particular interest to the water industry.
A project between Newcastle University and Northumbrian Water Group is developing a digital twin called ‘Twincident’ which responds to incidents in the water network, such as a burst pipe or a heavy rainfall event. These incidents cause significant disruption to customers and communities, and can affect the finances, health and wellbeing of those affected. Incidents also risk causing environmental damage and potentially come at large financial cost to the infrastructure company.
The idea for Twincident came about at the 2018 NWG Innovation Festival this summer. The aim of this digital twin is to run simulations of an area during an incident to show what could happen over a 24 hour period, in just 30 seconds. This is potentially a powerful tool for the water industry creating a new way forward for how it responds to events, specifically to enable employees in the field to respond faster and more effectively to emerging issues.
Newcastle University is already developing many of the building blocks for the Twincident digital twin, such as novel hydraulic modelling tools, spatial analysis tools and approaches for enabling simulation and decision support using different models working in concert. They also work on interfaces to allow non-expert users to use the tools, and for the city of Newcastle. Data from LIDAR maps, land cover maps, and buildings, are integrated with water company infrastructure information and dynamic weather, traffic and air quality data from the Urban Observatory (a digital urban sensing network of over 2,700 metrics freely available online).
Models provide insights into elements of the water and wastewater system, for example to identify leakage from the water network, blockages in sewers and drainage areas likely to lack capacity under more extreme rainfall. Typically models are only as good as the data available. They require considerable time and effort to accurately represent the real world, and considerable computing power to deliver results in reasonable timeframes.
Monitoring sewers and pipes is important to understand (both directly and through modelling) what quality of service is provided but the extent of monitoring is limited by access, power and communication – new sensors are needed for both physical, biological and chemical measurements to fully reflect network conditions. Also, individual models do not reflect the potential interdependencies of infrastructure systems: for example, a burst water pipe might lead to wastewater surcharge and pollution; surface water flooding could lead to traffic congestion, preventing operatives reaching the location and dealing with the problem. To provide real-time insight across the water system as a whole requires something considerably more advanced than what is within the industry’s reach at the moment.
Where the digital twin is especially powerful is when it comes to testing future scenarios. If you want to test the resilience of water infrastructure against a major flood event it could be simulated in the digital twin, providing an excellent tool for management and planning. Other possibilities include joining together rural catchment models, with land use and ground water models to anticipate and mange threats to resources, or wastewater and the sewer network. Some models are yet to be developed: the role of green infrastructure in flood mitigation currently needs to be parameterised, especially in response to rainfall; and the influence of customer behaviour on demand in the short and long term. Newcastle University works on a similar model for electricity that could be adapted for water and wastewater services.
As with any disruptive innovation, implementing the digital twin is not without challenges. Water companies are generally quite good at installing individual systems in the water network, and monitoring equipment, but they’re not often connected together in a common framework. For the digital twin to work properly these would need to be integrated in future.
Digital twin clearly provides numerous opportunities for the water sector and may revolutionise how the sector responds to incidents in future. It makes it possible. While a digital twin of an entire city may be some ways off, we now have at our disposal the modelling and data monitoring capabilities to identify many problems before they occur, and make better decisions about how to respond to incidents that impact people, communities and businesses. The integration of cloud computing, data science, urban sensing and internet-of-things makes digital twin an incredible planning and response tool for addressing the growing complexity of the water network, and the uncertainties surrounding environmental change.
