What Are Digital Twins? Digital Twins Simply Explained And Benefits
What are digital twins?
The “Digital Twins” concept is characterized by the fact that data is not collected for the general public on entities (e.g. “How are our vehicles used?”), But from the creation process (e.g. simulation, production) through logistics to use (e.g. usage data ) are recorded and assigned for each instance.
A digital twin is an individual digital representation of a physical object that is as detailed as possible. A lot can serve as an object: people, machines, vehicles or products.
The goal of digital twins is on the one hand to create traceability of the development of entities (e.g. “Why is the stove already defective after 2 months?”), On the other hand, of course, to interact directly with these entities and their individual history (e.g. “This So far, the heating element has always been used at maximum load, we have to shorten the maintenance cycle ”).
Simply put, digital twins are the collection and tracking of all data that arise in the life cycle of a product, person, or other entity. It is important not to transmit, record, and use the data unilaterally but bilaterally.
The entity (product, service, machine, etc.) transmits information and receives data to use it.
Examples of different types of digital twins
DIGITAL PRODUCT TWIN: EVERY PHYSICAL PRODUCT HAS ITS TWIN
A digital twin is handy for technical products. Whether the drill or the stove, the mobile phone or the heating system: All of these things produce data and can, in turn, be used more efficiently through individual data analysis.
The data acquisition of the digital twin starts at least in the production, often earlier in the development, then goes through quality control, logistics, trade to the end consumer, and their usage behavior.
All data sources must be coherently integrated, cross-linked, and brought together to represent the physical product’s digital image correctly.
DIGITAL TWINS IN MANUFACTURING: ONE MACHINE, LOTS OF DATA
One area of application for digital twins that many expect to have a high impact is manufacturing. In particular, large production machines often consist of a long history of simulations, components, partial repairs, uses, and much more.
If this data is combined with movement data (e.g., which products are manufactured, when, and how maintenance was carried out), this allows a 360 ° view of the machine and its current status.
This makes it possible to directly increase the efficiency of production: predictive maintenance, utilization plans, tracking of reasons for downtime, product quality control, and much more.
DIGITAL CUSTOMER TWIN: THE CUSTOMER AS A DIGITAL TWIN
Another interesting use is the holistic view of a customer and the combination of all contact points that this person has with the company.
Whether it’s orders, newsletter interactions, service contact, and other data sources: the consolidated history of a customer, including master data and its changes, allows a high degree of personalization (e.g., customer churn prediction ) and is the basis for a high degree of customer focus.
Today, customer data is mostly indifferent systems – or is not recorded in a structured manner.
DIGITAL SERVICES TWIN: HOW SERVICES ARE REPRESENTED DIGITALLY
But not only physical things such as machines, products, and people be represented by digital twins, but also services. From the creation process to pricing, from sales to use, from participants to ratings – services can also be significantly upgraded through comprehensive, linked data management.
This makes it possible to continuously improve its portfolio, manage capacity utilization and establish links with other twins such as the customer or product.
What are the benefits of digital twins?
Now that the principle has been presented, we would like to discuss a few advantages of digital twins:
- Data consolidation: Since all associated data is recorded for each digital twin, a kind of mini data lake is created technically and from a data governance point of view.
- Recording of the entire life cycle: The digital twin records usage data and from the beginning to the end of the life cycle.
- Process documentation and optimization: By building a digital twin, the underlying processes can be derived, and process mining can be carried out.
- High granularity: Individually recorded data also has a very high granularity, allowing more use cases than aggregated data.
- Use for data science: By consolidating all data belonging to an entity, you have a large field of application for advanced analytics, machine learning, and more.
Differentiation between digital twins and other prominent concepts
The idea of collecting data on products and people is not new per se. Of course, there have been master data systems for years that record information about entities and communicate more contemporary concepts such as individual things (IoT).
In the following, we would like to briefly discuss the positioning of the digital twins in some central technical pictures.
WHAT IS THE DIFFERENCE BETWEEN A MASTER DATA SYSTEM (E.G., MDM, PIM, CRM, PLM, PDM) AND DIGITAL TWINS
A master data system records the data on products (e.g., Product Information Management, PIM) for each customer (e.g., Customer Relationship Management, CRM) or on the development process of a machine or service (e.g., Product Lifecycle Management, PLM). Therefore, the idea of digital twins is very close to this type of system.
However, there are three significant differentiation criteria between digital twins and master data systems: MDM, PIM, CRM, PLM, or PDM.
- Master data systems often only record the “prototype,” i.e., the generic product (e.g., PIM), not every individual entity.
- Master data systems have rarely integrated all data sources that exist for an entity type (e.g., CRM rarely has data from service, web analytics, hotline, webshop, etc. integrated)
- Master data systems follow a specific operational process. They do not form the basis for a centralized, individual, complete data collection to use further this data ( advanced analytics, data science, machine learning, etc.)
As a result, depending on the master data system, there may already be a high overlap of data that can also be used for the digital twin. However, the goal of the digital twins is that this data is consolidated, recorded individually, and for further processing, which is often not congruent with master data systems.
WHAT IS THE DIFFERENCE BETWEEN THE INTERNET OF THINGS (IOT) AND DIGITAL TWINS
The Internet of Things (IoT) is a concept where every technical product individually sends and receives data from making predictions using artificial intelligence. This hyper-personalized data of an end device (edge device) is, of course, precious in the concept of digital twins.
However, there are two differences between IoT and digital twins: On the one hand, the Internet of Things focuses on the communication of devices in use; on the other hand, IoT does not cover historical data such as development or material used.
As a result, the Internet covers the middle part of the life cycle but does not consider all data sources (e.g., service requests, manufacturing data) and advanced information.
WHAT IS THE DIFFERENCE BETWEEN BIG DATA ANALYTICS / EVENT STREAMING AND DIGITAL TWINS
Big data analytics, especially in connection with event streaming, is another concept that is very closely related to digital twins. In general, however, it is only one application of the data volumes recorded in this way: If data is transmitted via event streaming, it can be analyzed directly; but the assignment, tracking, and long-term analysis in the context of the individual device is the extension of the concept through digital twins.
Example architecture of a digital twin
A digital twin can be built in many ways. In general, there are many providers of such software, but on the other hand, the concept can also be implemented yourself. In general, the architecture of a digital twin consists of three levels: the physical layer, the twin layer, and the utilization layer.
The physical plane is the natural manifestation of the entity. Whether human or machine: In any case, the physical level produces data over its entire life cycle. This data then has to be measured, recorded, and consolidated in the twin layer.
The twin layer is the heart of the entire architecture. It consists of several database systems for structured and unstructured data as well as extensive documentation. Likewise, an access block is usually inserted to make the digital twin available for the utilization layer.
All aspects that use the digital twin are summarized as the utilization layer, whether optimization within production, the use of artificial intelligence, or the implementation of individual software in the physical layer – all procedures that extract and use the data of the twin are based here.
Together, all three layers allow the physical representation to be transferred to the digital twin and used in various applications. Only when all three levels merge will it be possible to proceed efficiently when using the added value that has arisen.
The Relevance of Digital Twins in the Data-Driven Company
As a logical combination of data lakes and the Internet of Things, digital twins are one of the newer concepts in the world of data. They combine a range of technological and conceptual advances: individual data, consolidated and documented as the basis for use cases in data science, automation, and optimization.
As a result, the digital twins are a direction every data-driven company should behead. In the best-case scenario, the digital twins are based on other areas such as the data lake and are therefore low in their implementation but instead require an expansion of the scope.
Once digital twins are established, a range of applications opens up. From hyper-personalization to process optimization, the data transfer works in both directions. It focuses on what the data-driven company should stand for: the generation of added value through data.