My Digital Twins: from Simple Life Simulation to Development through Evolution. Part 1

Reads: 50  | Likes: 0  | Shelves: 0  | Comments: 0

  • Facebook
  • Twitter
  • Reddit
  • Pinterest
  • Invite

Status: In Progress  |  Genre: Non-Fiction  |  House: Booksie Classic

I plan to tell you a tale that describes my own more than 50 years of experience of creating and using digital twins in engineering, starting with a simple digital twin of sandwich panel and finishing with the artificial world of digital twins that are evolving via natural selection. Ready? Here are you go. Part one presents my simple digital twins of the 70s, 80s, and 90s. Part 2 is devoted to digital twins' development through evolution.

1. Introduction

 

If you read My Computers story (Booksie, Slava Zarubin) then you know that the words digitizing, digitization, and even digitalization are not strange for me, and I would add, for most of other people, you included :) because all our being today is related to digital formats starting from our morning conversation with Google Home, kind of “Hi Google, switch off the alarm clock, tell me the last news, please, and, by the way, what's the weather today” and finishing with the Netflix movies before going to bed. We got used to this, even to Tesla, the digitized box on the wheels steered by artificial intelligence. But during the last ten years, one more digitalized term the “digital twin” started to be heard often, more often, and very often now from different sides of our digital environment. I noticed that those, who are speaking on this subject frequently, present the digital twin idea in an emotional way similar to if they are opening our eyes to the new unprecedented phenomenon. But is this phenomenon really new and unprecedented? The answer is “Of course not.” And my story is about this. I plan to tell you a tale that describes my own more than 50 years of experience of creating and using digital twins in engineering, starting with a simple digital twin of sandwich panel and finishing with the artificial world of digital twins that are evolving via natural selection. Ready? Here are you go. Part one presents my simple digital twins of the 70s, 80s, and 90s. Part 2 is devoted to digital twins' development through evolution.

 

2. Simple Virtual Life of Sandwich Panel

 

Let's start with this:

Fig. 1. Sandwich Panel

 

My first digital twin, created as a student research work in the Kuibyshev Aviation Institute (KuAI), existed inside of exotic Promin-M computer of 1970. The physical original is shown in Fig. 1. You can build and test the panel like this. If the panel fails the test, you have to make suggestions about the reasons why it failed and build a new reinforced one to test it again. If it does not fail, you can make it lighter with thinner faces or core and continue to test. As you see, building different specimens and conducting natural tests in the search of the best panel is an expensive and time-consuming process. But if you can describe the behavior of the panel with formulas that predict the panel's reaction to the applied pressure by calculating displacements and stresses, estimating safety margin, then you may, instead of a natural test, make a digital one: create a digital panel with faces and core, load it with digital pressure, calculate stresses and compare them with allowable. If stresses are higher than allowable, then the panel fails, if not, make it lighter and repeat calculations. That was exactly what I was doing with my model, which is, as we know now, a digital twin or digital replica of the real sandwich panel.

This is a good example and an explanation of what digital twins are and the answer to the question “why do we need them?” As you see, the answer is simple: digital specimens are cheap, much cheaper than a real object, and we can test and break them virtually many times before finding what we need and only then after all these digital exercises, build the real one that will satisfy all imposed demands.


 

3. My Digital Twins of the 70s

 

Those years being a KuAI ONIL-13 researcher I was working by contracts with aviation original equipment manufacturers (aviation design bureaus) and was involved in a kind of digital revolution when finite element models changed the design process and even organizational structure of design offices. Such models, as an example of fig. 2, combining all aggregates and assemblies of the aircraft as one and whole entity, gave designers of different departments (wing, fuselage, empennage, etc.) the possibility to communicate, make their own impacts on the model, have access to the impacts made by others, and design an effective structure.

Fig. 2. Finite Element Model of Wide-Body Passenger Liner

 

Initially, in the sixties, aircraft designers didn't like finite element simulation as it was kind of their time-wasting, but they loved it finally when the understanding came that there are no other ways to solve complicated structural problems, for example, to find an effective wing-fuselage interface without such models and direct communication between wing and fuselage departments via these models. As a result in the 70s detailed models were created at the early stages of design, much ahead of drawings, which gave a chance to test the prototype digitally before manufacturing it in flesh. The model lived a long life in parallel to its metal sibling as a digital twin, providing results of numerical tests, helping to arrange and tune real natural tests, planning repairs, and modifications. Even the decision of decommissioning was based on the fatigue analysis where the digital twin is the source of information.
 

4. My Digital Twins of the 80s

 

Computerization of all engineering disciplines, aerodynamics, in particular, led to the ability for aeroelastic problems analysis, which is very important in aircraft design. The aircraft digital twins additionally to elastic properties obtained aerodynamic, inertia, and landing gear ones. Even runways were described numerically for a taxi, takeoff, and landing simulation. The digital twin of a regional turboprop is shown on the fig. 3 is undergoing natural oscillations. The digital twin combined geometric, elastic, inertia, aerodynamic, landing gear, control system, and other models now, fig. 4.

Fig. 3. Natural Oscillations of the Regional Turboprop Liner Presented by Digital Twin

Fig. 4. Models Combined by the Turboprop's Digital Twin

 

Information created by one discipline, by a specific model, is used in another one and vice versa. It is obvious that for effective cooperation of different disciplines at different design stages an integrated database is needed. This database has to provide designers with the possibility to make his/her own design solution in the project and has an access to solutions made by others participants. The database and CAD/CAE applications connect organizational departments into one information net breaking down the department’s fences and hierarchy, making the management structure flattened. Such database, proposed and created in the 80s, was a prototype of modern PLM (Product Lifecycle Management) systems, and our database similar to the PLM system lived a long parallel life as a part of a digital twin together with real aircraft. 

 

Artificial intelligence prototypes in the form of expert systems were created in that days too. For example, the expert system, created in 1986 supported the following decision-making stages of structural optimization: 

  1. Initial setting of algorithm parameters and their changing during the optimization process with the purpose of algorithm efficiency raising.

  2. Settle the situation when the algorithm cannot find a feasible solution.

  3. Make a heuristic optimization step in design variable space.

 

The scheme of the software used with a digital twin in the form of a set of aircraft models of fig. 4 is shown in fig. 5.

Fig. 5. Digital Twin Environment as a Set of Integrated Applications

 

course, car designers and manufacturers followed trends in engineering and finite element analysis with its digital twins came to the car industry too, see fig. 6.

Fig. 6. Car body digital twins created in the 70s-80s

 

The models were created at the earlier stage of design, used in detailed optimization, manufacturing of prototypes, their tests, and preparation of mass production.

The Perestroika started in 1985 and led to changes in the Soviet economy and even provided new opportunities for small businesses. But the economy and the state itself collapsed at the end of the 80-s and our customers from the aerospace and automotive industries became bankrupts.

 

5. My Digital Twins of the 90s

 

After collapsing of the Soviet economy the USSR collapsed too leaving its people to survive in their own ways without any state support. Most of the engineers lost their jobs and started small commerce operations of the “buy cheap – sell a bit expensive” type. But I managed to keep a dozen of young engineers in a private engineering company and even got sufficient investments for this business from a businessman who was impressed by our ability to create and use digital twins. We started our marketing campaign and entered absolutely new for us markets where, in comparison with aerospace and automotive ones, some life was still warm. Several examples are shown below.

 

5.1. Oil Industry

 

Digital Twin of Field Pipeline

The first analysis task was got from Surgut. They had a problem with field pipelines as shown in fig. 7.

Fig. 7. Digital Twin of the Oil Field Pipeline

 

The task was to estimate and predict the fatigue life for the pipeline that oscillates in the stream and has flaws in the form of pipe thinning due to wall wearing by the sand pumped with the oil.

Digital Twin of Oil Tank

The fragments of the oil tank designed by Samara Reservoir Company for a tank farm in Norway are shown below. We got this task because the design and manufacturing were done in Samara by our client, but the tank had to be assembled. installed and insured in Norway. The requirement for finite element analysis came from the insurance company, Lloyd in this particular case.

Fig. 8. Fragment of the Oil Tank Digital Twin, the Wall, and Branch Pipe Flange

Fig. 9. Fragment of the Oil Tank Digital Twin, the Dome

 

The model was created, analysis done, the report presented to Lloyd, the tank built and insured, and the model lived a long life together with the tank being ready to answer any structural questions, related to strength and fatigue.

 

 

5.2. Natural Gas Industry

 

Digital Twin of the Compressor Station Piping System

Fig. 10. Fragments of the Piping System Digital Twin

 

Samara Transgaz got a big headache when started the replacement of the old type of gas pipeline compressors for new ones. The interface between pump and piping, particularly the branch is shown in detail in fig. 10, which got a crack after several hours of work with a new machine. Our model and analysis helped us to understand the problem and find a solution. The geometry of the branch pipe was changed per our calculations, a new one manufactured, installed instead of the previous branch, and the problem disappeared.

Very soon after finishing the analysis presented above, we got a new problem, now from VNIIGAZ that was situated in Razvilka, Moscow. The digital twin, created for the problem solution, is shown in fig. 11.

Fig. 11. Digital Twin of Compressor Station with 5 GPA-16 Ural Pumps

 

The problem was unusual, very big oscillations of the 1.2 m diameter pipe happened on the overpass in a situation when one of the pumps (the last one on the right side of fig. 11) was idle. It was scary to watch how a huge pipe began to wriggle along the overpass. Our model helped to reveal the reason for such behavior. The branch of the idle pump worked as a whistle, the knee oscillated and pulled the pipe. A new effective damper system was needed and our digital twin helped to find it. The report was written and presented to VNIIGAZ. They read the report, discussed all details during the presentation, and told us that they knew all this stuff by themselves. Yes, after one month of our answering their questions and showing them the way of how to damp the oscillations, they announced: “We know all this without you” and didn't pay us. It was not a big surprise for me as we had even more severe experience with semi-criminal state structures. Business in Russia was dangerous at that time :) VNIIGAZ was deleted from our marketing list as a criminal organization and we never contacted them after this.

 

5.3. Chemistry Industry

 

Heat Exchanger

Samara Azot, the big chemical manufacturer, used huge Japanese heat exchangers which after several years in use started to be leaky and needed repair. The engineering staff of Samara Azot repaired by replacing old leaky pipes and welded new ones inside the cylindrical structure. During the thermo-treatment of the exchanger's body, the pipes inside it constrained cylinder extension and it was broken. Yes, the cylinder with a thick (6 cm of Japanese steel) wall failed. What to do? They had our telephones and addressed this problem to us.

Fig. 12. Digital Twin of Japanese Heat Exchanger

 

The model of the heat exchanger was created and the problem was visualized with a good graphical explanation of the reasons why the structure failed. Many numerical experiments were done with this digital twin to find the best process of heat treatment, reliable and safe. Samara Azot was very happy with the results and several more exchangers were successfully repaired. But they had no cash to pay us and instead of money we got two hundred plastic containers. I think that many Russians who lived in Russia in the 90s still remember the word barter. Yes, it was now our problem to convert plastic containers into money somehow via exchange in kind. Our marketing guy, Andrey, was very resourceful, and found the way to do this :)
 

5.4. Power Engineering

The Washer of Hydro-Power Turbine Blade

The Zhiguly Hydroelectric Station named after Lenin had a big rehabilitation at beginning of the nineties and during this event, they decided to solve their eternal problem of oil leaking through the turbine's blade washer. Yes, several tons of oil were going to the Volga river via the structure, presented in fig. 13.

Fig. 13. The scheme of a structure that should protect the oil leakage

 

The digital twin of the washer was created, fig. 14, analysis done, the reasons of leakage found and the variant of repair proposed.

Fig. 14. Digital Twin of the Washer

 

It was interesting, that the power station authority didn't have the cash to pay us too and the barter scheme was used again. This time we got gasoline in the amount of three standard Zil-130 tankers. And it was not a difficult task for Andrey to convert this gasoline into cash :)

There were many more digital twins and related R&D works for us in the 90s. I do not want to be boring with more pictures on the subject, just to tell you that civil engineering, particularly foundation-ground interactions problems were solved for several unusual high-rises on the Volga River bank, ski lift cable stations, automotive problems of car wheel fatigue, suspension dynamics, lots of other metal and composite structures, that were still designed and built during these dashing 90s. But one task, very special for me, has to be mentioned here.

 

5.5. Mir-Shuttle Docking Module

 

It happened that I spent May of 1994 in Washington DC having business training in the framework of the Gore-Chernomyrdin program and at the beginning of June moved to Houston TX, where had an internship in Tenneco Company with the goal to get some knowledge and experience of work in the natural gas industry. And I got some, see my picture in fig. 15 where I am shown during my visit to a fully automated offshore platform in the Gulf of Mexico. But my solid enough aerospace background was with me too and being in Houston I got aerospace contacts of course. One of them with legendary Angelo Miele from Rice University who is well-known in the area of applied aerodynamics, optimization, and numerical methods. He liked my presentation of what was done by our team in the area of structural analysis and optimization and presented me to NASA, particularly to NASA's Lockheed Affiliation, situated on the NASA 1 Highway, see me there in the picture of fig. 16.

Fig. 15. At the automated gas pumping platform in the Gulf of Mexico

Fig. 16. Lockheed Affiliation on the NASA Road 1 Highway

 

Visiting Lockheed and NASA, making presentations there, I managed to hook the actual task of dynamic analysis for the Mir Orbital Station – Shuttle docking module in the process of the forced landing of the Shuttle with this heavy cargo on board. The fact is that the docking module was designed and built by the Energy Space Corporation situated in Podlipki, Russia, and it was supposed to be put into orbit by the Shuttle. In short, the drawings of this module were given to me at Lockheed's branch. After negotiations the order for such analysis was received, model, fig. 17, created, analyzed, and in October 1994 I presented the results to NASA. We got very accurate data that were used later by Lockheed's full-scale research.

Fig. 17. Digital Twin of Mir-Shuttle Docking Module

 

It was an interesting internship. And there were some out of curriculum activities too. We had high-level official receptions and dined both in the Senate and in the Russian Embassy (vodka in the Embassy of Russia was poured by huge black guys in white gloves, which was really impressive). I worked in Washington DC and Houston, visited San Francisco Bay Area, stayed in Monterey and Carmel CA, sailed from Lake Charles LA to the gas pumping platform in the Gulf of Mexico, flew by chopper above Texas, made pics of alligators in Louisiana, met lots of new friends.

There were much more in the 90s for me. Yes, we got a contact with developers of CAD system worked on Analysis, Optimization and partly on PDM with expert system features. Most of this is mirrored in my full doctorate dissertation (you know, that in Russia they still have Ph.D. as Candidate of Science degree and full doctorate as Doctor of Science one) and in these two publications:

 

The first one contains a description the optimization tools:

, Structural Optimization, Vol.8, No. 2/3, Springer Verlag International, 1994

 

The second one is about integrated database and expert system:

, Structural Optimization, Vol.9, No. 1, Springer Verlag International, 1995

 

I know that there may be a lot of critics related to my explanation and my examples of digital twins presented above. Therefore at the end of this section, I would like to repeat again and stress that a digital twin is not a replica of an existing physical object and its information dossier. The digital twin is a multidisciplinary set of models that are created and optimized before the real object is built. The twin contains an integrated database that is a tool for information management and its storage at all stages of the object life, starting with the idea and finishing with retirement. The database plays the role of digital threat that provides a holistic view of the object's data through its lifecycle. AI tools in the form of expert systems are an important part of the interface between human beings and digital twins. And to finish with the 90s I would like to stress what was obvious in that time: an artificial object designed and manufactured as a high tech product is a replica of its optimized (all imposed demands and constraints satisfied) digital twin, which is a set of multidisciplinary models, their integrated database, knowledge base for software control and smart interface with end-users.

 

6. Conclusion of the Part 1: My Digital Twin Definition

 

There are six distinctive features of my digital twin presented and discussed in Part 1. They are:

 

  1. The set of multidisciplinary models is a key element of the digital twin.

  2. The set of functional features formulated in the form of requirements describes goals, demands, and constraints that have to be satisfied via optimization at every stage of digital twin life.

  3. The optimization tool that is used to meet requirements at all stages of digital twin life: design, experimental production, natural tests, fine-tuning for mass production, production, repairs, modifications, decommissioning.

  4. An integrated database related to models and analysis/optimization results is an information source for PLM.

  5. Knowledge base and expert system that provides a smart interface between digital twin and end-users.

  6. The physical twin is a replica of the digital twin at all stages of physical twin life.

 

In Part 2 I will introduce you to the digital world of digital twins where they can evolve via natural selection. You will see how the embryo of a digital twin is developing its skeleton, muscles, nervous and other systems gets training in accordance with the requirements of item 2 from the definition above, competes for the place in the mating pool, meet the mate, crossbreeding, the birth of new offspring, the mutation is probable too. It will be a fun reading :)


Submitted: May 25, 2021

© Copyright 2021 slava zarubin. All rights reserved.

  • Facebook
  • Twitter
  • Reddit
  • Pinterest
  • Invite

Add Your Comments:


Facebook Comments

More Non-Fiction Business Writings