The Climate Resilience Demonstrator (CReDo), a digital twin project first started in 2021, has moved into its second phase, now led by Connected Places Catapult.
I last reported on the CReDo project in Spring 2022, shortly after the first phase had completed.
In phase 1, our aim was to develop a digital twin of assets from Anglian Water, BT and UK Power Networks, and to investigate and model the cascade effects of assets across different infrastructures. For example, if power was lost, pumping stations could fail, and communications could fail. In phase 1, we focused on flood modelling and the failure models used were relatively straightforward – if an asset was dry it was working, if was wet, it was not working.
Data held on DAFNI, the Data & Analytics Platform for National Infrastructure, covers detail such as ground elevation, water sources, precipitation, and drainage. Maps included in the modelling show individual power station assets and communications infrastructure to a fine scale, allowing researchers to model the situation under average rainfall, as well as to identify the impact of a one in 100 or one in 1000 years ‘extreme climatic’ event.
As well as the real-life data for East Anglia, CReDo includes a synthetic dataset that’s available to any user and allows researchers outside the project to examine how they might build on CReDo and DAFNI for their own research. Developed by CMCL, it was designed to have features like the real data but with fictitious locations.
In the second phase of work we have expanded into more sophisticated failure models around flooding, and have added the variable of extreme heat. We are also now working more with Ofgem and Ofwat data – on the energy and water side.
There is cross pollination across energy and water. Both Ofgem and Ofwat are interested in the effects of heat, for example. If there are several days of unusually hot weather, what happens to your water pumping station? The processes could degrade, not necessarily fail, so now our modelling has been developed to be able to take into account more sophisticated impacts such as degradation rather than simply working/not working. The basic principle from phase 1, of investigating cascade effects across networks still applies.
DAFNI continues as the trusted platform which allows the project’s “integration” based on asset owner data – assessing impacts of cascade failures from one infrastructure to another.
The project is now reaching out to further electricity and water companies, and network operators, to encourage them to become part of this essential digital twin.
An exciting new development that STFC Hartree and DAFNI worked on recently is what we’ve named the “digital elicitation tool”. With this tool we sought to move expert knowledge that’s often in people’s heads, based on experience and learning understanding outside of a formal process. We wanted to enable this tacit knowledge to become explicit. For example, what could be warning signs of a substation asset failure, what circumstances could lead up to this, and what might be the likelihood of failure?
Often expert knowledge is simply in people’s heads. We aimed to capture the worries of people who run sub stations and transformers and networks, and their experience and specialist knowledge in areas such as rates of failure and degradation. They are the domain experts and one of the strengths of DAFNI is that we are the modelling and computing experts who can write workflows and complex Python codes.
The elicitation tool is a separate component which we will implant into the CReDo structure. In the Alpha stage of this, part of our CreDo work under the Ofgem strategic innovation fund, we created a prototype rather than a full production version, and focused on probabilistic elicitation – so we could capture the likelihood of an asset failing under certain conditions.
We also plan to make the elicitation tool available to energy companies who haven’t yet signed the data sharing agreement, to gather additional input from them.
We are now exploring integrating DAFNI’s BSRW work with the elicitation tool, and will be reporting on this at the 2024 DAFNI Conference in September.
In the future, we could add roads to the project and there is certainly interest from DAFNI users in working with the CReDo digital twin too – for example, if flooding occurs in an urban environment, where are the obstacles, and where would the water flow… what happens in a dynamic environment…