In the future, many components will come from the 3D printer, especially those that are custom-made for specific users – from prostheses to spare parts for vintage vehicles. Additive manufacturing processes, aka 3D printing, make it possible to produce individual pieces that are individually tailored to the application. In additive material extrusion, molten material is applied layer by layer. This process can also be used to process high-performance plastics used, for example, in medical technology, electrical engineering or aircraft construction.
But it is precisely with high-quality products, for example in medical or mechanical engineering, where no compromises as to the quality are acceptable. And to date it is difficult to verify the quality within the additive manufacturing processes. Many companies are therefore still reluctant to use 3D printing.
In addition to the external geometry, the internal structure is also critical for component quality. Honeycomb structures often ensure that the 3D-printed component is light and yet as stable as possible. In order to detect defects such as irregularities or cavities in the internal structure, the component must currently be X-rayed. So far, there is no functioning process-integrated monitoring for 3D printing processes based on the principle of material extrusion. The IPH research project Quali3D will change this. The goal of the scientists: In the future, it will be possible to check component quality during printing. Together with his team at IPH, project manager Alexander Oleff is developing an optical measuring system that can be integrated into an extrusion 3D printer.
The heart of the optical measuring system will be a camera that records images of each individual printed layer. An image processing algorithm will automatically evaluate these photos and detect errors. Errors can occur, for example, if too fast movements cause vibrations or if the material supply of the 3D printer is disturbed.
The biggest challenge for the researchers is that the quality inspection must be carried out without reference, i.e. without a comparative image. “3D-printed components are often unique,” explains Oleff. “Therefore, there is usually no reference – for example, layered images of an identical component that has already been printed without errors – with which the algorithm could compare the print result. Instead, it is possible to use texture analysis for troubleshooting. The algorithm evaluates the images and finds irregularities. Alternatively, it would be conceivable to read the machine code. From this it can be deduced, among other things, at which point how much material is to be applied. The planned print result can then be compared with the actual component.
The scientists now have two years to develop the measuring system and the associated algorithms: the “Quali3D” project runs until summer 2021. Both users and manufacturers of 3D printers should benefit from the results. The results will enable manufacturers to further develop their machines and increase the quality level of material extrusion. The researchers assume that this will also increase the use of 3D printing, especially in industries such as medical technology or for safety-critical applications. In the future, users will be able to monitor the production of each individual product and promise their customers tested quality – even for unique specimens. In addition, they can reduce production costs: If errors are detected during printing, the printing process can be readjusted or interrupted in good time. This saves time, energy and material.
Companies interested in process-integrated monitoring of extrusion 3D printing can still participate in the research project. A kick-off meeting is planned for the end of August. Interested companies should contact Alexander Oleff by e-mail at firstname.lastname@example.org