‘Collaboration and resilience through connected digital twins’ – CReDo Show-and-Tell Webinar

Question 

Your answer 

What is the approach to cyber security to protect critical national infrastructure? What are the benefits to the participants providing the data? 

[Matt Webb] Benefits to participants providing the data relate to long term planning, medium term event readiness and near real-time operational activities. 

The insight provided by the connected digital twin facilitates improved insight into asset criticality, risk, and the consequences of failure.  This aids in informing and justifying investment planning and intervention. 

Medium term event readiness can be considered in the context of the build-up to impactful events, allowing targeted, proactive mitigation and resource deployment to reduce impact and accelerate repair and restoration activities. 

Near real-time application facilitates in-event insight into present impact and cascade impact across critical infrastructure, enabling enhanced coordination between asset organisations. 

In all respects, network resilience is enhanced, investments optimised and asset damage and associated costs reduced. 

[Ben Mawdsley] In a CReDo only context: 

All asset owner data was transferred under encryption and passed to STFC through either a physical medium or internet transfer for storage on DAFNI. The decryption key was provided separately, and data only stored on a single, mutually agreed data store located behind the STFC firewall and requiring multi factor authentication to access. Users were kept to the minimum number of staff necessary.  

If working independently, no single asset owner would have access to all of the data present in the twin. Working together, we can identify the connections across these different datasets and learn the dependencies, allowing us to build models with greater functionality than three independent models would produce. If we can accurately model future impacts, we can make more informed decisions now to mitigate these. 

There’s also the case that sharing data now makes sharing data in the future easier- if we can make the changes now, we can make future applications easier to develop. 

Can you share more details of the thin slice approach please? 

Which data standards have been used to ensure data interoperability?  Breaking down the data silos? 

[Matthew West] There are two papers that are in draft awaiting publication by CPNI that will say more about the thin slice approach. They are; 

Managing Shared Data that talks about the approach in a wider context, and  

Developing Thin Slices that talks more about the bottom up methodology Tom Burgoyne described.  

Contact me if you want to see pre-publication drafts of these papers. matthew.west@informationjunction.co.uk 

The current available data standards are what we call Industry Data Models, like the Industry Foundation Classes, and Reference Data Libraries, like UniClass2015.  The IMF Team has done a survey of the ones available that have been brought to our attention: 

https://digitaltwinhub.co.uk/files/file/89-a-survey-of-industry-data-models-and-reference-data-libraries/ 

The problem here is that none of these IDMs are compatible with each other. So, the problem for the IMF Team is to develop a data model (the Foundation Data Model) that is able to integrate data from them. In this process we will be treating industry data models as thin slices. 

So far have you any indication from the Asset Owners as to how the information will be applied and any extra visualisation required 

[Matt Webb] My above response largely covers this.  With respect to visualise, a range of data visualisations could potentially applied – geospatial, schematic, charts, etc. – to cater for these various use cases. 

[Louise Krug] The ideas are that we can use the model to help direct investment – what sites are most at risk from different climate events, where risks are elevated due to interconnectedness 

We would also like to use the Twin to support operational teams during extreme events – help the teams from different organisations cooperate  

Other types of output are likely to be needed – lists of top 10 sites at risk for example rather than just visualisations  

Visualisations may be a good way to interact with the system to say “take this out” , “move this element” or “replace these values” and have that run through the model to see its impact. 

[Matt Edwards] Now that we have a working demo we can socialise internally we will use it as a means to develop more detailed user stories. We see potential opportunities for use already through all stages of the asset lifecycle; planning investment, design and engineering / construction, operations and maintenance, customer and service management, incident and safety management, and decommissioning.  

We are completely open to the need for further visualisations; as soon as data is presented in context (information), people and users always want more! Digital products should always be developed with opportunities for continuous improvement in mind. 

[Jethro Akroyd] Other types of output are likely to be needed – lists of top 10 sites at risk for example rather than just visualisations. 

Visualisations may be a good way to interact with the system to say “take this out”, “move this element” or “replace these values” and have that run through the model to see its impact 

Are you testing the model in collaboration with research institutions in other countries? 

[Chris Dent] On the application side, we are focused on the UK context. The methodology required for this kind of study is however universal, and we are developing links with analogous projects in Europe and the USA with a view to collaboration on future phases of underpinning research. 

[Jethro Akroyd] Yes. Many of the ideas underlying the use of the knowledge graph in the CReDo are inspired by the World Avatar project, which is a collaboration between CMCL Innovations (who implemented the CReDo digital twin), the University of Cambridge and Cambridge CARES, the university’s research centre in Singapore. 

{Ben Mawdsley]  All project out puts will be available as Open source or permissive licences that are accessible so anybody, here in the UK or overseas, could contribute to building models. 

– How can technology and data providers get involved and support the mission ? Would be interested to learn more about engagement channels, procurement and Collaboration opportunities 

[Yalena Coleman] Please contact yalena.coleman@cp.catapult.org.uk for any enquiries about how to get involved in Phase 2, we will be happy to set up calls to explore various routes for collaboration / procurement.  

[Chris Dent] In addition, the research organisation partners would be very pleased to discuss with interested parties how their needs can drive future research agendas, and how they might collaborate with and provide use cases for follow-on research. 

[Matt Edwards] At Anglian Water we have our Water Innovation Network (WIN) – WIN is a free partnership initiative run by not for profit organisation Allia and Anglian Water. It is the platform Anglian Water uses to connect and engage with potential supply chain, which includes innovators, individuals and businesses.  Water Innovation Network (anglianwater.co.uk). 

can you envisage having several levels of openness regarding simulations?- e.g at least one version where some data could be viewed by the public – where security and privacy re assets was not compromised 

[Matt Webb] Yes – I think that is both appropriate and necessary. 

Given the nature of the data, the extent of integration and the insight provided, limitations of access need to be considered for all use cases and end-users.  The degree of ‘openness’ will vary in line with this. 

[Louise Krug] That is the purpose of the simulated data system – simulated data used for all aspects that can’t be openly shared 

{Matt Edwards] Yes. Anglian Water increasingly develop solutions in this fashion; Digdat (digdat) and In your Area (In Your Area (digdat.co.uk)) being a couple where the public have access to visualisations and data that in reality are far more complex and enriched when used internally. The aim though is to present only views to the public that their user stories require. The products are always being evaluated for improvement. We are aware we need to constantly challenge what is open to the public, and we use data sensitivity classifications as one of the ways in which we understand access requirements as well as challenge ourselves. 

Are there any intentions to expand this example to include transportation networks? 

[Yalena Coleman] Absolutely – transport network information is already being tabled for discussion as an additional data source in Phase 2 – including road networks both national and local. 

Good to see progress over the past few months. Are the Data Sharing Agreements available to share? Also there was mention at the launch event of SLAs with Regulators being revisited. Any progress on that front? 

This is a huge search space to look for improvements to the various networks. What thoughts do you have about how to support the actors to search this? 

[Chris Dent] This is one of the directions in which we wish to expand the Operational Research activity within CReDo. The visibility across the three networks will alone provide new insights into where critical areas for reinforcement lie. We can also potentially use mathematical optimization tools to search potential options – in practice this would work by helping develop a list of candidate projects for evaluation by the asset owners. 

[Jethro Akroyd] Increase the level of detail in the digital twin so that it can describe things like switching between alternative supply routes, and link it to models that suggest possible mitigating actions. Automate the running of the digital twin to systematically investigate different options.) 

I’m aware that the role of bridges in carrying pipes and cables has become apparent during several flood events, and the bridge collapses have led to power cuts, severance of gas mains etc. Have you considered including bridges which carry pipes/cables in future versions of the model? 

[Matt Webb] This is an interesting point.  Bridges are just one form of civil asset that could be considered in this context, many of which are shared by multiple asset owners.  Having visibility of these ‘pinch-points’ would be of significant value from a risk mitigation point of view. 

A good real-world example of this is the Kingsway incident in central London several years ago.  In that instance, the interaction of telecoms, electricity, gas and water assets in a common subterranean tunnel system resulted in an extensive and sustained tunnel fire severely damaging all of those networks. 

[Jethro Akroyd] One of the next steps in the project is to assess what additional information might be needed in the digital twin. This is a very interesting suggestion. NB This naturally links to the question above [and below] about including transport 

It is great to see in the first presentation that the National Digital Twin is seen as a socio-technical change. While there is great technical development evident from the presentations, the social aspect of sociotechnical system is not as evident. Any work planned to analyse the socio-technical aspects of having this connected digital twin thinking ? for example, the social actors, and the dynamic relationship between technical and social aspects of a connected digital twin ? Thank you. 

 [Jethro Akroyd] I recently published a paper where we applied the same underlying design of digital twin to enable interoperability between socio-technical data (energy usage, climate and fuel poverty data) to investigate how the adoption of heat pumps might affect inequality: https://doi.org/10.1016/j.adapen.2021.100079. The connection with CReDo is that this work uses exactly the same design [i.e. ontologies and a knowledge graph] that CMCL used for CReDo 

What Open Source data did you use? 

 [Jethro Akroyd] By first intent, the CReDo digital twin was implemented using open source software with permissive licences. Full details are given in the technical reports. The digital twin was developed using confidential data about real assets. The demonstration at the webinar used synthetic data that the project team created and that will be published under a permissive open licence. The demonstration at the webinar also showed the ability to incorporate open data from the Environment Agency and Ordnance Survey. Details will be published in a forthcoming report. 

[Ben Mawdsley] Flood models HiPIMS are open source software too. Uses UKCP18 data.