Digital twins: the key to agile manufacturing in SMEs

Digital twins are essentially data-driven, interactive models that replicate the geometry of a machine or production process, but also replicate its behavior, performance and responses to control inputs in real time. This is what makes them fundamentally different to static CAD models or offline simulations. Digital twins are “live” – they are continuously updated with live data and can respond to commands in exactly the same way as the real system.

In an industrial automation context, this means that robots, production lines and entire workflows can be safely developed, tested and optimized in a virtual environment long before they go live on the shop floor. Twins offer enormous benefits for small and medium-sized enterprises:

• Digital twins make it possible to experiment with process modifications, optimize cycle times and train users without disrupting production.
• They reduce commissioning times by enabling virtual automated process validation.
• They prevent costly mistakes by detecting integration problems early on.
• They improve workplace safety by allowing teams to simulate dangerous scenarios before they occur in the real world.

Crucially, this technology is no longer the sole preserve of large corporations with big R&D budgets. The availability of simulation platforms, cloud-based infrastructure and intuitive robot programming tools means that SMEs can now introduce digital twins step by step, scaling their use in line with growing business requirements. The democratization of this technology is a help to SMEs in an Industry 4.0 landscape where agility, efficiency and rapid innovation are key to staying competitive.

How digital twins work

Digital twins are created in a structured process that combines data collection, modeling and real-time synchronization. It begins with data collection: CAD models of the robot or machine, kinematic parameters, payload data and the relevant sensor data are assembled to ensure that the virtual model accurately replicates the physical system.

This data provides the basis for a physically accurate simulation in applications like NVIDIA Isaac Sim that are able to replicate a robot’s movements, collisions and interactions with its environment under realistic conditions. Once the virtual model has been created, it’s time to move on to the control integration step. Platforms such as the Wandelbots NOVA robot programming tool connect the physical and digital worlds, enabling the same commands that control the real robot to also be sent to its simulated counterpart. It is at this point that the digital twin becomes an interactive, “live” system capable of replicating automated processes in real time. In the final synchronization and feedback stage, live process data from the physical machine (such as joint angles, sensor statuses and force-torque measurements) is fed back to the digital twin. This continuous exchange of data ensures that the virtual model remains up to date with changes in the real world and even allows it to predict future conditions. The addition of analysis layers or AI algorithms allows the digital twin to be used to run what-if scenarios, optimize curves or detect anomalies before they have an impact on production. This data collection, modeling and live synchronization workflow transforms digital twins into a powerful tool that can be used not only for simulations and training but also for process validation, predictive maintenance and system optimization, creating a live virtual factory.

From simulation to implementation

One of the biggest benefits of digital twins is their ability to bridge virtual development and physical implementation. Once a process has been modeled and tested in a simulation, it can be implemented with the real robot with minimal adjustments, significantly reducing commissioning times. Instead of programming on the shop floor, which can disrupt production and pose a safety hazard, engineers can validate processes virtually, optimize robot paths and identify potential collisions or inefficiencies in advance.
This workflow was also the focus of Andrea Bondin, Research Support Officer at the University of Malta, in a project coordinated by the Steinbeis Transfer Hub in Berlin. With funding from the EU’s ERA Shuttle project, he developed a digital twin of a collaborative UR5 pick and place cell at the Steinbeis Transfer Center Digital Workspace in Horb, in partnership with Wandelbots NOVA and NVIDIA Isaac Sim.
The first step was to replicate the process in a virtual environment where robot tasks could be defined, sequenced and optimized without using the physical hardware. “The next step will be to have these validated workflows performed by a real robot in the lab in order to demonstrate seamless and reliable transfer from the simulation to a physical environment”, explains Bondin. This stage is expected to provide valuable information about how digital twins can cut commissioning times and ensure process safety under real-world conditions.
“This case study illustrates the practical benefits of digital twins for SMEs. They enable rapid prototyping of automation solutions without needing to interrupt production, reduce the cost of experimenting with new layouts or processes, and allow employees to be trained in safety before they interact with the real machine”, concludes Professor Dr.-Ing. Tim Jansen, Manager of the Steinbeis Transfer Center Digital Workspace. Moreover, the ability to simulate fault conditions and recovery strategies improves reliability and prepares users for unexpected scenarios.

Benefits for SMEs

For SMEs, the introduction of digital twins no longer calls for massive investment or special IT infrastructure. Cloud-based simulation platforms, low-code robot programming tools and scalable computing power mean that even small manufacturing companies can start with a single process cell and gradually scale up as and when demand grows. Furthermore, open standards like OpenUSD ensure interoperability, allowing SMEs to integrate multiple tools without being tied to particular vendors, so that they can build up a complete digital twin environment gradually.

Digital twins also support HR development by providing safe, virtual training environments where employees can learn how to operate and maintain equipment without the cost or danger associated with the use of physical machinery. This has particular benefits for SMEs, where production stoppages are costly and skilled workers are in short supply. By enabling faster training, safer experimentation and better planning, digital twins can help SMEs to remain competitive in a rapidly changing market.

In the future, the integration of digital twins with AI-powered analytics and predictive maintenance systems will make them even more valuable – as well as assisting with simulation and planning, this will help SMEs to predict failures, reduce waste and improve energy efficiency. Projects of this type show how academia and industry can collaborate to make these technologies accessible and usable in practice, ensuring that the benefits of Industry 4.0 are within the reach of any company, regardless of its size.