Smart, rapid prototyping informs AM overhaul
Illustration of the steps taken by Dunbar, et al., To develop design guidelines for thin fins used in AM heat exchangers. Photo credit: Dunbar, et al.
Rapid prototyping has been the predominant use case for 3D printing and additive manufacturing for over 30 years. With metallic AM, being smart and fast is even more important when it comes to prototyping, given the cost of materials, machinery, and post-processing. Therefore, I introduced the concept of Intelligent Rapid Prototyping (RIP). In this column, I intend to delve deeper and share an example that may spark an idea or trigger a new way of thinking for users developing AM parts for production.
Consider a heat exchanger – a very “hot” app for AM right now (sorry, I couldn’t resist the pun). Typically, these are made by stacking layers of perforated or finned sheets on top of each other and then welding or brazing them together in an assembly. It is easy to make the sheets, it is easy to make the fins or perforations in the sheets, and it is easy to stack them together in an assembly. The end result, however, is often a large, square structure that must be “designed around” if you wish to integrate it into a larger system. Look no further than the radiator in front of your car’s engine to see what I mean.
With AM we can now reimagine what a heat exchanger looks like. For example, it can be designed to conform (i.e. wrap around) the shape of the system or component that is being cooled (e.g., a hot cylindrical pipe). An airplane engine oil cooler, for example, no longer needs to look like a metal shoebox; instead, the surface can be redistributed and adapted to natural spaces and gaps in an airplane engine if you are using AM.
When redesigning such a heat exchanger, users may start to wonder: how far can I make a wall or fin while printing it? In what orientation should they be placed? Which section should the fins have? Can I vary the fin sections, and if so, what is the best configuration for each fin and set of fins? What is the impact of surface roughness on the fins? What if the fins are not completely dense? would a porous fin create more surface area and be more efficient?
These are just a few of the questions one encounters when (re) designing AM heat exchangers. While they may seem straightforward and easy to answer, they aren’t, especially when it comes to laser powder bed fusion (L-PBF). The interactions between the raw material used, the processing conditions in the machine and the design geometry and print orientation are so closely related that it is impossible to guess – and currently too complex to fully model and simulate – them. best conditions for everyone. Worse, since designers have never had the ability to create such complex geometries and internal passages before, the heat transfer coefficients and friction factors they look for in a table or graph to size a heat exchanger do not exist. And now?
Enter what I call Rapid Intelligent Prototyping (RIP) to deal with every unknown in a smart and efficient way – in this case, to develop FA design guidelines for a thin wall or fin and, more importantly, mitigate the risk that it will not print or fail. during use. The following is a brief summary of a RIP-like study (although they did not call it that) performed in our lab by Alex Dunbar under the supervision of Ted Reutzel, Director of CIMP-3D.
As described by Dunbar et al., Any study of this type begins by defining the unknowns or uncertainties of interest: in this case, the thickness and the angle of the fins. The construction orientation of each fin relative to the coater blade was also of concern as a very thin wall built in parallel with a coater can be easily damaged depending on the L-PBF system. The team also had the opportunity to vary some of the process parameters, choosing to take into account a range of scan speeds, laser powers and exposure strategies. Experienced AM users know that varying process parameters is essential to pushing the boundaries of printable geometry. Look no further than Velo3D or SLM Solutions, two L-PBF companies that have mastered the use of process parameters to enable metallic AM without support.
With multiple design geometries, construction orientations, and process parameters to vary, it’s time to use Experimental Design Techniques (DoE) to intelligently vary different combinations of each variable to maximize the information obtained with a minimum. effort and cost. Users should also consider the number of builds to run as part of the study, as these add time and cost. Finally, users need to think about how to measure the outcome of each generation, as it’s easy to get overwhelmed with too many samples given the capabilities of AM – another area to be “smart in”. When using RIP to address unknowns and mitigate risk.
Once the DoE is planned, the construction – or constructions – are executed and the results are analyzed to identify the best combination of parameters for, in this case, the thin wall geometry relative to the process parameters and construction orientations. . In some cases, parameters may be easy to control during a construction (eg laser power and scan speed) while others should be treated as uncontrollable “noise factors” (eg, orientation. of the construction of a thin wall in the heat exchanger). In the latter case, choose settings that minimize variability to reduce risk as much as possible.
This study gave the team the knowledge to develop a design guideline to fabricate fully dense and robust thin walls at different construction orientations. In the example of Dunbar et al., Varying process parameters, along with design geometry and construction orientations, allowed them to achieve a 70% thinner wall with no porosity using custom parameters versus to standard processing parameters. Not a bad result for a profitable, fast and intelligent prototyping study. What would you do differently if you were armed with such an approach?