Utility Unlimited

On 13 April 1970, an oxygen tank on board the Apollo-13 spacecraft exploded, seriously jeopardizing the lives of three astronauts. NASA mission controllers tinkered with the 15 training simulators and, despite the zero margin for error, quickly came up with fool-proof procedures that brought back all three astronauts safely to earth [1].

Although the term “digital twin” was not in use back then, this was perhaps its first application. That’s just about how useful even rudimentary digital twins were half a century ago. “Rudimentary” because simulators are not digital twins by themselves. The tinkering to replicate conditions in space brought them close to digital twins [1].

Cutting across to the present, Deloitte utilized digital twins to cut by 20% the changeover time for their manufacturing customers [2]. Boeing improved first-time part quality by 40% using digital twins [2]. GE’s digital wind farm promises to hike electricity generation by up to 20% enabling a single wind turbine to create $100 million more across its lifetime [3].

Digital twins first caught the mainstream in 2017 when technology consulting-research firm Gartner, Inc. named it as among the top ten trends for strategic technology. The firm repeated a similar rating in 2018 [4]. This explains why the global market for digital twins is predicted to jump from $3.1 billion in 2020 to $48.2 billion by 2026 [5].

What are Digital Twins?

Digital twin is a virtual or digital model / representation / copy of a physical asset. The twin [5]:

• Represents the physical asset for its entire lifecycle.
• Receives real time data inputs from the physical asset.
• Employs machine learning, simulation, and reasoning to analyse the data and boost asset performance.  

Asset can be a product, a system, or a process. More specifically, the asset can be a manufacturing process, car, airplane, cargo ship, turbine, engine, city, bridge, building, offshore oil platform, or a human being, among other things.

In short, digital twins are computer programs relying on real world data inputs to forecast asset performance [6]. Internet of Things (IoT) sensors capture this data that is also utilized to predict failure [4] and other problems.

Digital twin could also be based on a physical prototype of the asset, made to refine the performance of the full-scale asset before it is built. Alternatively, the twin itself can serve as the prototype [4].

Unlike simulations, digital twins study multiple processes and receive real time data. This empowers them to run numerous simulations to analyse multiple issues from a number of perspectives [5]. Predictive twins use recorded data from other physical assets [4] as opposed to real time data.

Except component and part twins, others handle data of multiple entities that operate together. Digital twin classification is based on the level they represent [5]:

• Component Twins replicate individual components. Part Twins are those of somewhat less important components.
• Asset Twins represent multiple components that work in association.
• System / Unit Twins model the exchanges between numerous assets.
• Process Twins are the digital copies of two or more systems.

Pros, Cons & Applications

With assets getting more complex, the need for intricate systems to develop, utilize, and maintain them is crucial. Customer expectations too are evolving rapidly and it makes sense to know these beforehand [7].  

Advantages are related to the data they collect and evaluate. Digital twins:

• Position manufacturers ahead of digital disruption by forecasting asset performance and providing data on how customers use assets after purchase. The latter helps understand changing customer expectations [7]. 
• Improve efficiency of manufacturing processes that are already in place for product making [5].
• Eliminate unnecessary functionality, products, or components through analysis of data on how clients use assets after purchase. This saves money and time [7]. 
• Enable superior product design via research and development on the possible performance scenarios before production begins [5].
• Redefine age-old assumptions in product development and process improvement [7].
• Allow better decision making on which process to utilize for products at the end of their lifecycle. Manufacturers can also determine which materials they can gainfully extract from such products [5].
• Simplify maintenance by permitting technicians to check if a certain solution will work on an asset before actually applying it [4].
• Upgrade traceability by connecting separate systems via digital thread [7].
• Fix equipment from remote location [7].

Digital twins are already in use for the following:

• Power Generation Device operators can better schedule maintenance operations for locomotive engines, jet engines, and turbines [5].
• Manufacturing Operations churn out improved products because all stages from design to manufacturing are streamlined. Remember: digital twins represent the entire lifecycle of the product or system [5].
• Automobiles are a natural application for digital twins, being a complex arrangement of several, interconnected systems. Virtual models boost automobile performance and the efficiency of production processes [5].
• Large Structures such as buildings, bridges, offshore drilling platforms and the like which have to stringently adhere to engineering rules [5].
• Health Services employ digital twins to virtually model patients to track health parameter data and analyse the same [5].
• Town Planning makes use of real time, 3- or even 4- D spatial information as well as augmented reality [5]. This enables better prediction of the impact of proposed changes [8]. The fourth dimension is a bit abstract. It is taken to mean time or something that is perpendicular to a cube [9].

Not all assets merit the application of digital twins. As Gartner warns, they could sometimes be a technology overkill. Plus, they open up concerns related to cost, privacy, security, and integration [4].

Considering the involved complexities and costs, digital twins become viable only for large and complex products, processes, and systems such as [5]:

• Systems Engineering
• Manufacturing Processes
• Power Generation & Distribution Equipment
• Aircraft & Automobile Production
• Railcar Design
• Large Constructions

Digital Twins in Cybernetik

It is a standard practice in Cybernetik to use digital twin to check system performance before actual manufacturing followed by design modifications, if necessary. The objective is to:

• Improve system performance.
• Minimize project execution time.

To attain these goals, we employ a rigorous, tried and tested procedure that starts with the creation of a 3D model. This is then uploaded into process simulator software to generate a virtual model. Linking sensor inputs and PLC software to the virtual model gives a digital model. This model is then run at real time parameter values to test its performance. Gathered feedback serves to improve performance.

Finally

Health and capabilities of assets and the performance parameters of processes are critical for all and particularly asset-intensive sectors. With analytical capacity of digital twins expanding as a result of greater resource allocation to improve their cognitive abilities, their role in manufacturing will deepen further [5].

Cybernetik is a 30+ year old company that uses innovative automation to solve challenges of throughput, efficiency, hygiene, and safety. Write to us at [email protected] to know more on how we can be of assistance.

---

References

1. Stephen Fergusson. Siemens. “Apollo 13: The First Digital Twin.” 14 April 2020.
2. Anna Sidyuk. Softeq 25. “Five Examples of Digital Twin Technology in Different Industries.”
3. GE Renewable Energy. “Meet the Digital Wind Farm.”
4. Keith Shaw et al. Network World. “What is a Digital Twin and Why it’s Important to IoT?”
5. IBM. “What is a Digital Twin?”
6. TWI. “What is Digital Twin Technology and How Does it Work.”
7. Maggie Mae Armstrong. IBM. “Cheat Sheet: What is Digital Twin?” 4 December 2020.
8. Adina Solomon. Smart Cities Dive. “Are Digital Twins the Future of Urban Planning.” 1 November 2021.
9. Molly Edmonds. How Stuff Works. “Can our Brains See the Fourth Dimension?”