They enable to virtually map physical products and processes, to monitor and predict their behavior and to optimize their further development: Digital twins open up application scenarios that would not be possible with static documentation and are the key to the benefits of the Industrial Internet of Things (IIoT). A definition.
A digital twin is a virtual model of the state of a product or process. It connects the real with the virtual world. Initially considered as a realistic mathematical model, it reflected the behavior of these products or processes as accurately as possible. Meanwhile, this definition has been extended to include simulated and visible dynamic 3D models. In the automotive sector, for example, of specific vehicles and their components. These are also called “vehicle shadows”.
How does the digital twin work?
Creating a digital twin requires three essential elements: a real object to be represented, a virtual representation space and context-specific information or data on the environmental conditions.
The digital twins use real (time) data from installed sensors that represent, for example, the working conditions or position of the object to be mapped. It is the basis for recognizing patterns in these data with the help of machine learning models, performing extensive analyses and creating simulations.
Complex applications are often based on a modular concept in which the entire digital twin is composed of many individual twins. For example, the digital twin or vehicle shadow of a car consists of individual twins of the engine, body, chassis, tires, and so forth. (In detail, the engine consists of digital twins of individual components as well.) Since this generates large amounts of data from different sources and in different qualities, data repository applications and big data technologies will be used.
In the actual application, we then go deep into the details. During the usage phase, for example, sensors record the data of a vehicle or driver and report it back to the digital twin. Here they are compiled, compared and evaluated using the machine learning model. On this basis, individual and detailed analyses are possible in order to inform the driver at an early stage and, therefore, recommend maintenance intervals or offer and execute individual comfort services without the driver having to take any action.
What are the advantages of the digital twin?
Digital twins and their numerous applications are not limited to the automotive sector. The entire manufacturing industry can benefit from their principle of using machine learning models for data-based development and optimization of products and processes:
Concepts can be validated in advance and processes or products extensively tested in virtual environments. The risk of errors or disturbances in real processes is reduced, quality and efficiency in production and operation are increased. In addition, development and implementation times are shortened, whereas flexibility increases significantly.
An example of production engineering and mechanical engineering: While physical prototypes are usually created late in product development, digital twins can be used throughout the entire life cycle of a product. They enable complex product requirements to be taken into account and implemented without actually creating or modifying prototypes.
An example of a wind farm: If the digital twin of a wind turbine is already in place, machine learning models can recognize patterns in the sensor data. The comparison of data from history and ongoing operation provides important insights into possible weak points and thus enables predictive maintenance.
What happens next with the digital twin?
Gartner names the digital twin as one of the ten most important strategic IT trends of the year. The market researcher estimates that by 2021, half of industrial companies will already be working with virtual avatars, which will increase productivity by up to ten percent.