Maturing Your Digital Twin

Many of us struggle with the implementation of a Digital Twin. It seems unimaginably difficult to pull together everything we know about every physical product instance. How can we ever track down all these objects and interrogate them in the wild? The answer is to start with an immature Digital Twin and nurture it to grow into that wonderfully rich and complex Digital Twin we desire over time. Yes, we always eat the elephant the same way – one bite at a time.

Simple Definition

To me, the definition of a Digital Twin is the use of a collection of electrons to give insight into the behavior of a collection of atoms. Any such representation in the digital realm that gives insights into the physical realm is a Digital Twin. Complexity or breadth is not a constraint, only that it is complex enough to give insights to solve real world problems.

Humble Beginnings

A Digital Twin can start from humble beginnings. A very simple Digital Twin might start with a 3D CAD model and maybe a FEA simulation of stress or heat. This is a very low maturity Digital Twin, but a Digital Twin none the less. Information about the physical product is available to us without the bother of building a physical product.

Smaller Companies have Smaller Needs

For many smaller organizations this might be all the maturity they need. If we assume Lean Techniques apply to Digital, then more is not always better. The organization needs to assess its performance and decide if more maturity is warranted and in which direction to mature. These are difficult questions, but they can be answered. The best place to start is usually to gather data on business performance or, if you are not collecting any, to start collecting it.

Directions to Mature

To mature your Digital Twin, you can better model the physics, add data on what was actually manufactured, add data on changes in the field, add data on usage, add data on environment, add data on actual failures, or add introspection. Each of these are key areas of maturity.

Improved Physics

We can expand our exploration of the physics of the part in many ways. We can expand our reach from static to dynamic. We can explore how the part reacts when it is part of the larger system of the product. We can explore how the part reacts more realistically as the part tolerances are adjusted from ideal to a more realistic view of the tolerances we can achieve in the factory.

In fact, we can actually simulate not just the part, or the system, but simulate the factory as well. And why stop there? We can simulate the product working in the real world. We can make the Digital Twin very robust just by looking at the theoretical physics implied by our design choices.

Exploring all Solutions

And why stop with just our final design choice. We can model and simulate solutions we did not chose. We can start second guessing ourselves and exploring the design solution space to see if we missed something. These “what if” solutions don’t even need to be feasible – we can search for areas where we do not have capability today, but if we did, the innovation would be a game changer. This is fertile ground for R&D efforts.

What was Actually Manufactured?

Moving beyond simulation we can start to have our Digital Twin collect data on what really happened as we realized our product as individual product instances in the factory.  We can add data on what was manufactured and how the tolerance distributions came out. This data can be used to better inform our physics simulations. It can also be used to anticipate issues as the products are put into the field and prepare the field service group to deal with them.

Service and Maintenance in the Field

For many complex products we can repair them as they are used, thereby increasing their useful life. This also changes the product instance to make it more unique than it was originally. This data can be added to the Digital Twin and used to anticipate future repairs. The failure mode data can be used to further inform our simulations.

We can also use this data to develop upgrades and service kits to anticipate future failures or prevent them. And, of course, we will need to track the application of the upgrades and service kits as well.

Real World Usage

For very complex and expensive products we can have them “phone home” and tell us what they are doing. This usage data can be added to the Digital Twin to expand our knowledge of not only what the product is doing, but what people are doing to the product.

Many times, users of the product dream up new usage scenarios that were never anticipated when the product was designed. We can either use this information to make the next generation of the product more robust, or we can rush out a service package to make sure the product can survive the unexpected usage pattern.

Real World Environment

With a little additional effort, sensors can be added to the product so that when it “phones home” it can inform us of the environment it is operating in along with the usage data. This is a valuable addition to the Digital Twin because it gives us additional insights into why the product is behaving as it is.

We all make assumptions about where our product is going to be used and we are occasionally surprised when our clever users find new places to use the product. Is it being used in the arctic, or in a mud bog in the Amazon? Is there a huge dust cloud or salt spray we need to navigate? These insights can provide the next level of understanding without us trudging out to the usage site to see what is going on.

What Good is all this Data

The Digital Twin is just an expensive pile of electrons if you cannot gain insights into your Product and Product Development Process. A classic mistake in maturing the Digital Twin to keep adding data and not trying to gain insights into how to improve things. Then the Digital Twin becomes the Digital Money Pit and impedes your ability to move forward. There are many ways you can wring value from all this data you have collected. The key is to focus on your current problems and what data could help you address them.

Key Focus Areas

As you focus on Maturing your Digital Thread it should be envisioned as a problem-solving tool and not a status symbol. The data you gather should be what you need to solve problems. The bigger the problem, the more urgent the need for data. Also, be aware, your data needs may change over time. As your Product and Product Development Process mature, your Digital Twin will need to evolve to keep up with them.

Priorities

The priority of implementation will probably follow the path outlined above. In general, the path we followed was from the easiest to the hardest to implement and maintain. That is another good point: You not only need to create the Digital Twin, but you also need to maintain it over time. The data will need to be curated and in many cases Quality Assurance will need to be implemented to ensure the data is correct.

Drivers for Digital Twin Strategy

As you develop your Digital Twin strategy you will be focused on solving business problems. You will also be making tradeoffs between ease of collection and value to the business. You should also keep in mind that your data needs will grow over time and collecting some seemingly less important, but easy to collect, data now may make it available in the future when solving some new issue requires it.

And the Digital Thread

The Digital Twin needs to support your business needs and be cost effective and flexible enough to change as the business changes. All this movement of data around the Digital Twin is your Digital Thread, so you will need to invest in that as well.

Conclusion

Have you had a good experience maturing your Digital Thread? Do you have any tips of tricks to create it and keep it growing to support your business? We would love to hear your insights in the comments.

Digital Guideposts is written by Mark Pendergast – retired Data Junkie, Deep Thinker and Innovator. He worked with product data for over 30 years of his 41-year career in Automotive Components Manufacturing. His background includes work in Engineering, Operations and Information Technology. He is also an Electrical and Computer Engineer (BS-ECE) and a Certified Project Management Professional (PMP). In his spare time, he mentors a High School FIRST Robotics Team, reads and plays on his computer.