The global Digital Twin market size was valued at USD 3.8 billion in 2020 and is projected to reach USD 48.2 billion by 2026. An IoT implementation survey by Gartner in 2019 states that 13% of organisations implementing IoT projects already use digital twins, and another 62% are either in the process of establishing digital twins or plan to do so. Gartner defines Digital twin as ‘dynamic software model’ of a physical thing or system.
What is a Digital Twin?
A Digital Twin, as the name indicates, is the digital version of an entity – this could be a product, component, process or a system. Thus in a digital twin pairing, there are two components of something – the physical thing and its digital version. Connected data acts as the third component and ties the former two. Digital twin is a 3D living model that has the ability to take the virtual representation of all the elements and the dynamics of how its physical real life counterpart operates and lives through its lifecycle.
Consider the analogy of e-books as the digital version of paper books or MP3 as the digital format of music – the theory of digital twin is similar, but also slightly different in that, it uses smart components, such as sensors, joined to physical entities to receive and broadcast data, monitors and optimises performance of those physical devices based on lessons learnt.
Simply put, a digital twin is a virtual replica through a dynamic computer program, that uses real data of the physical thing and creates simulations that can predict anomaly and provide early heads up on how the physical counterpart would perform.
Data scientists can use digital twin run simulations before actual devices are built and deployed. The technology may be used to provide input on a potential product or also to simulate what might happen to a physical version as it is designed as a prototype. As an expanded technology, digital twins have been extended to include large objects such as constructions, factories, and even cities. As a result, such a setup empowers complex physical systems to be intelligently operated with transparency and valuable insights for a better decision making. Lately Digital Twin Technology is increasingly linked to ‘smart manufacturing’ and ‘Industry 4.0” with a focus on automation, connection, IoT and Big Data.
Twin Model Requirements and Challenges
For successful implementation of digital twin technology, the first step is the adoption of digitisation and industries need to strategically equip themselves to embrace digitisation. The initial step in developing a digital twin is to generate 3D geometric drawings and most of the industries are still working on 2D drawings. Creating a digital twin requires sensors, communications networks providing secure and reliable data transfer from physical devices to the digital world and a cloud-based digital platform that serves as a modern data repository pooling and storing the sensor data with high-level business data. By combining these data sources and by using advanced AI algorithms, actionable insights can be derived for data driven decision making.
As explained above, a digital twin model depends on the inputs from thousands of remote sensors and this necessitates managing quality of the data and integration & use of only the relevant data. Conversely, limitations associated with not capturing non-measured data will lead to approximations when automating operations in say, a smart factory. This could lead to downtime, an inefficient workforce and loss in productivity. Another challenge is building a robust engineering and administration process in place to ensure storage and handling of the digital data, so that the digital twin perfectly manages the practical and physical arrangements of its counterpart. Again, the digital twin evolves and accumulates growing historical data, such as geometric models, simulation data and IoT data. As a result, the digital twin owner may run the risk in becoming increasingly locked into the vendor with the authoring tools. Because total life-cycle visibility is a critical feature of the digital twin concept, individual asset (the physical model) data must represent its total life-cycle too. Asset data must include information about production of its various components, not just their use. Obtaining the component data from suppliers, historically considered to be out of the scope of manufacturing usefulness, requires close collaboration with multiple tiers of the supply chain. The opportunity to implement digital twins across different sectors and share knowledge also creates the challenge of creating and using a standard language and frameworks across sectors.
Digital Twin and AI – their relationship?
Digital twins and artificial intelligence draw mutual benefits from each other. AI enables Digital Twin to run analytics in real time or faster, provides a high degree of prediction accuracy and integrates data from a collection of incompatible & distinct sources. Conversely, a digital twin can go through an infinite number of repetitions and scenarios. The simulated data thus produced provides potential real-world conditions to train the AI model.
Let’s understand the relationship from an application perspective – collaborating Digital Twin and AI for planning and construction of future buildings. Data collected on the physical building through individual sensors & control systems is modelled and centralized before transmitting in near real time for predictive analytics by AI. Inputs of previously unseen patterns in data of the structural twin and its simulation of different scenarios of different systems and elements on the virtual twin can now allow remote subject experts to obtain greater insights into the digital twin model. Managers then can evaluate multiple action paths and measure their possibilities and the resulting cost functions in order to make the right option for the next steps. For instance, if a section of HVAC system needs repair- using a digital twin Engineers can, not only find the fault location on their smart devices, but also better troubleshoot the problem using learning curve intelligence captured in the twin. Another example could be a simple deployment of conference room sensors for an entirely optimized environment in one building, which can then be replicated and modelled in other locations within a portfolio for owners to more easily quantify deployment and logistical costs. Thus AI & digital twin will help decision makers visualize how a building works in real-time and enable them optimize their design, construction, and performance over the entire lifecycle – in other words, take Digital Twin driven informed decisions.
References:
Video courtesy: What is Digital Twin? How does it work? geospatialworld.net
https://silo.ai/digital-twin-ai/
https://www.offshore-mag.com/production/article/14185502/offshore-industry-embraces-digital-twin-technology
October 13, 2021, 11:44 pm
I blog quite often and I seriously thank you for your information. Your article has truly
peaked my interest. I will take a note of your website and keep checking for new details about once per week.
I subscribed to your RSS feed as well.
Also visit my web blog :: artikelpendidikan.id