Fresh strength for industry

Electronic products, printed

Instead of having products made in lowwage countries and then carting them halfway around the world, 3D printing could make it worthwhile to bring the entire production chain back to Germany in the future. This would mean faster production with less climate impact, but that is not all. It would also eliminate the need for fragile global supply chains. “Our goal is to make it possible to print complex products at the push of a button – locally, right where they are needed,” says Kausch’s colleague Lukas Boxberger, who also heads a department at Fraunhofer IWU. Just feed the data into the machine, and it instantly goes to work producing a coffee maker, Bluetooth speaker or robot vacuum right near the customer.

“This kind of machine essentially has to handle four subprocesses: holding non-printable things like textiles, film or wood veneers in place, setting up structures, integrating wiring and cords for electronic products and inserting other elements, like motors or displays,” Boxberger explains. The researchers achieved this with different tool heads that can be swapped out as needed to allow the machine to handle the next task. A prototype version called multi-material additive manufacturing, or MMAM, already exists. It's not only the machine itself that marks a big step forward in production. In particular, the ability to integrate wires opens up a range of new options. That is because the design freedom famously afforded by traditional additive manufacturing has only applied to plastic components so far. “Our system adds maximum functional flexibility to 3D printing. We can incorporate electrical properties and data lines along with antenna functions or thermal features,” Boxberger says. The print head they developed puts the wire and plastic together. It can combine any kind of metal wire – copper, constantan, nickel titanium – in various diameters with all kinds of thermoplastics, thereby realizing any desired function. The machine has already automatically printed its first product without any human participation at all: inserting the film, winding a spool, placing magnets and a circuit board. The outcome was a speaker, ready to use.

Another advantage of this method is that it allows for a complete rethink of production. Boxberger explains: “Cords can be printed into the housing, and switches can be incorporated directly into the housing via pressure-sensitive areas. Having wires and cords inserted into the component or printed directly onto it makes the components more shock-resistant and increases their level of integration of different functions while also reducing material consumption.” The team is currently working to scale the system up to the size of a shipping container. The new machine will be capable of 12 operations, up from four: When complete, it will also be able to print complex wire harnesses, optical functions, fiber optic data lines and lighting elements. The prototype is scheduled for completion in late 2026.

Lasers in the powder bed

While 3D printing is an obvious choice for production of plastic-based products, laser processes are the additive method of choice for products made from metal. This includes laser powder bed fusion, in which a laser beam is used to melt metal powder layer by layer. Developed by the Fraunhofer Institute for Laser Technology ILT in Aachen in 1996, it is now the dominant technology on the market, accounting for more than 80 percent of additive manufacturing of metal parts. It can also be used to print plastics.

Nevertheless, the method is far from being exhausted. Dr. Tim Lantzsch, a department head at Fraunhofer ILT, views beam shaping – literally changing the shape of the laser beam – as having particular potential. A Gaussian laser beam is the default, but it overheats the material locally, melting not only the layer of metal powder being worked in the process but also the two to three layers underneath it that have already been processed before. That is not an efficient use of energy. “With a more complex beam shape, we can distribute the energy around the powder bed in a completely different way,” Lantzsch says. “Even simple beam shapes with rectangular or circular distributions double the process speed while improving component quality.” Simulations help the researchers to gauge how different beam shapes will affect temperature distribution and the weld pool. The team is also building up the necessary hardware. Machines for additive manufacturing that offer this kind of versatility – such as one of the world’s largest powder bed machines, located at Fraunhofer ILT – are undergoing continuous development and refinement to make it easier to translate the researchers’ findings to industrial use.

Researchers at Fraunhofer IWU aim to push the limits of laser powder bed fusion by adapting their scanning strategies. “In the current practice, the laser’s scanning paths are pre-set by the machine’s software, with only a small scope for modification. But they don’t always reflect the best possible choice. Components with a complex shape are often produced with just one scan strategy and one set of laser parameters, which affects the component’s dimensional accuracy,” explains Dr. Juliane Thielsch, the technology manager at Fraunhofer IWU. Thielsch and her team worked with the Chair of Virtual Product Development at TU Dresden to devise a software solution that makes it possible to modify specific individual laser scan paths and associate them with separate parameters such as laser power or speed. “The software gives us a lot of freedom in the processing workflow,” Thielsch says. “The user can choose between different scanning strategies and edit or delete scan paths themself or add new ones.” This has allowed the researchers to reduce deviations from dimensional tolerances in short-stem shoulder joint implants from 21 percent to 3 percent.

Metal meets metal

In some cases, it can be a good idea to combine different kinds of metal and different functions. Take a tool used for injection molding, for example: Steel guarantees high strength, while a copper alloy improves heat conductivity. At present, these kinds of metal-and-metal products are often produced using electrochemical processes that fall under the EU’s regulation on the registration, evaluation, authorization and restriction of chemicals (REACH). As a result, researchers are looking for alternatives. “To be able to additively apply two different materials accurately and separately from each other, we developed a powder bed-based process with two material chambers,” says Dr. Georg Schlick, who heads the Additive Manufacturing department at the Fraunhofer Institute for Casting, Composite and Processing Technology IGCV. “We apply the first material, melt it with the laser, suction up the unused powder and then start the same process with the second material.”

FIDENTIS, a spin-off, plans to use this method to produce telescopic crowns for dental prostheses. Producing crowns like these has previously involved a lot of manual work, which makes them expensive and tends to put them within reach only for those with private insurance. The new method could change that, with the first few patients receiving this form of care as early as at the end of April. Schlick also sees a wealth of potential for other applications, especially where thermal and mechanical stresses coincide. “We can use this method to improve the functionality of a component such as a tool, thereby lowering the costs of the finished product,” he explains. “Hopefully, this will be one way we can bring manufacturing back to Europe from Asia.”

Wire-based laser cladding times eight

Instead of using lasers to liquefy powder, the material that is to be melted can also be used in wire form. Unlike in a stream of powder, one hundred percent of the wire fed into the system in this method ends up in the weld pool. So far, the process has been too slow for high-performance industrial coatings. In the future, though, it could come to seriously compete with powder-based laser material deposition, also known as cladding. This development is being spurred by a method called COAXquattro, which was developed by researchers from the Fraunhofer Institute for Material and Beam Technology IWS. There are several things that set it apart. While it is customary to feed the metal wire through the center of the machine’s head and melt it with laser beams from the sides, the team simply flipped the setup around. Now the laser is in the middle, and multiple wires are fed in from the outside. “This lets us use standard lenses, which are currently usable up to 20 kilowatts of laser power. Compared to previous laser welding heads, the system is three to four times more productive in terms of the melt rate,” says Dr. Elena Lopez, a department head at Fraunhofer IWS.

The specially developed welding head also allows the team to incorporate up to four wires and four streams of powder into the process. “By using different met-als in a single work step, we can create coatings out of in-situ alloys. The starting materials aren’t mixed together until during the coating process,” explains Holger Hillig, a project manager and col-league of Lopez’s. This means desired properties can be introduced into the coating on a localized basis without any additional processes, achieving aspects such as locally higher hardness or thermal conductivity or a particular coefficient of expansion. COAXquattro moves wire-based laser cladding into an economically efficient range, even for very large components such as planetary gear mechanisms – with laser-coated sliding bearings – for wind turbines or to armor-plate extruders for tire manufacturing to guard against wear.

The method is opening up new potential applications in additive manufacturing: “The side surface of a train used by Deutsche Bahn used to take seven different production steps. We can take that down to three,” Lopez points out. Because additive manufacturing makes different geometries possible, the researchers are fo-cusing on a bionically inspired structure similar to the veins in a leaf as a way to stabilize the side portion. This makes the component 30 percent lighter. In the long term, COAXquattro may also replace materials affected by resource shortages by allowing manufacturers to achieve the same properties through alloys. How exactly to do that is being explored by the researchers involved in ORCHESTER, a Fraunhofer flagship project.

Replacing critical substances

Another laser method from Fraunhofer ILT that has become an industry standard is extreme high-speed laser material deposition (EHLA). In this process, the powder is melted directly in the laser beam before it makes contact with the component. This allows for thin coatings and makes it possible to coat large areas in a short time. “EHLA is especially at-tractive when high process speeds, thin coatings and sustainable material use are needed,” says Dr. Thomas Schopphoven, a department head at Fraunhofer ILT. EHLA is already an established alternative to hexavalent chromium coatings, which are harmful to the environment, and going forward, it could also substitute for critical per- and polyfluoroalkyl substances (PFAS). The technology also promises to help with meeting the Euro 7 emission standards: “EHLA is a central solution for coating brake discs for use in cars. The strong, firmly bonded coat reduces abrasion, thereby lowering particulate emissions,” Schopphoven explains. Industrial capacity is currently being built for series production.

Wetlaid nonwoven technology only for paper? Much more is possible!

Innovative leaps are sometimes possible even with longstanding methods, as shown by wetlaid nonwoven technology, which has been known for some 2,000 years and is currently primarily used in production of paper. Fraunhofer IGCV research scientist Violetta Schumm sees other possibilities as well: “This method opens up a way of reusing fibers from composite materials.” Fiber-reinforced composites are used in a host of different fields for their potential for lightweight construction, from aviation to wind turbine blades and from automotive applica-tions to manufacturing of sporting goods. But the end of the product life cycle brings one key question: How can these materials be recycled effectively?

It is currently possible to reclaim the individual carbon fibers from waste streams and process them. The challenge, however, lies in efficiently further process-ing these recycled fibers. An innovative modified wetlaid nonwoven pilot plant at Fraunhofer IGCV is used to do this. Much like in paper manufacturing, the recycled fibers are separated in large tanks of water. The fiber dispersion this creates passes through several process steps and is ultimately transported on a continuous belt that acts like a sieve, producing a web of nonwoven fabric that can then be rolled up. The plant in Augsburg can use this method to process recycled fibers up to 30 millimeters in length at production speeds of up to 30 meters per minute. The nonwoven fabrics produced are used as reinforcing components in second-life composite materials, where they offer a broad spectrum of functional properties, for example in relation to electrical and thermal conductivity. By continuing its research and development activities relating to this process route, Fraunhofer IGCV is making an important contribution to closing the material loop, thereby support-ing economically sound and industrially usable recycling concepts.       

Robots dismantle car batteries

Recycling is also a dominant topic these days in the automotive sector, which still accounts for a large portion of German val-ue creation. But the switch to electric cars has shaken up the worldwide automotive market, leaving German manufacturers struggling to maintain their once-leading role. “One crucial factor when it comes to competing successfully is the availability and cost of the raw materials needed for batteries and electric motors,” says Prof. Alexander Sauer, head of the Fraunhofer Institute for Manufacturing Engineering and Automation IPA.

“That means it is all the more important not to simply shred end-of-life batteries that still contain valuable raw materials.” The most fundamental requirement for reusing battery components is that it has to be possible to dismantle the various parts and separate them by material. In the DeMoBat project, 12 alliance partners worked together, coordinated by Fraunhofer IPA, to devise concepts and applications for doing just that. “We set up a largescale system for dismantling car batteries for the first time,” says Anwar Al Assadi, a research scientist at Fraunhofer IPA. “Cutting cords, loosening and milling screws – robots carry out the entire dismantling process automatically.” Sensors and 3D camera systems check the interim results after each step. The DeMoBat has concluded; the researchers are now retool-ing the system to use it for dismantling a different automotive industry component. This approach could also be translated to large appliances like washing machines or dishwashers. For small appliances, the researchers involved in the Desire4Elec-tronics project are working on automated dismantling solutions.

From sensor to cloud

Teasing out the very best from the in-dividual manufacturing technologies is vital if the “Made in Germany” label is to retain its shine – after all, they are the backbone of production. But it isn’t enough on its own. “Our job is to combine existing approaches in effective ways and turn all those individual solutions into one big harmonious whole,” says Michael Fritz, head of office of the Fraunhofer Cluster of Excellence Cognitive Internet Technologies CCIT. Researchers from the cluster are studying how to do that in the Edge Cloud Continuum 4 Production (ECC4P) project.

Their key idea is to combine edge and cloud computing to create an end-to-end data space. Once that is done, computations will be performed where it is more efficient and cost-effective to do so. The “edge” in this scenario means the area where the data is first collected or generated – local sensors, machines or devices. Especially for processes that need to be concluded in split seconds, there is a need to use the computing power and storage space avail-able in this edge. Other tasks that are lengthy and compute-intensive – training an artificial intelligence, for example, or performing simulations – take place in the cloud. In order to transfer data dynam-ically between the edge and the cloud, it is crucial to connect them with what are known as data space connectors. Their importance became apparent in the Catena- X project, an initiative of the German automotive industry that aimed to create a shared data infrastructure for the entire value chain. “A bunch of different tech-nologies were developed as part of Cate-na- X. But the main takeaway from the companies was: We need a connector that can be used in industry,” Fritz recalls. Based on the IDSC connector, which Fraunhofer developed years ago, experts from Fraunhofer and industry created the Eclipse Dataspace Connector, or EDC. “The EDC is now on its way to becoming the industry standard,” Fritz says.

The data space is created once a con-nector is in place. To fill it, the researchers developed various sensors and combined sensors together to generate true added value. They tackled three generic applications in the process. First was adaptive gear grinding, which is needed during production of gears. A structure-borne sound sensor was incorporated into the grinder for this; where previously, it was necessary to wait until after the finished gear was installed inside a gearing mechanism to see whether the process had gone well, the process can now be evaluated right while the work is in progress. The second sensor solves a similar problem with forming and pressing: Controlling the pressure exerted during pressing significantly reduces waste and shortens retooling time. The third sensor monitors the drilling and milling process. It is an ideal way to retrofit old machines and incorporate them into digitalized value creation. To expand on the value added by the sensor data, this information is analyzed locally in an edge cloud, together with the machine data, and then transferred via EDC to the cloud, where it is used to train an artificial intel-ligence. The entire processing chain is structured much like an MLOps pipeline to develop and operate powerful and scalable AI/ML solutions. Security is provided by a tool called Clouditor, which was developed at the Fraunhofer Institute for Applied and Integrated Security AISEC. Fritz explains: “We’ve already implemented the entire approach as proof of concept for the grinding process. It is beneficial in all forms of production where milling, forming, drilling and grinding take place, regardless of the industry, plus it is especially suitable for retrofitting existing machines.”

Matrix Production Instead of Line-Based Methods

Traditional line-based production is highly efficient, but it doesn’t have much to offer in terms of flexibility. “If there are multiple production lines, even machines that don’t see much use have to be available for all production lines, which creates redundancies,” says Dr. Simon Harst, head of a business unit at Fraunhofer IWU. He and his colleagues from Fraunhofer IWU and Fraunhofer IPA hope to replace this rigid system with what they are calling a matrix production. Instead of traditional lines, they create islands. The product makes its own way through the production area, carried by driverless transportation systems. “You only need one of the machines that aren’t used often, and capacity utilization is much more efficient as well,” explains Dr. Marcel Todtermuschke, also the head of a business unit at Fraunhofer IWU.

The production architecture that con-trols how a product moves around was created in the Fraunhofer flagship project SWAP. Some of it is available in an open source version. The matrix approach to production is now being further developed in the Robotics Engineering Application Lab for Matrix Production (REAL-M) research platform with the ultimate goal of “first time right” production that generates no waste or scrap at all. Infineon was able to use the matrix approach to reduce the time needed to produce wafer structures by about 20 percent while increasing throughput by 50 percent.

Can AI solve mechanical engineering problems?

Maintaining the excellent reputation of the “Made in Germany” brand in the long term also means tapping into the benefits of artificial intelligence. LeakAIr, a project at Fraunhofer IPA, shows just how big an impact AI can have. It focuses on com-pressed air, a perennial issue in production. There are about 60,000 compressed air systems in operation in Germany, accounting for 7 percent of all the electricity consumed by German industry. However, it is estimated that around a third of the compressed air leaks out unused – and typically even unnoticed. The costs of this waste can quickly run into the tens of thousands of euros per company per year, not to mention the impact on the company’s carbon footprint.

“We wanted to shine a light on this issue and locate leaks automatically,” says Daniel Umgelter, head of a business segment at Fraunhofer IPA. Working in tandem with the Institute for Energy Efficiency in Production (EEP) at the University of Stuttgart and with sensor company SICK, Umgelter developed a way to use a smart algorithm to pinpoint leaks. To collect enough data for training purposes, the researchers poked holes in a machine to allow the compressed air to escape and then ran it through various load cycles. This produced an approach that SICK plans to market under the name LeakAIr. “End-toend automated detection of leaks could reduce energy losses by about ten percentage points on average,” Umgelter says. “Extrapolated to the whole of Germany, that would work out to between 80 and 160 million kilowatthours of electricity saved per year. That’s the equivalent of the average yearly energy used by 22,000 to 45,000 households.”

Thanks to AI technologies, data is viewed as the “gold” of Industry 4.0. Researchers from Fraunhofer IWU provided impressive proof of just how accurate this comparison is through their work on the EmulDan (Energy Efficiency in Production through Multivalent Data Usage) joint research project. An improved understanding of processes made it possible to lower energy consumption by as much as 70 percent in the case of rotary swaging machines, which are used to produce lightweight construction components. In the case of press hardening, a method that combines the benefits of forming and heat treatment in a single step, hybrid process models revealed energy conservation potential of up to 20 percent based on a digital twin – that is, in a dynamic virtual visualization of the process.

These quantum leaps became possible because there is still a certain amount of mystery associated with many processes. Data is collected, but its potential goes unused in many cases. This results in grinding wheels being changed much too early and thinner sheet metal being heated with unnecessarily high amounts of energy for forming purposes. To change that, the researchers incorporated sensors into areas where this was not possible before, such as in rotating components during machining. Strain gauges there measure the tool’s deformation on three axes down to the micrometer range and then transfer the data wirelessly to the outside. However, the majority of the researchers’ development work lay in combining the data from the various sources and analyzing it using artificial intelligence. “We use the results of the AI analysis to control the machines automatically and adjust them to actual circumstances,” says Stefan Polster, head of the Sheet Metal Processing and Tool Design group at Fraunhofer IWU.

Documentation and searching for information? Let the AI do it!

The artificial intelligence can also handle tedious, timeconsuming tasks that stand in the way of value creation. Studies have shown that 25 percent or more of people’s work time is lost to activities like these, which include looking for information and the obligatory documentation. The response from the Fraunhofer Institute for Production Systems and Design Technology IPK is an AI production assistant based on a generative language model, the same technology used for ChatGPT. “Companies have heterogeneous IT landscapes. This means employees have to interact with a large number of different IT systems, which inevitably means their work becomes more fragmented,” says Prof. Julian Polte, head of a business unit at Fraunhofer IPK. “Our aim is to replace them with a personal virtual assistant for employees that automatically provides the necessary information and handles complex repetitive tasks. We expect this to significantly boost productivity.” Ap-plications based on these principles are already being successfully transferred to industry − where they are doing their part to help keep Germany’s reputation for superior craftsmanship going strong.