Utility Week

29 ISSUE 03 MAY/2019 days using conventional techniques, says Hawes: " e power goes exponential when you start to move past data collection and visualisation on its own to link your models to the real system and unlock value from those additional insights." Digital twins are still a nascent technology and various technical, economic and behavioural challenges mean it may be many years before utilities have fully functional models of their entire assets. B i r d ' s e y e v i e w e most advanced twins today typically focus on small-scale high criticality problems, such as a specific issue with an energy substation, or long-term planning scenarios that are more tolerant to a lack of real-time data sources and rigorously accurate engineering models. Building an all-encompassing 'bird's eye' view of a network, including live operational data, would require major investment to retrofit IoT sensors into existing infrastructure and upgrade IT systems. Some networks are constrained by older infrastructure that's been in place for 40-plus years. However, that level of forensic detail is not necessary to benefit from the technology, says Samuel Chorlton, project lead at the Data and Analytics Facility for National Infrastructure (DAFNI): " ere isn't one digital twin that solves all things, there are multiple types that a business might look to develop to answer different business questions and the associated data required to drive that is going to vary too." When working with clients, Smart Infrastructure tries to break down the need for sensors into a problem-solving exercise around 'data management, sense making and decision making'. "When clients follow that process they often find they have some of what they need already, and where there are gaps data science takes care of it and we can build the insight piece from what we have," says Hawes. e success of digital twins is largely reliant on data interoperability and the ability energy generation and the increasing penetration of renewables. Rohit Banerji, global lead on the development of big data analytics platforms at Accenture, observes: "Our clients are beginning to ask for twins to dynamically plan generation, operational routing, and to run scenarios to understand how the congestion will play out in the bi-directional grid. For example, if solar was to grow in a particular part of the country, or a new wind farm was introduced, how will it impact on congestion?" Power distribution companies have used twins to simulate the future impact of electric vehicles and related high- volume power storage and fast charging on a network. Digital twins are very efficient engineering problem solvers. Where a traditional desk study might require extensive manual data gathering, engineering calculations and analysis using software or spreadsheets, twins can combine real-time data streams from various sources and run them through sophisticated algorithms and machine learning to produce rapid results. Mott MacDonald's Smart Infrastructure business used this type of approach when developing the digital twins for Auckland Council and a "number of water and sewage companies in the UK", which cannot currently be named for non-disclosure reasons. Oli Hawes, head of smart infrastructure, says: "We're trying to tap into real-time data from across their businesses, whether that's from rain gauges, their GIS [geographic information system] or their customer relationship management system. Many data sources are themselves already digital twins, for example a digital representation of a hydraulic model built in software. We connect all that together, then layer in other whole system data sources, such as the national weather, river level data from the Environment Agency or tidal information, to create a digitised desk study that runs in real time." e Moata platform is able to run around 30 different scenarios in three seconds, where previously it would have taken 14 SPACE TWIN STARTED IT ALL e Apollo 13 space mission had the world on tenterhooks in 1970 when the oxygen tanks exploded early in the mission and the crew became engaged in a fight for survival. Key to the rescue mission, which required technical issues to be resolved from up to 200,000 miles away, was a digital twin model of the spacecraft, which was used by NASA engineers on Earth to test out possible solutions. Many data sources are themselves already digital twins, for example a digital representation of a hydraulic model built in software

• Cover