In day two of our Frontiers in additive manufacturing series, Victorien Menier and Pieter Coulier look at some reasons why metal additive manufacturing hasn’t lived up to certain promises for some adopters, and how certain issues can be addressed.
“Complexity is free” is a common saying in additive manufacturing (AM). The idea is that the manufacturing cost of an AM part remains roughly constant as you increase its geometric complexity, contrary to conventional manufacturing methodologies (casting, forging, milling, etc.) for which the cost of added, integrated features tends to increase exponentially. The concept of reliable, low-cost complexity sparks a lot of interest and can mean different things to different disciplines when talking about AM.
For innovators, it evokes the promise of freedom—freedom from the constraints of traditional manufacturing—and the opportunity to solely focus on part performance. Incredibly complex parts that might only exist in engineers’ imaginations can now be manufactured in a matter of days instead of months or years, which is a huge paradigm shift in design cycles and lead times.
For legacy part manufacturers, the saying evokes the promise of a supply chain management revolution. Some standard components consist of thousands of parts produced in various locations before assembly. For diverse reasons (e.g., wear and tear) components in the field regularly need to be replaced or upgraded. Doing so in a timely manner is a formidable tooling, storage and shipping challenge that requires an early buildup of inventoried parts. With AM, however, part files (i.e., digital models) can be stored in digital warehouses, without the need for costly physical inventory, and used for production just-in-time, or with only a short lag, as needed.
Unfortunately, most conventional metal AM technologies fail to deliver on these two promises.
With the proliferation of metal AM technologies, both innovators and producers of legacy parts have now come to realize that they have traded one set of design constraints for another.
Trying to print most parts “as-is” can result in an unacceptable part –poor dimensional accuracy, failure to meet required mechanical properties, visible defects etc.– or worse, cause a catastrophic print failure that could waste expensive material or even damage the printer.
Thus, a successful print often requires an expert user to employ “Design for Additive Manufacturing” (DfAM) guidelines, that is, to re-design the part to accommodate a broad set of limitations that are common in AM. In addition, this often requires fine-tuning print parameters for a particular part and a particular printer.
Three typical limitations of conventional metal AM printers are overhanging features, residual stresses, and thin walls.
Unsupported overhang features can have a devastating effect on a metal 3D printed part’s manufacturability and performance. If the angle between the part’s surface and the horizontal plane falls below around 45 degrees, the result will be a misshapen part or even catastrophic print failure.
Users of conventional AM systems are asked to orient their part in a way that minimizes such low-angle regions and to add supports to the remaining ones. After printing, removing these supports can be prohibitively costly and, in many cases, even impossible (e.g., a supported surface inside a shrouded impeller).
The melting and cooling of metal at different temperatures and time scales can cause residual stresses to build up in a part. This stress accumulation can distort or crack the part—or even the build plate. Users are typically asked to avoid large areas of uninterrupted melting and/or to prescribe lasing hatch patterns. Thin walls typically need to be thickened, even more so when they are angled.
As an alternative, an end-to-end AM solution can overcome the need for DfAM compromises with a combination of software, hardware, and sensor-guided controls that are tightly integrated in a single production stream. From start to finish, processes such as print preparation, powder deposition, laser pattern and more are customized to every geometric feature in a design. Controls monitor the underlying physics of each process to balance metrics such as print speed, residual stresses, dimensional accuracy, and surface quality.
This kind of advanced solution can be used “out of the box” by even non-experts. You simply import the native CAD model of your part into the set-up software on the AM machine—without any re-design needed—and it automatically generates a file of production (laser) instructions for that system. These instructions can also be sent to any compatible printer, anywhere in the world, with confidence that the result will be identical.
Such extensive automation nevertheless allows for flexibility. If the user desires a characteristic that deviates from the default on a region of a part—to identify trade-offs between print speed, surface quality, or dimensional accuracy, for example—they can realize that intent through a variety of selection options. As print preparation can be a computationally intensive procedure for some of the most complex parts, an available workflow allows offloading to the cloud or a local cluster. Once the print run is completed, users can inspect the outcome layer by layer. Special attention is also paid to the balancing of the workload across multiple lasers to maximize print speed, while avoiding any undesired interference between lasers that are active simultaneously.
The promise of metal AM to deliver innovation, without the compromises of DfAM, is finally being realized by the more advanced 3D systems that are now available. Product designers, engineers and supply chain managers are now free to move beyond the limitations of what conventional technologies have been offering and achieve their ambitions using today’s fully integrated AM solutions.
Picture: A 3D printed heat exchanger (Image credit Velo3D)
Victorien Menier is Senior Software Engineer and Pieter Coulier is Principal Software Engineer at Velo3D.
Subscribe to our free @AuManufacturing newsletter here.