AspenTech’s Mike Brooks opined recently that ‘the current [digital twin] technology focus is all over the place and leads to many questions’ like ‘what is a digital twin?’, the title of his blog. A digital twin is a virtual model of a process, product or service represented in computer code, logic, equations and algorithms. AspenTech has been in the business of creating digital twins ‘since well before they were ascribed the DT moniker’. They originated as the computer models used to design and simulate chemical processes. No single digital twin can address all scenarios. Brooks advises plant personnel to ‘think about the problem they are trying to solve, and get to the heart of the matter by asking the right question of their twin(s)’. More on AspenTech’s twin philosophy in the white paper, The Digital Twin and the Smart Enterprise.
If we can just push Brook’s reasoning a tad further in the interests of de-mystification, the ‘digital twin’ is simply a rebranding of a the model/simulator.
The March 2021 Issue of the Industrial Internet Consortium’s Journal of Innovation is devoted to the digital twin. The 85-page publication has three sections, on DT and the web, open source/standards for the DT and on an oil industry specific DT from Osprey Data. The latter, titled ‘Design and implementation of a DT for live petroleum production shows how an ensemble of DTs can represent a system of assets and can be used to determine optimal operational set-points. The use case presented is artificial lift where production control is traditionally dependent on ‘manual’ simulation design and expert recommendation. Current oil and gas DTs allow for broader, overall process optimization, but do not support fully automated set-point recommendation. This requires a data processing component integrated with a simulation engine that can manage, process and generate large volumes of data. This is a ‘deficiency’ that Osprey Data sets out to address. The process leverages ‘live data, in-cloud processing power alongside simulation and data science tools’. This is one of the best presentations we have seen covering the nitty-gritty of how a commercial simulator is used to generate training data for an ML model. Tools of the trade include Apache Airflow (scheduling) and Spotify’s Luigi (workflow authoring in directed acyclic graphs). Streaming sensor data is captured in systems such as OpenTSDB or TimescaleDB, purpose-built data stores for store and query of time series sensor data. This kind of infrastructure paves the way towards the next step in the DT, closed loop dynamic set-point optimization. A highly recommended read!
Another section proselytizes for DTs based on ‘open standards and open source software’, authored by ABB, Bosch, Microsoft and SAP, all now apparently converted to ‘open’. The authors note the lack of a common definition of the DT that has led to ‘many flavors’ of the concept, leaving some to wonder if DT is just a buzzword. The IIC and the German Plattform Industrie 4.0 have both provided DT standards. A third set comes from the Industrie 4.0 spin-off, the Digital Twin Consortium (DTC). These cover information models, APIs, connectivity to physical twins, data ingestion mechanisms, security and interoperability. While DT vendors offer proprietary solutions, complex systems of systems are likely to involve components from different vendors. DT interoperability is now desirable. Here ‘multiple new organizations’ have been set-up to address the issue, including the Industrial Digital Twin Association* (IDTA) and the Linux Foundation’s Open Manufacturing Platform (OMP).
The situation in the DT interoperability space reminds us of the title of Evelyn Waugh’s novel ‘Put out More Flags’. It looks like DT interoperability is to be achieved by ‘Putting-out more standards’!
* For more on the IDTA’s Asset Administration Shell (AAS) see Oil ITJ 2020/5.
OneGeology, the BGS-backed initiative of the geological surveys of the world reports in its latest Newsletter the outcome of the first OneGeology digital twin workshop. Representatives of 21 geological surveys explored a long term vision for global geoscience DTs. Presentations covered DTs for river dikes in the Netherlands, nuclear waste storage in France and others. OneGeology has a role to play in ensuring that geology and the subsurface are included in future DTs such as the Deep-time Digital Earth (DDE), EPOS, AuScope, LOOP and Earth Cube. DTs are seen as ‘a positive and challenging framework for our community to engage between our organizations and potentially in a stronger way with our stakeholders’. More from the OneGeology portal and the code base on Github.
In a recent webinar, Esri presented its ‘Digital twin for petroleum’
a.k.a. ArcGIS Velocity (AGV). The cloud-based solution takes real time
data from the field and adds big data and analytics capabilities to
ArcGIS Online. The solution is claimed to be ‘scalable, resilient and
managed by Esri’. Esri’s pitch for digital twin status stems from its
not unreasonable claim that ‘everything has location’ and the
observation that many of today’s operations control centers fail to
leverage the ‘power of location’. AG Velocity is an ‘end to end’ system
for the oilfield leveraging Kubernetes-based scalability and
resilience. The system is controlled and configured through a simple
web interface. AGV talks to multiple IoT platforms – MQTT,
AZURE/AWS/CISCO IoT and more. The big data captured can be pushed on
for analytics, all driven from a web interface. The video demo (see
below) showed a connection to the Microsoft Azure natural resources event hub.
One thing is for sure, the new ArcGIS functionality only received a
brief mention in the ‘software shorts’ section of our last issue.
Raising the stakes to a ‘digital twin for petroleum’ certainly gets
more attention. More from the Esri Petroleum Team.
An 88 (small) page whitepaper, co-authored by Accenture and Dassault Systèmes, found that ‘virtual’ (a.k.a. digital) twins are an ‘untapped opportunity to help companies unlock combined benefits of $1.2 trillion of economic value and 7.5 Gt CO2e emissions reductions by 2030’. The white paper reads like a pitch to access some of the EU taxpayer’s largesse via the ‘Destination Earth’ initiative (see below) that is to use computer modeling to ‘slow or stop global warming’.
Given that what passes as a twin is likely to be a hotch-potch of product-based simulators, interoperability of these would seem like a desirable characteristic. To which aim, the Digital Twin Consortium, a ‘program’ of the Object Management Group has partnered with LF Edge on DT/Edge platform Interoperability.
Esri too has joined the Digital Twin Consortium albeit from a slightly different direction than the DT for Petroleum. Esri’s utilities and GIS for architecture engineering and construction (AEC) unit, along with partners Autodesk and Microsoft are to demonstrate how users can benefit from GIS when developing digital twins.
The EU-backed Destination Earth initiative is to develop a precision
digital model of the Earth to monitor and simulate natural and human
activity, and to develop and test scenarios that would enable more
sustainable development in support of EU environmental policy. The
project is to contribute to the EU Commission’s Green Deal and Digital
Strategy and ‘speed up the green transition and help plan for major
environmental degradation and disasters’. The juggernaut is to be built
on a ‘federated cloud-based modelling and simulation platform’,
providing access to data, advanced computing infrastructure (including
high performance computing), software, AI applications and analytics.
The ‘7-10 year’ project starts in 2021. More on the EU Destination Earth boondoggle.
The Digital Twin Consortium and the Fiware Foundation are joining
forces to accelerate digital twin technology adoption. Fiware is a
not-for-profit organization with 370+ international members including
Atos, NEC and Red Hat. Fiware has defined standard APIs and models for
digital twins in multiple sectors including cities, utilities and
others. More in the press release.
Now for a really serious digital twin, co-developed by Altair and
Gruppo Cimbali, that puts data, simulation and the IoT at the core of a
platform that blends physics and data-driven twins of Gruppo Cimbali
coffee machines. The DTs, powered by Altair Activate, capture information on each espresso/latte served to a database, along with insights about drink quality. More from Altair.
Our last look at the DT was in 2018 when we concluded that digital twin buzzword did not represent anything really new and that building an overarching model of a facility was hard since each component model comes with its ‘own scope, granularity and time frame’. Interoperating models, DTs or whatever they are called has been considered a holy grail of IT for many years, notably with the IEEE’s High Level Architecture. It appears as though not only are there very many DTs, there are also quite a few different interoperability ‘frameworks’. Everyone seems just to be rolling their own amongst these plethoric ‘standards’, interoperability pipe dreams and rebranding. Esri’s ‘digital twin for petroleum’ gets a shout out for astute marketing of a product update. A digital twin? Not really, more of an entreaty to ‘cut the crap, use a map!’
© Oil IT Journal - all rights reserved.