How to use digital twins to reduce risk

I’ve been following news of Microsoft releasing a new version of its Flight Simulator more than my usual tepid interest in computer games. It’s been a good decade since its last Flight Simulator, and I have a special connection with the game as it was the first packaged computer game I’d ever played. 

I was in elementary school, my father had recently purchased a PC clone, and at some point I was quite sick with the flu. My dad came home with a strange package that contained version 1.0 of Microsoft’s Flight Simulator to lift my spirits, and I credit the game with stoking the flames of an emerging interest in technology. I also learned a critical lesson about flying: Being a grade-schooler I assumed that to make an aircraft go up, one would press the up arrow on the keyboard. Based on this flawed assumption, I spent days rolling my simulated Cessna into Lake Michigan, until in desperation, I hit the down arrow instead, pulling back on the virtual stick, and launched my plane into the sky for the first time.

SEE: The tech pro’s guide to video conferencing (free PDF) (TechRepublic)

Using computers to simulate the real world is nothing new in everything from gaming, to media, to social interactions. But these common tools are woefully underutilized in the corporate environment, with most simulations rarely extending beyond the ubiquitous spreadsheet, and that tool quickly reaching the limits of its inherent technology.

What is a digital twin?

In the last few years, the term “digital twin” has entered the lexicon, likely as a result of overzealous consultants applying a complicated name to a simple concept. A digital twin is nothing more than a computer simulation of something in the physical world. The Cessna I careened through the skies of Chicago on my monochrome monitor as a youth was a digital twin, just as a spreadsheet predicting next year’s sales can also be a digital twin, as they both aim to simulate a future outcome using data and logic.

Digital twins are incredibly valuable for the rather obvious reason that they can help you gather key insights and model potential future outcomes at a fairly low cost, thus de-risking larger investments. Consider my early experiments in flying an aircraft. For $40 or so I was able to crash my “digital twin” of a $200,000 aircraft multiple times before gathering the critical insight that I needed to pull back on the stick instead of pushing forward.

SEE: The driving force behind IT salaries in Global Knowledge’s report (free PDF) (TechRepublic)

In a more relevant recent example, I worked with a client who was trying to determine if the logistical costs of a complex distribution network could be sustained at a price customers were willing to pay. This model quickly got outside the capabilities of Excel, so we created a simulation of the distribution network. In the course of a couple of months, we learned that the mix of products flowing through the network was critical to keeping it financially sustainable, an insight we hadn’t anticipated. By building the simulation, and being able to run decades of different permutations of logistical traffic, we avoided spending millions on a network that might fail.

It’s more than just game time

Many technology leaders dismiss simulations and digital twins out of turn, thinking that they’ll be wildly expensive endeavors requiring armies of specialized skills. This can be the case if you require an overly complex model or high degree of certainty. Simulating the weather, a nuclear explosion, or complex interactions at the sub-atomic level certainly requires supercomputers and wild degrees of complexity. However, if you’re designing a new factory, help desk ticket routing scheme, logistical network, or something else that can be distilled into a relatively simple set of rules that operate on a cyclical basis, you can likely execute a simulation that’s simple enough to build, yet provides valuable insights.

SEE: Augmented reality for business: Cheat sheet (free PDF) (TechRepublic)

Consider the board game Monopoly. While it’s not a perfect simulation of the New York real estate market of the 1920s, it does provide valuable insights into economic trends. Key economic activities are distilled into a set of basic rules, and an element of randomization is introduced by the dice. If you automated one of the dozens of computer-based Monopoly games, you’d quickly find that there’s a tendency toward certain economic outcomes.

When considering your own digital twin, start by determining what a “turn” looks like. In most cases, a physical day is a good starting point. Figure out the key activities that occur during that “turn.” For example, if you’re simulating a new production line, there’s potential downtime for maintenance, a set of inputs provided to the various machines, transportation of items between the machines, and some amount of working time for each machine. At the end of the turn, you’ve accumulated some finished goods, spent some inventory, and accrued some costs, all of which are the starting point for the next “turn.”

SEE: What IT teams can learn from Prince Harry and Meghan (TechRepublic)

Even with this level of simplicity, you can simulate what happens when various parts of the production line have extended downtime, modeling the broader impact to the system and using those insights to determine the cost of prevention versus the cost of mitigation. Just as a high-school kid could program a Monopoly game, so can your team, if you can distill the physical process into a turn-based set of rules. The power of the simulation comes from its ability to track long-term performance of the process and run thousands of different scenarios in an afternoon.

Try it in your business to limit some risks

Next time your company is faced with a large investment rife with potential risks, consider the use of a digital twin. It just might be a simple insurance policy that lets you learn and make mistakes in a zero-risk digital world before you begin work in the physical.

Also see