Share

The Digital Twin, an enabler of flexible, dynamic production lines

© PCharlot/CEA
On-demand production—one of today’s key challenges in industry—has the capacity to reduce both costs and environmental impacts. CEA-List used its Papyrus4Manufacturing modeling platform to design a new software architecture to provide the flexibility and automation needed to make on-demand production a reality. Tests with CEA-List partner APS confirmed the solution’s potential.

At a time when we are already looking ahead to Industry 5.0, with its focus on people, sustainability, and resilience in production, many of the challenges of Industry 4.0 remain unresolved. One of them is order-driven production. To satisfy a wide variety of orders—whether in terms of number of units or type of product—order-driven production lines must be extremely flexible and automated.

This requires three key capabilities: monitoring production asset availability; dynamically assigning production operations to the assets; and automatically integrating new assets into the existing production process.

Today’s standards and concepts only partially provide these capabilities.

  • The AAS (Asset Administration Shell) standard for creating Digital Twins of industrial assets includes a management and communication interface.
  • CSS (Capability Skill Services), a concept that represents assets (robots, conveyor belts, etc.) by capability, enables the dynamic selection of the most suitable asset for a given job.
  • PnP (Plug-and-Produce) refers to the ability of an industrial system to be reconfigured, enabling new assets to be integrated into a production process and collaborate automatically and dynamically with other assets.

Figure 1: Software architecture of CEA-List’s XAAS solution.


 

However, AAS presents certain limitations that prevent it from fully integrating CSS and PnP. For example, AAS cannot detect new assets or provide detailed descriptions of asset capabilities.

CEA-List used its Papyrus4Manufacturing platform, with its AAS-standard Digital Twin modeling and generation tool, to develop XAAS, an extendible AAS software architecture designed to enable flexible, dynamic production lines (Figure 1).

XAAS extends AAS with two major additions:

  • Asset capability description parameters and
  • A suite of four new services: Discovery for asset detection, Matchmaking for capacity/job matching, Configuration to configure assets, and Integration for the integration of assets into the current process.

 


Figure 2: Demonstrator


 
These new features make Plug-and-Produce a reality.

CEA-List implemented the XAAS architecture on a sandbox use case: filling packaging boxes. During the experiment, new assets were automatically integrated into the current production process as soon as they were turned on, and the system correctly assigned tasks to the right robots, confirming the relevance of the XAAS concept.

With XAAS, several asset configurations for a given job can be tested and the best one chosen according to criteria like production time, energy consumption, and the like.

Our Papyrus4Manufacturing modeling platform enables the flexibility and automation that the industry of the future will need.

Rebecca Cabean

Quang-Duy Nguyen

Research Engineer — CEA-List

The INTEROP project has confirmed the potential of our architecture using representative production assets. Building on these results, APS has renewed its partnership with CEA-List.

Rebecca Cabean

Maxime Morin

R&D Manager, APS — APS

Learn more

Prototype

  • «Plug&Produce on a production line» – TRL 5.

Industrial project

  • Objective: Develop a Plug&Play software component for an assembly line.

Partner

  • APS, based in Béziers, France, provides plug-and-play interoperability for all components of production systems.

Flagship publication

  • «An AAS-based Architecture for Plug and Produce and Order-Driven Production». Nguyen, Q.-D., Morin, M., Dhouib, S., & Navarro, N. (2025). IFAC-PapersOnLine, 59(10), 927–932. https://doi.org/10.1016/jifacol.2025.09.157