Process monitoring needed for 3D printing, says Photonics West panel
Greg Blackman reports from a metal 3D printing panel discussion held at Photonics West in San Francisco, where the plea to the photonics community was for better in situ monitoring tools for laser machines
GE’s additive Advanced Turboprop engine could be considered the poster boy for metal additive manufacturing. The design takes 855 parts and combines them into 12 3D printed components; the engine is 5 per cent lighter with 20 per cent better fuel consumption, and the development process was accelerated by 12 months.
‘No technology like this has ever existed that can do these things all at once,’ stated GE Additive’s technology development leader Andy Martin, speaking during a panel discussion on metal 3D printing at SPIE Photonics West at the end of January.
In his statement at the beginning of the discussion, Martin noted that the kinds of improvements traditionally made in jet engines have been incremental: in materials, say, in one place in the engine. Additive manufacturing, he said, is poised to transform the rate at which GE can improve its product. ‘It’s [AM] very important technology to us and we’ve embraced it utterly.’
In autumn 2017, GE Additive unveiled a beta machine for its Atlas project with a powder bed measuring 1m2 by 300mm high. It can build parts 600mm in diameter. The internal architecture has a gantry so that the laser spot can be focused accurately on the powder. The systems can also be updated with additional lasers as required.
Despite the leaps in technology being made by GE Additive, Trumpf, Siemens, and others, the panellists were restrained about praising the capabilities of metal additive manufacturing too much.
‘I think additive manufacturing is real; it will change manufacturing, but it will take longer than many claim,’ commented Peter Leibinger, CTO of Trumpf. He said that expectations need to be managed when it comes to the technology, as it will take time for the machines to become productive and robust.
‘I think there will be disillusion coming our way in the coming years where a lot of the expectations – many of which are unrealistic – will not be fulfilled, and companies that are hopeful about this technology today as customers will turn their back on it. And they will make a mistake by doing so,’ he said. He added that it is the industry’s task to educate customers as to realistic expectations, along with how to design parts to get the best out of 3D printing.
One way system makers are working on improving machines is through process monitoring, such as sensing what is going on inside the melt pool. Martin said the photonics industry has a big role to play to develop the tool for this job. In situ metrology would give better understanding of how thermal distortion occurs in the part, as well as whether powder is being applied correctly.
‘Saying it boldly, you cannot print [geometrically] accurate parts today,’ Leibinger said. ‘It will depend on metrology if this ever will be possible, but at this point it’s not possible.’
Leibinger observed that Trumpf has been working on laser cutting – a relatively simple process compared to 3D printing – for 30 years and it is still being optimised. He said that to accelerate the development of 3D printing, engineers need ‘to find the levers to make it more robust’ – this could cover any number of parameters, although Leibinger mentioned specifically that it was important to make the machines operate with lower grade powder, as only being able to use high-quality powder made the process expensive and wasteful. Building robust machines involves sensing and process monitoring.
Karsten Heuser from the additive manufacturing part of Siemens’ digital factory division remarked that engineers should not underestimate post-processing, even if the printers are not perfect. He said that a lot of costs in additive manufacturing come from post-processing and that sometimes it might be better to have a lower resolution in the printer, because the part will go through post-processing anyway. ‘You need to take care of the whole value chain in the factory for additive,’ he said. Siemens has built gas turbine components for power plants, which have been proven for 4,000 hours operation.
Standards and simulation
The panellists generally agreed that metal additive manufacturing was at too early a stage to try and impose standards on the process or in any one industry. What is needed is education about AM’s capabilities and also, to some degree, a change in engineering mindset when designing the parts, according to Heuser, which he said software can help with. Siemens is developing simulation software to give an understanding of what will happen when the laser hits the powder bed before the print run.
However, accurate simulations need high quality data, which goes back to in situ monitoring. Martin commented: ‘I’d like to understand the depth of the melt pool, the shape of the melt pool, and the temperature distribution – those are all really important so we can validate those simulations,’ he said.
As to the future of metal additive manufacturing, Leibinger commented that he is both optimistic and pessimistic. ‘If I look at the curricula of German engineering universities, they tend to be very traditional. However, if I look at the kids they all have 3D printers at home. Many young people today work and play with plastic 3D printers. They take this for granted.’