Cross Paths: Digital Twin

“How would you know the difference between the dream world and the real world?” — Morpheus, The Matrix

Ayyüce Kızrak, Ph.D.
10 min readNov 9, 2022


Machines, internet, and data…

Where our paths cross with digital technologies, expectations are at the peak and this is called the “digital twin”.

In order to take a more holistic perspective on artificial intelligence (AI) applications, let’s take a closer look at the digital twin technology, which has not only increased its popularity in the past five years but has also found application in different sectors.

Image Source: The Matrix Resurrections-WarnerBros

What is it?

A digital twin represents the structure, context, and behavior of a unique one or group of physical objects, creatures, processes, or organizations. It is a software model that allows it to dynamically update, process, and extract insights with data from its physical twin throughout its lifecycle.

In short,

it is a virtual representation of everything that happens in the “real” world. *

What is not?

The digital twin is not a simulation. A simulation is just a starting point. This initial phase is static and requires parameter updating. These updates are essential to adapt to real physical conditions.

However, the digital twin is not independent of the outside world. The most important feature distinguishing the digital twin from the simulation is its real-time, continuous, and dynamic.

When did it begin?

Although the concept of the digital twin has been in the literature for about 20 years, it has found a place in our lives with the development of different repulsive technologies. With the widespread use of Internet of Everything (IoE) technology, the ability to collect faster and higher quality data has been gained. As a result, the digital twin has been appearing more and more in the last five years.

Why now?

Digital twin; AI was less popular 20 years ago than it is today due to the need to develop Machine Learning (ML), IoE, blockchain, cloud computing, Augmented Reality (AR), and communication technologies.

While today’s world is so data-oriented, an ecosystem wants to take advantage of the opportunity in design, customization/personalization, production, and operation stages by adding a new dimension to the data. In addition, The technology ecosystem is now more mature and ready. It is a harmonious, efficient, real-time, dynamic, and risk-reducing approach with a high economic impact in many sectors, including healthcare.

How does it work?

Although the definition or description is somewhat reminiscent of science fiction novels, if we go to the basics, it is a part of applied mathematics. A substantial data collection process is required to create a complete and individual representation of the real-world entity and its components. Once the data is collected, it is transferred to the virtual environment in real-time and continuously. While the model in the virtual environment continues life just like in the real world, it can make predictions about the processes in the real world.

The exciting thing is that this scenario could be for anything. A similar procedure may work for the body of living beings, the design of a machine, the life of an airplane’s engine, the operations of a city, the services of a government, or procurement processes.

The digital twin is a process developed and tested with feedback. So it requires working with a prototype in development to see what will happen with the physical release.

Representation of the Digital Twin Concept

What are the types of it?

Although the digital twin concept is basically the same, its purpose and scope may vary according to its real-world counterpart. In this context, there are four types of digital twins, from narrower to broader coverage.

  • Component twins: Represent a part of a system. For example, The gearing of a rocket engine requires design for specific stress parameters. Designers and engineers can use a digital twin for this procedure.
  • Asset twins: They can be thought of as virtual representations of the product itself, not part of the physical product. An example would be the rocket engine with the gear in the model in the component twin. Another example; a wind turbine is a twin from which the operation of this asset can be monitored.
  • System twins: Represent assets/products working together. It is a whole structure made up of components and assets. For example, It can be considered a rocket with the engine itself or a manufacturing robot.
  • Process twins: Digital representations of systems working together. For example, Modeling of the production line used to manufacture the rocket is performed with a process twin.

More than 42% of executives across the industry verticals understand the benefits of digital twin technology, and 59% plan to incorporate it into their operations by 2028. — BusinessWire

The digital twin approach lays the foundation for data integration at different levels, as illustrated in the figure below.

  1. Solving how to integrate the main components of the architecture in the smart solutions vertical;
  2. Supporting the integration of different systems within an intelligent organization with a system of systems approach;
  3. Establishing the primary “common language” that the systems connected to the organizations speak and understand for the Data Space concept allows secure data sharing between organizations.

Data types and providers must be standardized to support these three levels’ effective integration.

Levels of Integration supported following a Digital Twin approach

What are the benefits?

When thinking about the use cases of the digital twin, production will come first mentioned. The most important reason is that there are environments/factories where regular and high-quality data are collected through sensors for a long time, and processes are standardized with specific procedures. As can be expected, Industries that have met most of the requirements will adapt faster to the digital twin technology. Today, essential applications of the digital twin can be seen in different and dynamic environments, from aviation to health.

The digital twin’s workflow and relationship to other technologies

Manufacturing is a sector that always needs to reduce costs and optimize processes. Using a digital twin to improve performance at many stages, from the R&D process to logistics, is possible. It assists engineers and designers in all phases of a product, from design to testing and even customizing the product. The digital twin can also be used to analyze the strength and lifespan of the product ready for production.

Unilever tested the technologies in a single facility and saved $2.8 million by reducing energy consumption and increasing productivity.


75% of Air Force executives stated that they trust the use of digital twins. — BusinessWire

There is no room for error in this industry. Therefore, it is necessary to choose the most precise calculation techniques. Digital twins are also used in the automotive industry to create virtual machine models.

For example, developing/modeling a car starts with a clay model, but then automotive engineers use Siemens NX CAD to turn it into an actual product. With this product, a digital twin environment is provided.

It allows virtual design before starting to develop the car’s parts. The difference between this tool from other design tools is; It is also helpful for modeling and analyzing production stages and problems that may occur when the car is on the road. For sure, digital twins have also become a vital test tool in the autonomous vehicle industry.


It is the second most common use of digital twin technology in supply chains and logistics. By virtualizing product packaging and checking the process for errors, they can predict the performance of packaging materials. Carbon footprint can be minimized with optimizations that can be made in the supply chain.

It is possible to optimize operations in procurement processes and predict errors based on historical data. In addition, optimizing warehouse designs can increase productivity by simultaneously evaluating many parameters such as inventory, logistics network, and road conditions.

With this perspective, DHL has implemented a digital twin project for warehouse design.

Digital twin technology helps retailers minimize capital expenditure by 10%, reduce excess inventory by 5% and improve EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) by 1–3%. — Boston Consulting Group

French supermarket chain Intermarché has created a digital twin of its stores using data from the shelves using IoT technology. Digital twins allow store managers to manage inventory and develop different store strategies.


Digital twin-empowered solutions in smart cities will reach $4.8 billion by 2027. — BusinessWire

Combining digital twin and AI technologies is important in designing agile and sustainable smart cities. Singapore and Shanghai are two notable examples of fully digital twin smart cities. In cities redesigned with this vision, determining energy needs, waste management, emergency management, parking, and charging problems are no longer unsolvable. Improvements in energy consumption, traffic flow, and city planning have been implemented. Smart cities are also a topic studied in digital twin Türkiye to reduce pollution and increase the well-being of citizens.


It is estimated that energy consumption in buildings can be reduced by 30% to 80% by using digital twin technologies. — Accenture

One of the goals of the digital twin today is to help design green and sustainable buildings and their associated businesses. It allows for simulating hotel workflows and guest circulation and providing personalized services to each guest. There are important examples where infrastructure resources and human needs are planned more effectively in living spaces.


Every data-oriented technology is making radical transformations in the health sector. It benefits people and the environment in healthcare processes and in numerous areas such as bioengineering and drug discovery.

Digital twins have also begun to be used to develop personalized patient care and genome-coded drugs/treatments.

The fact that most data is sensitive and personal raises ethical debates in the digital twin, just as in other destructive technologies. In this field, it is essential to follow ethical codes, as we have seen in AI.

Digital Twin and Artificial Intelligence

It is not difficult to think of digital twin technology and AI together. If a material such as data can create a perfect twin, any application that can be done with AI, such as predictive maintenance, classification, and regeneration, can be done in virtual space, as we do in the real world. Moreover, this environment will contribute to the development of AI itself.

For this, in 2018, Microsoft launched Microsoft Azure Digital Twin, a product that allows users to create digital twins with various data collected from IoT devices and analyze these models using AI algorithms.

Some digital twin and AI solution products and applications (and some open source):

Where to Start to Develop an AI and Digital Twin Project?

What if we want to develop a digital twin project successfully:

  • Comprehensive, diverse, systematic, and quality data
  • A well-thought-out use case and goal
  • An acceptable set of rules
  • Ethical, human, and environmental-oriented integration of AI
  • Human resources prone to new technologies
  • Technology investments
  • Legislation and ecosystem readiness/maturity

What awaits us?

By 2023, approximately 75% of digital twins are expected to use five different types of integration endpoints for Original Equipment Manufacturers (OEMs) connected to the IoT. — Gartner

  • The data collected at the endpoints/edge will increase drastically. These increased data points will lead to the need to update some security protocols.
  • Data governance will become more critical than ever.
  • Inevitably, the need for applied training to provide expert human resources in data science and engineering will increase.
  • Project managers who are familiar with ethical values and principles by design will be sought.

A similar future awaits us in digital twins and in all technologies with developing administrative, technical and legal regulations. With a common sense approach, we can design a more powerful future.

🏃‍♀️ 🏃 Feel free to follow me on LinkedIn, Twitter, and GitHub!



Ayyüce Kızrak, Ph.D.

AI Specialist @Digital Transformation Office, Presidency of the Republic of Türkiye | Academics @Bahçeşehir University |