Marc Westphal, Bayer
In a nutshell: We are right in the middle of it! ... The digital transformation is an omnipresent topic for us. At Bayer, the digital transformation is a top priority that has our top management’s full attention. It is being addressed by several cross-divisional initiatives. Let me mention our CORE project as a global example. This is a multi-year, business-driven, IT-supported strategic project for the transformation of our end2end processes. Its goal is to increase customer value. When we say 'customer', we are referring to both external and internal customers.

Our individual divisions – in this case, Crop Science –, have set their own priorities for the digitalization projects. However, they have a common set of goals: To create transparency, optimize processes and avoid waste. At Bayer’s Dormagen site, there are currently 18 projects explicitly focused on digitalization. They relate to the areas of production, technology, logistics and procurement.

In the area of technical asset management at that site, we are currently working on building digital twins that are digital replicas of our technical equipment and individual devices. They incorporate all related aspects and data flows across the entire asset lifecycle. The scope extends from equipment planning and design to operation start-up and maintenance, and through to decommissioning and final disassembly. In this context, we are challenged by the fact that there are no standards that have been implemented across the industry. After years of hearing about “Industry 4.0”, respectively, the fourth industrial revolution, the Internet of Things (IoT) and the Asset Administration Shell (AAS), there is still plenty of room for development.

This is evident in the varied approaches taken by manufacturers, service providers and ourselves as operators. Related interest groups and consortiums such as NAMUR, the Digital Data Chain Consortium (DDCC) or the Industrial Digital Twin Association (IDTA), provide meaningful support to help standardize processes, data models and tools.

Marc Westphal, Bayer
A digital twin is a comprehensive digital representation of an identifiable asset. It is an entity that not only represents the current state of the asset but can also store historical data, for example data covering the entire lifecycle of the asset. It may also reflect links to digital twins of other assets.
Even though the term "digital twin" was first defined some 20 years ago and became a hot business topic five years ago, we are only now able to begin reaping the potential of this technology. Open platforms and industry standards such as the Asset Administration Shell, or AAS, support these efforts. They are actually prerequisites for implementing digital twins.

Unfortunately, the definition of an asset administration shell is technically complex, scaring off some companies and getting in the way of industrial digitalization.
There are two approaches that can help simplify this.
The theoretical approach defines the digital twin on the basis of specific maturity levels during the product lifecycle. It differentiates between how an asset was designed ("as engineered"), how it was manufactured ("as built"), and finally, how it is being maintained ("as maintained"). At each maturity level, new or additional information becomes available that is digitally linked to the asset. If we apply all this to safety valves made by LESER, it would mean that for every safety valve delivered to us as the customer, a digital twin would be available at the lowest maturity level, the "as built” level, which comprises manufacturing bill of materials information (MBOM) as well as information about spare and wear parts. When LESER provides us with maintenance information, the maturity level will advance to "as maintained".

What is actually of greater interest to the industry as a more pragmatic approach would be a version of the digital twin based on its possible implementation, i.e. as an Asset Twin, Product Twin or Line Twin. 
This approach places greater emphasis on the application aspect and the interlinking (or collaboration) of the digital twin with various "partners". An Asset Twin, also referred to as a Master Data or Information Twin, reflects the initial and most basic composition of the asset. It is a purely digital representation of the entity as manufactured and supplied. This implementation, along with other information collected from master data, documents, CAD documents and bills of material, enables new application scenarios. Examples include the digital nameplate or ID link.

The next development step, called Product Twin, opens up opportunities for predictive maintenance when combined with additional IoT data (from sensors and controllers). The highest level, or Line Twin, is interlinked with all digital twins of a process line, enabling the simulation of entire processes. The implementation of a digital twin we are currently addressing is the Asset Twin. Together with LESER, we have collected and reconciled all information needed for a digital representation of safety valve master data. In this process we apply industry standards such as ECLASS and VDI2770. This work has prepared the basis for using an information platform to generate and use future Digital Product Passports (DPP).

Marc Westphal, Bayer
As I mentioned earlier, digital twins are a key aspect of every digitalization project. Merging all relevant information (characteristics, status information) covering the entire asset lifecycle is impossible without a digital twin. This is why there is great potential for digital twins. To realize this potential, digital twins need to be used collaboratively between the manufacturer, the operator and the service provider. This requires introducing unified data standards such as the AAS mentioned earlier. But other aspects of data sovereignty must be standardized too.

Once there is a clearly structured process for the interaction between all the stakeholders, it will open up many possibilities for additional services along the asset lifecycle. Even options for creating semantic networks based on available information are no longer utopian. With the digital twin as a basis, the use of AI can be advanced efficiently.

Marc Westphal, Bayer
Regrettably, many things we are talking about today are still mere theory.
What is the precise definition of a Digital Twin when we are talking to our partners? What is the basis of this definition? Is the digital twin interoperable? How can the required data sovereignty be made possible?
Plus, there are system- and data-related challenges inherent in IT systems, in particular information-sharing platforms needed for collaboration.

To tackle these challenges, we have identified the following key aspects we need to address:
Interacting with all stakeholders (manufacturer, operator, service provider, platform operator) as partners and cultivating an open dialog. We cannot hope to arrive at an effective and sustainable solution unless everybody has the same understanding of the boundary conditions and requirements.
Sharing information in committees (such as the ones mentioned earlier) will support the development of solutions that can be applied on a broad basis. 
After all, including users in the development process, especially with regard to potential new services and E2E processes, is crucial. Use cases and user stories must be developed by involving everyone.

Marc Westphal, Bayer
It is our goal to be able to see the current status of all safety valves that are integrated in our equipment or kept as backup, in a system overview. A digital twin enables us to not only see all technical properties but also track all changes and repairs performed on safety valves by service providers. In this scenario, service providers are under obligation to document all modifications made on the information sharing platform and provide them to the operator.

Furthermore, a digital twin provides a basis for standards on the operator side for inventory management and engineering purposes.

Marc Westphal, Bayer
It will automate the digital collaboration between the manufacturer, the operator and the service provider. This means that all stakeholders will be able at any time to access the information they need at the given stage of the asset lifecycle.
For the future we especially expect a strong increase of data sharing between stakeholders. This will result in additional, innovative services. In particular, condition monitoring of equipment deployed in the field will allow service providers to offer new kinds of services. Manufacturers will be able to derive key information for product development directly from application data.

This will enable manufacturers to improve the longevity of products as well as the accuracy of end-of-life predictions for scheduled replacements.