Part 3 of the 4 part series: Digital Twins through the full lifecycle

How to Sustain an Operational 3D Digital Twin

"The operational lifecycle of a 3D Digital Twin must incorporate ongoing reality capture in order to sustain its accuracy and fidelity"

While reality capture is a key requirement for sustaining a 3D Digital Twin system, it must be broadened to incorporate three important features (outlined below) to ensure this critical operational tool can live on and support the operation through its long, useful life.

 

  1. Continuous Detection / Continuous Integration (CI/CD) management of change: Learning from best-in-breed software engineering techniques, sustainment requires a structured, human-in-the-loop, automated process that monitors and detects changes in the physical operation as they occur, including, but not limited to digital reality capture.

  2. Granular, modular version control and governance: Model version control and governance over historical, work-in-progress, and updated versions of the model, with rich, searchable visual metadata templates and catalogs.

  3. Always-on orchestration and synchronization: A way to synchronize and orchestrate access to models as they are being maintained while maintaining 24×7 live access to the source truth model of record. This is a key requirement for it to be “Operational”.

Let’s dig into each of these.

 

Continuous Detection / Continuous Integration management of change process

 
Over the past 15 years tremendous benefits have been realized from agile software engineering methods. Key among these developments have been continuous development / continuous integration (CI/CD), which has allowed for the unprecedented acceleration of bringing functions to market. 3D Digital Twin Systems are a collection of data and software that can leverage that same agile philosophy. A robust, CI/CD-enabled model management of change should:
 
  • Use best available techniques to detect and log changes: whether your organization has the latest and greatest in reality capture / change detection (see sidebar), or relies on simple scratchpad note taking, change detection should be visually and contextually integrated into the model.

  • Schedule new visual content capture: whether new visual content is available, or must be manually gathered, the tasks should be integrated to follow scheduled and managed workflow.

  • Apply AI and rule-based logic: Newly captured visual content must be taken through the process in a structured, way, with the ability for AI-based automation to convert input to model structures, and augment models.

  • Human in the loop: AI-based methods must be backed up by human judgement, particularly in critical or sensitive areas. Probabilistic evaluation and reinforcement learning allow systems to alert when changes are simple and highly probably accurate or when human review is needed. Rules and thresholds supported by configurable policies are critical.

Granular, modular version control and governance

 

3D Digital Twins are a rich, multi-dimensional reflection of physical reality. The information they encapsulate is not monolithic—that is—the model reflects many individual components, and systems of systems, each of which can be separately involved in changes and differing states of completion.

For example, you may be planning a small construction or modification of the plant and implementing infrastructure changes incrementally. Or you may simply want to reflect the actions being planned in a turn-around or new setup separately from what is “live.” Critically, each individual component must be maintainable and traceable separately, without impacting the use of the model for daily operations. Each component must be versioned separately, with the visibility of who and how it was changed. That is granular, modular version control.

Governance is the process of defining and structuring how model changes can be made, who has the right or privilege to make those changes, and how that authority is managed and enforced. The model owner and stakeholders must be able to define a customized process that suits its organizational security and other institutional needs.

That customized process ought to be software defined – a configuration, rather than reprogramming or redesigning the model software. Configuration-based governance allows some organizations (and some parts of the model) to be managed with a “light touch” and others to be more heavily governed, as required by regulation, management policy, or other concerns, such as safety. Finally, the model, its versions, and its change history should be visible and auditable.

Always-on orchestration and synchronization workflow

Often overlooked in the sustaining process for a 3D Digital Twin is the ability to manage between the work in process changes that may be taking place as parts of the plant and its digital twin are modified. Orchestration and synchronization of the live model goes hand in hand with the design requirement for modularity and granularity. No Operations Leader wants the activity of updating the model to render it useless during this activity. Therefore, a structured orchestration and synchronization process, integrated with the change management workflow as defined, and the ability to do small, incremental changes at a fine grain make managing change robust, low effort, and non-disruptive.

What operators should take away

Managing change and incorporating it into the model for the useful life of the physical asset is critical capability of Digital Twins, particularly 3D Digital Twins. Making the sustainment easy, and not cost prohibitive, operators should work with a 3D Digital Twin that:

  • Establish a 3D Digital Twin as your System of Record for visual information during the operational life of the asset/system.

  • Work with a 3D Digital Twin partner who has a robust Model Management of Change that can be adapted to your unique operational, regulatory and management needs.

  • Don’t underestimate the up-front effort to define. governance for your 3D Digital Twin, just as other systems of record require. Software tools that automate detection and integration tasks will minimize the cost and effort.

  • Work with a 3D Digital Twin partner who have experience with the creation and ongoing, lifelong maintenance of the model over time.