Imagine that we know an object well when it starts out, and we can measure activity and forces applied to it, and we can simulate how it would react and change. In such a case, we’ve created a “Digital Twin” for our object.
Consider the critical structural elements of an airplane in which actual fatigue is evaluated by removing parts from the aircraft and performing X-ray examinations. What if our Digital Twin for a particular aircraft could accurately predict what we would see without making us remove the item and X-ray it?
Digital Twins have been applied to aircraft for some time, but are now poised to explode in usage thanks to the Internet of Things (IoT) and the availability of enormous compute power to do the simulations. I attended a discussion group on the topic of Digital Twins at a conference earlier this year, and I was struck by the interest level among those who didn’t appear to have a compelling need. Not surprising, though, given that this technology was listed on Gartner’s Top 10 Strategic Technology Trends for 2017.
Gartner predicts billions of things will have Digital Twins in the next few years, in large part thanks to the enormous number of interconnected IoT devices. Digital twins will come to life due to billions of IoT devices, coupled with real-time data feeds and massive compute power in data centers to ponder what the data means. We appear to be on the verge of trusting Digital Twins to model physical reality with high accuracy.
Apollo 13
Whenever the topic of Digital Twins comes up I can’t help but think of the “physical twin” available during the Apollo 13 crisis. This was dramatized in the movie Apollo 13 when the parts “available to the astronauts” were dumped out on a table, with the demand that engineers on the ground find a solution and relay it to the spacecraft before the astronauts ran out of oxygen. That was nearly 50 years ago. NASA has moved on to more simulations than ever to understand their “objects,” which definitely get far beyond our reach as they travel throughout our solar system. You might say they have the ultimate problem of not being able to examine the “real” object and having great interest in what a Digital Twin can tell them.
Summary
Michael Grieves, at the University of Michigan, first wrote of the concept using the Digital Twin terminology more than a decade ago. I’ve provided a link to a piece he wrote, plus links to a number of articles written on this topic with different speculations about the future.
Thinking about Digital Twins requires considering the possibilities that vast data acquisition has when coupled with modeling objects “in the field.” Can we predict failures better? Yes. Can we change usage of the objects to make them more effective? Yes. What additional uses for this concept will emerge? It will be interesting to see!
Additional reading to learn more:
- Digital Twin: Mitigating Unpredictable, Undesirable Emergent Behavior in Complex Systems (Excerpt), by Dr. Michael Grieves and John Vickers (with many citations on related works)
- Gartner:
- General Electric Corp.:
- Forbes article: Digital Twins: The Business Imperative You Might Not Know About
- Oracle: Digital Twins for IoT Applications – A Comprehensive Approach to Implementing IoT Digital Twins
