-
Towards a Web of Digital Twins discussion
-
2 minutes ago, olivierthereaux said:
This is incredibly important, and I agree that the first connections will need to be between similar twins and relevant sub-systems. The whole approach (and I hope what the talk conveys) is that over time, these connections will grow and scale up to be a whole web.
Yes, fully agree! To develop these standards and protocols we need to first understand the use cases that these interconnected twins need to facilitate. There may be a need, for example, for protocols to support different levels of aggregation to suit these different use cases.
-
3 minutes ago, Nicholas said:
Thanks Chris and Andrew. I think that you have been doing some interesting work to look at connections between twins in different sectors/areas (or twins that connect across these sectors). Are there any specific thoughts on what might help networks of cross-sector twins?
I think the main limiting factor on this has been how few projects we’ve seen that involve more than one twin (either cross sector or within a sector). I think we have lots of ideas on how twins could be connected, and some ideas where connected twins could add more value, but most of the investment is still focused on individual twins and how we get data into and out of them. However, some of the protocols for ingesting data into Twins could be directly applicable to sharing data between twins in future (e.g. web service protocols).
-
17 minutes ago, Andrew Myers said:
That was a really interesting talk.
Northumbrian Water explored some similar themes at our Innovate East event last year.
Some of the outcomes/discussions points are below:
Standards need to be simple and broadly applicable. We don’t want a high barrier for adoption
- Core mandatory data items (as few and simple as possible)
- Optional but standardised data items (e.g. BIM/Uniclass)
- Any other data items (flexible and extendable standards)
Aim to keep the data sharing as simple as possible – how many types of data do we need? What level needs to be shared?
- Time Series Data – e.g. temperature sensor readings
- Event Data – e.g. an engineer visit, a status change
- Configuration or Meta Data – e.g. date of manufacture
Agree with this also. One of the things we’ve been doing within BIM4Water is using Uniclass as a common classification for data items at an equipment level (pump, valve) – but one of our groups run by a colleague David Bell, has started to look at how we can apply this common classification at higher levels of our Water Industry data hierarches, so facilitate potential standardisation at higher levels. It would be interesting to look at how this approach could be applied at a cross-sector leve.
-
4 minutes ago, Guest Yu(Steven) ZHANG said:
The intermediaries reminds me of Data Trust. “A data trust is an intermediary between data providers, data users and other stakeholders in the sharing and use of data. ” ref: https://theodi.org/article/data-trusts-decision-making-report/
Indeed.
Data Trusts are one of the many approaches for data stewardship though – not all use cases will require fiduciary responsibility. We have been trying to map the variety of approaches here: https://theodi.org/project/the-data-access-map/
-
Hi @olivierthereaux
What is the relationship between a network of twins and a National Digital Twin for the built environment please?
-
Just now, Nicholas said:
What is the relationship between a network of twins and a National Digital Twin for the built environment please?
The two concepts are very similar! The NDT programme is in effect helping build the web of twins, with a focus on creating positive impact at the national infrastructure level.
The notion of a “web of twins” attempts to go beyond national boundaries however: we found it useful in our research to test how far the concept could be scaled up, and given that we will likely have use cases for collaboration and data flows across borders, the “web of twins” we envisaged in our research is worldwide.
-
17 minutes ago, Nicholas said:
Thanks Chris and Andrew. I think that you have been doing some interesting work to look at connections between twins in different sectors/areas (or twins that connect across these sectors). Are there any specific thoughts on what might help networks of cross-sector twins?
Just reading through Olivier’s ODI Data Access Map – a bit too much to digest immediately on first glance. What does strike me about that, and the conversation in the thread, is that we’ve very quickly got hung up on data, which is essentially the lowest level of the twin concept (ref the triangle diagram in the Gemini Principles). Far greater challenged will emerge as we seek to connect twins as this will require co-modelling across vastly different systems, with huge differences in temporal and spatial scales, it will require much more complex decision support and solution space optimisation and for twins to be designed and operated in a manner that makes the benefits they bring to be made available to non-expert users.
-
5 minutes ago, Chris Jones (NWG) said:
What does strike me about that, and the conversation in the thread, is that we’ve very quickly got hung up on data, which is essentially the lowest level of the twin concept (ref the triangle diagram in the Gemini Principles). Far greater challenged will emerge as we seek to connect twins as this will require co-modelling across vastly different systems, with huge differences in temporal and spatial scales, it will require much more complex decision support and solution space optimisation and for twins to be designed and operated in a manner that makes the benefits they bring to be made available to non-expert users.
Agreed! The presentation tries to cover that in some respect.
When we started exploring this “web of twins” I also expected to mostly be thinking about (raw) data flows, but we quickly realised that there is just as much value, if not more, in getting collaboration (and sharing) at the higher levels (models, insights, decisions).
-
1 minute ago, olivierthereaux said:
Agreed! The presentation tries to cover that in some respect.
When we started exploring this web of twins I also expected to mostly be thinking about (raw) data flows, but we quickly realised that there is just as much value, if not more, in getting collaboration at the higher levels (models, insights, decisions).
Great. Have you any thoughts on how that might work?
-
19 minutes ago, Richard Stirland said:
Agree with this also. One of the things we’ve been doing within BIM4Water is using Uniclass as a common classification for data items at an equipment level (pump, valve) – but one of our groups run by a colleague David Bell, has started to look at how we can apply this common classification at higher levels of our Water Industry data hierarches, so facilitate potential standardisation at higher levels. It would be interesting to look at how this approach could be applied at a cross-sector leve.
Thanks for this Richard. Pipes and valves (e..g. steam pipes in power plants) are one of the examples we are seeing on need/value for cross-sector consistency on taxonomies and classifications.
-
1 minute ago, Chris Jones (NWG) said:
Great. Have you any thoughts on how that might work?
(that = sharing at higher levels)
To be fair, we didn’t explore that in as much depth as I would like. This still feels “early days” with too few well-developed use cases to truly speak authoritatively.
I can certainly picture sharing of insights and predictive models across an industry based on operation of digital twins. I think in some of our research material, we mention the use case of wind turbine repair, and the fact that operators in similar situations may want to share insights about the right way to run maintenance based on the operation of their DTs, but that may not require any data sharing at the raw level.
-
I’m interested in the point about the need for tools for a multi-decade vision. I agree that there is a long journey ahead to a connected eco-system of digital twins. Who is doing a good job at articulating this multi-decade vision and building tools that are open data/Gemini principles aligned?
-
2 minutes ago, Tom Hughes said:
I’m interested in the point about the need for tools for a multi-decade vision. I agree that there is a long journey ahead to a connected eco-system of digital twins. Whose doing a good job at articulating this multi-decade vision and building tools that are open data/Gemini principles aligned?
Hopefully that is the role of the DT Hub.
Members of the DT Hub include asset owners who will help shape the solution to suit their needs which will need multi-decade persistence to be achieved. Personally it is my hope that a drive towards open data will negate the need for tool-dependency (value is in the nail, not the hammer you used!). However, this’ll depend on what suits the members so that any solution developed will be one that is appropriate.
-
12 minutes ago, Guest Luke O’Rafferty said:
I do think your point on pooling/aggregating information, and how we federate to be the most important. If you have a common language for inputs/outputs, an easy (conceptually at least) use case could be how has Covid-19 impact infrastructure energy usage? I can imagine querying each major infrastructure provider to understand where, when and how much energy is being drawn. I imagine the transport networks draw down will have dropped (significantly?) but there would be an uptick in draw from local substations. How does this balance? What does this tell us about the resilience of the network to major shifts?
Yes! Energy/Transport (and Energy/Housing) have been high on our list of sector crossovers and “linking” use cases for digital twins. I agree with you that the current situation provides a powerful incentive to seriously explore them.
-
4 minutes ago, Guest Luke O’Rafferty said:
This is bearing in mind that Uniclass2015 is 5 years in and still “dynamic”, so even if you had used that to name everything in your power plant 6 months ago, comparable components might not match mine today.
Hi Luke. While Uniclass has been dynamic recent changes have typically expanded the classes (i.e. added new sub-codes) as opposed to altering its structure and taxonomy. In addition, Uniclass has been considered as part of a CIH, BEIS, CDBB, UKBIMA BIM interoperability report:https://www.cdbb.cam.ac.uk/news/bim-interoperability-expert-group-report
The report is currently out for public consultation, feel free to submit your views on its recommendations around Uniclass and other areas of interoperability.
Log in to reply.