TRICK: An Easy Acronym to Conceptualize Digital Twin Technology

Summary

The device you’re reading this on now is one of the many machines, devices and tools that surround you and support your everyday life–at work and beyond. At Altair, we say our technology is always around you, integrated seamlessly into all these tools, empowering you to be safe, connected and successful in today’s fast-paced world.

How does Altair accomplish this? Increasingly, the answer is digital twin technology, which has an immense impact in business and beyond. What can digital twin accomplish? For starters, the technology:

1. Drives operational efficiency
2. Facilitates a model-based ecosystem, providing access to authoritative data and knowledge for faster, better decision-making
3. Drives superior engineering quality and product development performance
4. Flattens the disruption curve for production

 Digital twin is no longer a hypothetical, “nice to have” technology–it has has fully arrived. Five years ago, few organizations were talking about digital twin technology, let alone creating it. But according to a recent research report Altair published surveying more than 2,000 global respondents, 69% said their organization currently leverages digital twin technology. And of those who said their organization isn’t currently employing digital twin technology, more than half of respondents (58%) said they believed their organization will adopt digital twin technology within the next two years or sooner. This is a game-changing technology very clearly on the path to rapid, widespread adoption.
 
However, in my view, there are still too many organizations that don’t understand what digital twin technology is and how they can harness its power. Luckily, there’s a TRICK to digital twin.

The TRICK acronym

 “TRICK” is a simple, compelling acronym to guide you on your digital twin journey. TRICK is neither sequential nor mutually exclusive–it’s merely a framework to help you distinguish true digital twin technology from fancy-looking interactive CAD models, which are not digital twins but rather mere digital shadows.

T: Think in systems
What defines true digital twin technology? Primarily, its function as a system of systems. Digital twin technology is not merely the sum of its parts, a conglomeration of numerous individual components, processes and systems. Rather, it’s the interaction of all these things acting as one in a holistic environment.
 
Digital twin technology empowers us to better conceptualize and improve interactions between components and systems. As designs continue to become more complex, you find that the most important details dwell in the spaces where systems–not mere components–interact. Engineers and designers should always be thinking at this level to minimize issues and maximize efficiency.

R: Reduced-order modeling
When we want to share insights across different teams and people with varying specialties, we want everything to be as simple and actionable as it can be. Real-life processes are often very complex and hard to distill into something like models or equations. But this is the essence and beauty of mathematics: it allows us to abstract, simplify and standardize real-world knowledge. This is the goal of reduced-order modeling (ROM).
 
For people to scale insights and achieve digital twin breakthroughs, they can’t be inundated in complex, unorganized data. ROM helps reduce the computational complexity of models in physics-based or statistical simulations. True digital twin capabilities are far too complex without the assistance of ROM–it’s essential in any digital twin initiative.
 
There’s an important detail I’d like to add about ROM as well. No matter how you approach it, teams should always be using explainable, transparent solutions, otherwise known as “white box” solutions. White box solutions allow users to see exactly how AI-based models and algorithms are working, including:

• Metadata sources
• What datasets were used
• What underlying equations were used

This approach enhances AI transparency and makes it clear to everyone how models function and why. This is compared to unexplainable “black box” solutions, which offer AI capabilities, but ones that are fully opaque to users.

I: Integrate
Digital twin approaches work best when users of all levels–including those without specialized engineering/data skill sets and non-coders–can leverage advanced tools without needing to rely on data or engineering experts. This requires formalizing and democratizing computational intelligence.
 
With integrated digital twin capabilities, teams can virtualize products, processes and equipment by integrating the best available models informed by sensor and metrology data, guided by domain knowledge, bolstered with software and hardware compute and continuously synchronized with physical counterparts.

C: Convergence is key
Digital twin technology is a living, breathing model of systems that’s grounded in physics assimilating data from inspections, sensors and elsewhere. These systems can be physical, biological, or business process-related. Developing and sustaining a digital twin ecosystem that connects the infrastructure, environment and methodology (tools, methods and processes) requires the convergence of simulation, high-performance computing (HPC), Internet of Things (IoT) and AI. Simply having these technologies in isolation doesn’t constitute a true, integrated digital twin–it merely represents a nice interactive CAD model.

K: Knowledge-driven workflows
As teams progress through development and testing phases, they generate a heap of knowledge at each step. But they’re often losing a lot of knowledge and experience in the spaces between each step because information is trapped in silos. These silos are detrimental–they’re filters that sift valuable nuggets of insight from between each phase until hardly any surplus is left at the end of the development cycle.
 
When I say digital twin means creating “knowledge-driven workflows,” I mean that it can capture information and experiences into a model-based ecosystem that can then be used to develop traceability and governance and improve product quality. A model-based systems engineering workflow breaks horizontal silos by providing a model as a common language of communication.
 
These knowledge-driven workflows reduce design errors once you finally get to the product testing and release stages. There’s no such thing as a universal mistake-proof workflow, but you can certainly minimize errors and the time spent correcting them by harnessing previously lost knowledge.