Bernat Morato
Innovation Technologies Expert at NTT DATA
In a world where the intersection between the physical and the digital becomes increasingly prominent, digital twins emerge as a powerful tool to capture the essence of this convergence. The seed of this technology was planted by David Gelernter in 1991 when he introduced the concept in his work Mirror Worlds. Since then, digital twins have evolved from being a futuristic vision to a tangible reality that is transforming the way we interact with our physical environment.
In this article, we will explore in detail this powerful technology, the different types that exist, their diverse applications, the technology and protocols behind them, and finally, the challenges and opportunities they present.
Deciphering Digital Twins
Digital twins are virtual replicas designed to accurately reflect physical objects, processes, or systems.
This concept comes to life with the democratization of microcontrollers and digital manufacturing, which has allowed for the massive collection of data on real processes and facilities. When these data are integrated and represented in a three-dimensional digital environment, they create an enriched virtual version of our physical world, known as a digital twin.
This technology has transcended imagination limits to become a fundamental tool in a wide range of sectors. Furthermore, it requires a greater or lesser degree of initial investment depending on detail level, quantity, and quality of data to be consumed, as well as the process context itself and the visualization engines needed to be used.
As a general rule, a digital twin can be integrated with a wide range of visualization technologies, from a multi-screen system to a variety of immersive technologies such as Augmented Reality (AR), Virtual Reality (VR), or Mixed Reality (MR), with which it creates a striking synergy (if you still have doubts about the difference between these technologies, delve into a detailed article here).
This versatility stems from the cloud-connected nature of digital twin systems, meaning that data is stored and processed in the cloud, enabling visualization to be tailored to the specific needs of each user or application.
Now that we’ve explored the foundation and versatility of digital twins, it’s crucial to understand that not all are created equal. Next, we’ll delve into the different types and their specific applications, from visualization to real-time control, and much more.
Diversity of Digital Twins
Classifying the various digital twins can be complex due to the multiple perspectives from which this concept is approached: whether by its content, its development in relation to the physical product, its level of interaction, among others. However, these classifications can lead to ambiguities and difficulties in identifying the specific model required by collaborators or users.
Frequently, the first question that arises regarding digital twins is: What are they for? and consequently, How can they be applied in a particular sector? Therefore, a useful approach to distinguish the different types is to adopt a functional approach. This classification methodology presents self-descriptive categories independent of the technology used, making it easier to understand their purpose by simply reading their name.
This functional approach not only allows for describing the most innovative and unconventional functionalities of digital twins but also eliminates exclusivity among different types, making them a more accessible set of possibilities. Next, we will explore the diversity of types that can be found.
Visualization Digital Twin
This type is the most common and focuses on providing a visual and graphical representation of a physical system or process. It allows users to observe and understand the state and behavior of the system in real-time or over time, whether through numerical observations, colors, movement, and other visual indicators.
A practical example could be a surveillance system used in a multi-level parking garage. Imagine an employee is responsible for monitoring the parking garage and has access to a screen displaying a three-dimensional (3D) model of it, detailing each of its 500 parking spaces. In this model, each parking space is clearly represented, indicating whether it is occupied or vacant in real-time.
Therefore, with just a glance at the screen, the employee can obtain instant information about the occupancy status of all parking spaces in the garage. This simplifies parking management by allowing quick and efficient monitoring of available and occupied space, without the need for constant physical inspections.
This example illustrates how a visualization digital twin facilitates the understanding and monitoring of complex physical systems through intuitive and real-time visual representations, serving as the foundation for the following.
Control Digital Twin
This type is designed to enable interaction and direct control over the physical system or process represented by the digital twin. Users can make adjustments and decisions based on the information provided by the digital model.
In this case, the employee responsible for the parking garage in the previous example can control the raising and lowering of barriers and shutters on all levels, and also lock and unlock certain reserved spaces using their digital twin software.
Diagnosis or Maintenance Digital Twin
This type is focused on identifying problems, anomalies, or failures in the physical system represented by the digital twin. It uses historical or real-time data to detect and diagnose issues, facilitating decision-making for problem resolution. This type can also make use of AI and Deep Learning algorithms for predictive maintenance analysis.
Continuing with the previous example, the parking manager is able to monitor the water pressure in the fire protection systems, the air quality on each floor, the status of ventilation systems, and much more. Even the software is capable of providing alerts when the ventilation system shows any signs of imperceptible deterioration, thanks to the sensors it incorporates and the predictive maintenance algorithm in the digital twin application.
Simulation Digital Twin
This type is used to simulate the future or hypothetical behavior of a physical system or process. It allows users to explore different scenarios, forecast outcomes, and assess the impact of changes before implementing them in the real world. This type may also include optimization and prediction functions, typically using AI and Deep Learning algorithms.
In this case, if maximizing parking efficiency and avoiding queues is desired, the implementation of a simulation system can be crucial. The software can simulate various scenarios of peak hours, storage strategies, and changes in lane directions, among others. The data from each simulation compared will provide key insights (longer wait times, slower parking, etc.) to add future improvements to the hypothetical parking setup.
Metrics Analysis Digital Twin
This is one of the most cutting-edge types and is focused on the collection, analysis, and interpretation of metrics, in addition to providing data related to the physical system or process represented by the digital twin.
It’s ideal for use in fleet management, personnel control, or production environments. It focuses on continuous monitoring of key variables, pattern identification, and generating insights to improve performance and decision-making. Furthermore, it can integrate advanced data analysis techniques such as data mining, time series analysis, and AI to extract relevant insights from the collected data.
Continuing with the same example, let’s say a storage and distribution area is added to the parking lot, rented out to other companies for their operations. To ensure at all times that this operation is functioning properly or profitable, a sophisticated, state-of-the-art digital twin software proceeds to cross-reference the data from all previous examples. This includes the presence and location of people, vehicles, and goods, among others, in such a way that it extracts useful business metrics. This level of virtualization provides enough data to maintain a series of real-time KPIs and thus perform analysis and simulations in much greater depth and with data correlated from multiple sources.
Operations Digital Twin
This type is used in the most ambitious use cases to visualize, coordinate, and execute operational procedures in real-time, including multiple sources and types of data (people, vehicles, facilities, etc.).
While it is one of the most complex and expensive to execute, it is especially useful in large-scale constructions, mining explorations, defense operations, and other types of environments where the dataset is critical at all times and a high degree of control is required.
As an example, let’s consider the construction of the parking lot, which is carried out with state-of-the-art technology. In this case, the digital twin would provide an exceptional level of control, precision, and security in construction. The system would consist of a complete and real-time situational map 24/7, including many elements such as the location of people, vehicles, machinery, transportation routes, ground metrology and equipment, supply chain status, safety checklists, AI-driven alerts and control actions, communication transcriptions, air toxicology measurements, operators’ vital signs, and many more elements depending on the type of operation being performed at any given moment.
In summary, after understanding how a functional approach can clarify the purpose and applicability of Digital Twins in different sectors, we will explore in detail the platforms and protocols that support their implementation and operation.
What lies behind a Digital Twin
All the feature sets mentioned earlier are enabled by various IoT-focused protocols and platforms, among which MQTT, HTTP/S, CoAP, and OPC UA stand out due to their extensive support communities and widespread implementation.
However, some proprietary developments from major companies have been instrumental in driving these technologies at scale by creating complex integration and management tools. Examples include IBM, Cisco Kinetic IoT, Siemens Advante, Azure Digital Twins, or AWS TwinMaker. Additionally, various 3D rendering engines like Unity and Unreal Engine, which are becoming increasingly accessible, complete the picture by providing an accessible and efficient development platform.
Challenges and opportunities
Despite the promising future of digital twins, they are not without risks. Among them are the high costs of specialized hardware, the diversity of types of processed data (such as positions, states, values, signals, among others), the wide variety of vendors and protocols, and the inherent complexity of large-scale systems.
However, the magnitude and versatility of these systems provide the opportunity to establish long-term evolutionary projects that can be improved and refined as they develop, adding new layers of information or new ways of interpreting existing data. Thus, a project that makes use of these technologies can become an invaluable resource for strategic decision-making, process optimization, and continuous innovation across a wide range of industries, from manufacturing to urban infrastructure management.
To conclude, it is worth emphasizing that the digital twin market niche has been growing steadily for some years, and there is no indication that it will decrease in the near future, as it is still in the discovery phase. Moreover, application predictions for this technology revolve around energy generation equipment, structures and their systems, manufacturing operations, healthcare services, the automotive industry, and urban planning. However, they will also have an impact on business methodologies and strategic decision-making. In the coming years, we will see an increase in demand for these developments, driven and challenged in equal measure by recent advancements and the democratization of AI in simulation and prediction environments.
Digital twin systems not only represent a powerful technological tool but also a paradigm shift in how we conceive, design, and operate complex physical systems in the digital era.
To delve deeper into other extended reality trends that will shape the future, click here.
