Conference on digital twins

Ron Kenett writes:

This conference and the special issue that follows might be of interest to (some) of your blog readers.

Here’s what it says there:

The concept of digital twins is based on a combination of physical models that describe the machine’s behavior and its deterioration processes over time with analytics capabilities that enable lessons to be learned, decision making and model improvement. The physical models can include the control model, the load model, the erosion model, the crack development model and more while the analytics model which is based on experimental data and operational data from the field.

I don’t fully follow this, but it sounds related to all the engineering workflow things that we like to talk about.

6 thoughts on “Conference on digital twins

    • I follow, or think I follow, this part: “a combination of physical models that describe the machine’s behavior . . . with . . . the analytics model which is based on experimental data and operational data from the field.” I like the idea of trying to align physical and statistical models.

      • So “digital twins” are the in thing in engineering right now. Bit on the high side of the hype curve. Tastes like cold fusion and flying cars.

        But here’s how I understand it. Traditional engineering models used to only model the core behaviour of equipment. Say you had a chemical reactor model the reaction kinetics and heat trasfer.

        The Digital Twin camp wants models that capture EVERYTHING. So sitting on your comfy armchair the model should tell you how fast the metal is corroding, who you buy spare nuts and bolts from, when did you last give it a wash.

        Pretty much a Grant Unified Model.

        The idea is tantalizing, but we are far from it in the real world. There’s some niche flagship projects.

        Color me skeptical for now.

        • Doesn’t sound right to me.

          Take the West Seattle Bridge. Yeah, we have plans for the bridge, but who knows what was actually built? No one, but we are sure We also have a set of varying measurements from inspections. We have records of repairs, cracks noticed by annual inspections, etc..

          So we should start with a model for the bridge and update it with the data as it comes in. That’s our “digital twin” of the bridge. And one can project the model forward in time more quickly that projecting the real thing forward in time, so we know it’s doomed to fall and we have to debate how to replace it for a year. Oh, and close it, please, asap.

          I suppose the air traffic controller system is a digital twin of the actual planes in flight, and the JPL is in the business of maintaining digital twins of space probes.

          30 lashes for who dubbed it a “twin.”

        • I guess the analogy is that if you built a NEW bridge you would feed everything about it into the digital model. Eg all the drawings, test data, part info, vendor info etc.

          In principle then if you have a very good model then ANYTHING you want to know about the bridge you should be able to get from the model. So maybe you also factored in the rate at which the bridge rusts so you can plan your painting and replacement from the model.

          Almost anything. Within reason.

          In that sense the model is a twin since anything you want to know about the bridge you could get it from the model.

          Again, that’s me paraphrasing the idea.

          Personally, I think a lot of it is hype for now.

  1. Bridges need to be maintained. They now include sensors https://www.mdpi.com/journal/sensors/special_issues/Bridge

    Statistics and analytics is used in monitoring (identifying changes over time), diagnostics (picking up what caused a change) and prognostics (assessing the impact of a change). To derive these capabilities, sensors can feed a digital twin model that is used for monitoring, diagnostics and prognostics. These are not off line simulations but simulations fed by on line sensor data and incorporating physics based models.

    Next time you drive your car, think about the same thing. The fixed schedule maintenance schedules do not make sense any more. Some people need to have their car serviced in shorter periods than the ones listed in the manual, some in longer periods. The sensors you have in the car, and analytic models, can help achieve that. This is much more than the indicator lighting up when your tires are low in pressure.

    An now imagine what can be done with health related data. And educational data on students learning experience in a classroom. All these can be personalized, with data and analytic modeling. The other side of the coin is that it also requires ethical and privacy considerations…

Leave a Reply to Andrew Cancel reply

Your email address will not be published. Required fields are marked *