Fraunhofer Institute for Production Systems and Design Technology
Better Than New
Where massive forces are at play, even the hardest component eventually gives way. When in use, turbine blades develop cracks, shafts develop dents and punching tools start showing signs of deformation at the edges. Modern additive processes are used to efficiently repair such damage and make the worn component ready for use again.
Laser Powder Directed Energy Deposition (LP-DED) in particular has become an efficient process of choice for these kinds of repair tasks. A laser beam is used to create a weld pool into which a nozzle conveys a powdery, usually metallic filler material. This melts in the weld bath, and movements of the nozzle or the component produce weld beads, two-dimensional coatings and more complex 3D structures layer by layer. In this way, LP-DED can be used to repair a wide range of components by local welding. For simple geometries such as running or sealing surfaces of shafts, the process is established and widely used in industry.
But repairing localized damage using DED can also be worthwhile for expensive components such as stamping or molding tools, like those used in automotive engineering. However, complex geometries and individual damage patterns, as in indentations, edge chipping or deformation, make it significantly harder to use the process efficiently in such scenarios.
Repair welding using LP-DED
Generation and job simulation of toolpaths in the CAM software
Scangineering software as a tool for the additive repair process chain
The Repair Process Chain
For such repair tasks, a process chain like the following is usually run through:
1. The defective component is 3D scanned, including its damaged area.
2. The component is prepared for repair (for example by grinding or milling out the damaged area). 3. The prepared component is 3D scanned again.
4. The 3D scan data thus generated is processed, detecting the defective areas and generating a difference volume that distinguishes the nominal and actual states.
5. An additive LP-DED repair process is planned in a CAM program.
6. The repair process is carried out.
7. The repaired part is heat treated and reworked.
In this chain, steps 4 and 5 in particular pose challenges. A weld path cannot be derived easily from a simple 3D scan. It is true that the capture and measurement of geometries by laser scanning or photogrammetry is standard for many applications today. But the resulting large data volumes are often only used for visualization or measurement purposes. To use them for repair processes, the scan data must be processed, aligned and converted into parameterized 3D models. Only with these models can path planning for the repair process be carried out in a CAM program. This process of converting 3D images into CAD models is called reverse engineering. It is still carried out manually to a large extent, which requires trained specialists and a great deal of time. For highly individualized components, the repair effort is therefore often very high. As if that were not enough, the actual repair process is complicated by welding and material challenges. Forming tools are often made of cold-work or hot-work steels with relatively high carbon contents. While the carbon content contributes to good hardenability, it also reduces the weldability of the material, which is problematic for repair welds. This is all the more true since the repaired areas must meet the requirements for resistance and hardness in the same way as the original component did as a whole.
Researchers at Fraunhofer IPK are addressing these challenges with an end-to-end automated repair solution. Thanks to Scangineering with automated component recognition and geometry-based modeling as well as modern AM processes, defective components can be made ready for use again with little technical effort.
3C Scanning and Reverse Engineering results in Scangineering
The reverse engineering process known as »Scangineering« is an in-house development of Fraunhofer IPK. In this process, intelligent algorithms are used to preprocess, align and parameterize 3D scan data of components. This means that a geometrybased and thus modifiable 3D model is generated from the point clouds of a laser scan, which, for example, can be loaded into a CAD program.
Users can intervene in the conversion process at any point as input providers and analysts. At the same time, they are relieved of manual and repetitive steps. By means of Scangineering, complex components, but also other objects such as machines or even buildings can be turned into virtual models quickly and easily.
CAM Planning: Calculating Toolpaths
For automated component repair, in the next process step the additive repair processes are planned on the basis of the models. The detected geometric defects are used to calculate the tool paths and welding commands for the additive building process.
Mathematically determined volumes, surfaces and curves reduce the need for additional auxiliary geometries to be created, thus simplifying and accelerating the programming process: Coating surfaces are clearly defined and can be selected, complete difference volumes can be programmed by building them up layer by layer and complex curves can be used as support curves for the alignment of tool paths. By simulating build-up and checking traverse paths, the tool paths can be verified and any collision points can be detected and eliminated in advance.
Additive Repair Process and Finishing
The final step comprises the design of the process parameters and the actual component repair. This phase requires materials engineering know-how to ensure a metallurgically high-quality and durable repair: Material-specific properties must be considered. The high carbon content in tool steels, for example, favors high hardness values – desirable to increase service life. However, in conventional repair welding, such as arc processes, the high heat input in combination with the carbon content and other alloying elements can lead to cracks in the component. Laser-based processes such as LP-DED can be advantageous here: The high energy density and the resulting low heat input reduce the cracking tendency, the size of the heat-affected zone, the influence on the base material as well as the degree of base and filler material mixing. Additionally, the selection of the filler material is not limited to the alloy of the base material.