AI in production is mandatory

The birthday slogan of SmartFactory Kaiserslautern (SFKL) "20 years of the future" expresses what it stood and still stands for: For the production of the future. "The future of yesterday is our present today," says press spokesman Dr. Ingo Herbst. "Industry 4.0 has arrived in the application. When we look around the Hannover Messe, we see digital technologies everywhere, for which the foundations were laid 20 years ago."

At the Hannover Messe, the research network and its industrial partners showed which production developments are becoming important for the industry. This year, the SFKL addressed visitors directly with questions. "We try," says Herbst, "to make the benefits of the latest technologies clear to the various target groups. We want them to realize that they can quickly save money and resources. In recent years, we have carried out industrial projects where companies have been able to significantly reduce costs by implementing AI, i.e. small software snippets." This is what SFKL emphasizes time and again: IT and OT flow into each other. "The way we think about cables in control systems must change in the direction of modern programming in high-level languages," emphasizes Herbst.

60-year-old milling machine with digital twin

To show the connection between yesterday, today and tomorrow, the experts have equipped an old milling machine with a feed motor, digital twin and CNC control. The skills of the milling machine can be called up as skills, for example the milling of rectangular or circular pockets. They can be controlled and parameterized via a standard OPC UA interface. Before starting work, the digital twin calculates costs and energy requirements, plans the trajectory and checks it for collisions. It thus assumes the role of a CAM system and returns information that can be used, for example, to create a quotation. The digital twin thus helps to automate and optimize the entire planning process.

"Geometric features of a component can be manufactured directly from CAD using the latest Industry 4.0 paradigms," explains Herbst. "This saves time-consuming programming work." When searching for suitable machines for a specific job, the milling machine, called SkillMill, automatically reports back, as do all other machines with similar capabilities in the machine park, in accordance with the skill-based approach. To ensure that a skill can be executed with its specific parameters, the digital twin carries out a feasibility check in advance to check the feasibility of the work order.

"Old machines are still frequently found in production halls, but are sometimes little used and are often incompatible with modern production systems," says Andreas Wagner from RPTU Kaiserslautern, who is responsible for the project. "This example shows that such machines can be transferred to the future through retrofitting." Integration into data rooms, as envisaged with Gaia-X or Manufacturing-X, would also be conceivable.

With the Phuket production island, SFKL demonstrated how the skill-based approach can work in practice in modern production systems. In a matter of seconds, the production plan adapts dynamically to the availability of components or existing machines. "Currently fluctuating supply chains or material bottlenecks can be absorbed by our software agents," says Herbst. "Time-consuming and therefore expensive planning and control work is no longer necessary."

AI, robots and humans become colleagues

In the Phuket assembly module, a cobot carries out various manufacturing processes. It can switch between cobot and industrial robot mode. The integrated safety concept was developed by an SFKL working group in collaboration with TÜV Süd and Sick. It is based on Sick sensor technology, which ensures that operation is adapted in accordance with safety requirements and that the worker cannot be injured. The cobot is controlled via human gesture recognition, which enables direct interaction.

An automatically generated worker assistance system is also implemented, which is derived from the information in the administration shell (AAS). The workstation configuration, such as table height or assistance complexity, is adapted to the worker's personal preferences and qualifications. This means that employees can perform any work flexibly, or even unskilled workers can be deployed more easily in the production process.

Every success story has its prehistory

What is everyday life today began as an idea much earlier, of course. The Internet became a reality in the 1990s. Legend has it that it was Prof. Detlef Zühlke who thought about networking production machines in order to take manufacturing to a new level. Researchers had been working on the Internet of Things since the early 2000s. "There were also smart homes, so it wasn't really that big a leap to say we also want to have a SmartFactory," says Zühlke. In 2005, seven companies and research institutes (DFKI, BASF, KSB, Pepperl+Fuchs, ProMinent, Siemens, TU Kaiserslautern) founded the "Technologie-Initiative SmartFactory KL" (SFKL).

Work quickly began on building a demonstrator to put the idea into practice. "We didn't just want to do laptop research, we wanted to do real research and work together with industry," says Zühlke, emphasizing that it was essential to work in a manufacturer-neutral way. "In the end, it was actually relatively easy to put the whole thing on track and thus prepare the ground for what would later be referred to as the big bang, Industry 4.0. That only came a few years later. But all the work that was necessary for this was started at SFKL."