A new project hopes to develop a way to enable widespread 3D printing of scrap metal.
Normally, metal 3D printing is highly controlled due to the precise nature of the print parameters. Slight variations can throw off a print job, resulting in poor-strength parts or even print failures. Often, jobs require some iteration to determine the optimal print parameters.
Now, there’s an initiative undertaken by the Worcester Polytechnic Institute to get around this problem.
Why do this when there are perfectly good methods of 3D printing metal today? It’s because of fear of a future where it might become more challenging to obtain pure metals. This is indeed a growing concern given the geopolitical games being played these days.
The idea is to create a new, renewable source of metals for 3D printing — from scraps. Normally, scrap metal is not considered for 3D printing because of the purity required. Purity is required because in order to produce highly consistent metal parts, the input material must be identical, as well as the print parameters and hardware used.
The “Rubble to Rockets” project, sponsored by the US Defense Advanced Research Project Agency, hopes to change that. The concept is to use a machine learning system to identify the varied input scrap types and then develop a prediction for the output 3D print materials behaviour.
Associate Professor Danielle Cot explained:
“This work is crucial as it allows us to build high-quality components from unknown source materials with new confidence. Our goal is not just to build a single solution but to create a framework that guides future innovations. By improving our predictions and understanding of material performance, we can pave the way for new advancements in additive manufacturing from diverse and unpredictable sources.”
In other words, input materials are analyzed, and the strengths of the output material are predicted. This material is then converted into 3D printable form and sent to the printer.
To demonstrate the feasibility of this approach, the project intends on 3D printing a sounding rocket using the new materials generated by the process.
I’m a bit skeptical of this approach, as it varies considerably from today’s standard procedures, which are known to work well for production parts. Yes, you could print a rocket using the random material, but could you print a thousand of them? Would they each operate in an identical manner?
It seems to me that they would have to derive new 3D print parameters for each material, and that is known to take much time and iteration. However, the report says that the process is intended to run at a “rapid pace”. If so, it may be that they are also generating likely optimal print parameters as well as predicting the material’s properties.
In any case, this is a good step towards more full recycling use of metals. All too often, our cranky metal 3D print processes require near-perfect materials to work, and anything that removes that constraint should be welcomed.
