When I was a child, I was jealous of twins. I thought how cool it would be to have a twin. I had twin cousins a few years older than I was that I saw on a frequent basis. I always got them confused. Which one was Roger, and which one was Ronny?

A buzzword floating around recently is "digital twin". Sounds cool. What is it? According to a webinar by Hexagon I listened to today, a digital twin is a "virtual representation of a physical facility".

It is easy for me to envision how this applies to CAD or traditional drawings. Whether 3-D, pseudo-3-D (an isometric), or in diagrammatic form (e.g., a P&ID or PFD drawing), I can see how these can be digital twins.

Let's dive into what this does or does not mean for simulation software like AFT. And whether simulation software is an identical twin or a fraternal (non-identical) one. Or just maybe, (gasp) a triplet!

The webinar I watched today was very interesting and was led off by Michael Fry, whose company (Deepwater Subsea) was involved in the blowout preventer that failed on the Deepwater Horizon. He provided a compelling discussion based on his experience with the Deepwater Horizon failure investigation. Here is a link if you are interested in the webinar: https://connect.hexagonppm.com/digitalTwin_recording_pilot_episode.

In the early days of AFT we got involved with a customer developing a transient model of a blowout preventer using AFT Impulse. Most of what I know about blowout preventers came from that project.

I will not get into all the details of the digital twin webinar, but here are some key things I took note of that relate to simulation software:

  1. Ability to simulate the system in question
  2. Support for not only current design and operations, but past design and operations as well
  3. The ability to check simulation vs. operations past and present
  4. Provide "a single source of truth"


When a buzzword gets out into the wild like "digital twin", you can pretty much guarantee that everyone is going to talk about how they support that buzzword. It is a good way to get, well, "buzz".

So, let's talk about pipe flow simulation and digital twins. I have read with amusement other pipe flow software companies talking about how they provide a "digital twin". Really? Do they?


1. Ability to simulate the system in question

Pipe flow comes in many forms. Fluid phase (liquid, gas, multiphase). Steady vs. transient. Single component fluid vs. multicomponent. Isothermal vs. adiabatic vs. heat transfer. Newtonian vs. non-Newtonian viscosity behavior. A true "digital twin" should be capable of simulating all of these that are relevant for the pipe system being modeled.

A pipe flow software that only models steady flow is not a digital twin at all. What happens when a pump trips off? And a check valve slams? I learned 30 years ago that this causes waterhammer, which can cause all kinds of bad things to happen including pipe catastrophic failure. See this link for a discussion of waterhammer.

If a pipe flow software cannot model waterhammer, is it a digital twin?

If a system has heat loss and gain in the system and piping, and the pipe flow software cannot model this, can it be considered a digital twin?

If a system has a positive displacement device which causes pulsation and vibration, which can lead to problems such as damaged supports and fatigue failure, and the pipe flow software cannot model pulsation, can it be considered a digital twin?

If a pipe system is handling gas or steam, which is compressible, and the pipe flow software does a poor job of simulating compressible pipe flow, and cannot handle heat transfer effects and sonic choking, can it be considered a digital twin?

I think on all the examples above we can agree the said pipe flow software is not a digital twin at all. At least it is not an identical twin, as it does not accurately reflect the physical behaviors of the system being modeled. At best it reflects some of the physical behaviors of the system. Which would make it a non-identical, fraternal twin.

But if the simulation is not an identical twin, is it a digital twin? No, it is not. What is the use of a non-identical digital twin? I can tell you the answer. It is useful as a buzzword for simulation software developers who prefer to ignore their limitations. For engineers dealing with real systems, pipe flow simulation software, even with the limitations I have listed, is useful. But don't fool yourself that it is a digital twin.

The bottom line is that AFT software, which can accurately model waterhammer, pulsation, heat transfer and highly compressible flow, is closer to being a digital twin to the real system than (most) others who cannot do any of these things.


2. Support for not only current design and operations, but past design and operations as well

Pipe systems evolve, their design changes, pumps are added, or new ones installed, same thing with valves, etc. And they change operations (flowrates, pressures, temperatures, tank levels, control setpoints, etc.). AFT software has been used on systems over a century old. Very cool!

What happens to the digital twin pipe flow model when the above things happen? For most, it means creating a new model file to hold the model changes.

Once you do that, you have two digital twins. Wait, two digital twins? And both fraternal (non-identical) to boot. That makes them non-identical triplets, not identical digital twins. Is that not likely to confuse people?

What you would want is a way to keep all the design and operational changes in the same model file. And with the changes linked together within the model. AFT has perfected this capability over the last 20 years and we call it the Scenario Manager. With all changes in a single model file, then you can still have a digital twin. You can find out more about the powerful Scenario Manager feature in AFT software here.


3. The ability to check simulation vs. operations past and present

How does one check the simulation vs. operations past and present? Well, one way to do it is to send both to Excel and compare them there. AFT software can import from and export to Excel. The Excel export is all automatic and can come from different model file scenarios. The Excel import can come into different scenarios in the same model file. Read more about it here.

A true digital twin would live in an ecosystem where these comparisons could be performed and, even better, performed automatically. AFT software can do that. Not all other pipe flow simulators can.


4. Provide "a single source of truth"

This is a hard one for a digital twin. Here are some things that come to mind.

First, like the first point, can the software accurately and completely model the system? If not, it is not a single source of truth.

What about software updates? Does the software developer provide regular updates to resolve problems and make improvements? I know of one pipe flow software developer who provides an updated version once a year! Yes, really. Nothing in between.

Also, how accurately is the data in say, CAD systems, reflected in the pipe flow model? AFT software can import data files from numerous formats including CAESAR II Neutral files, Piping Component Files, GIS files, and EPANET files. See Piping Layout Wizard here. If you can build your model from a previous source of truth (a CAD or pipe stress file), you have a better chance that your model is providing the same source of truth. If not, then you may now have a second source of truth. And that would be, again, a non-identical digital triplet. Not good.

In summary, providing a true digital twin is hard. Simulation software companies claiming to provide a digital twin are usually providing something much less – while jumping on the buzzword bandwagon. They are providing something that sort of looks like the physical system in certain situations.

The breadth of tools and capabilities offered by AFT is frankly unmatched by other software providers. We are much closer to providing a true digital twin than lesser software companies who spend more time on buzzwords and less on creating software tools and capabilities.