The timing of this partnership between AM and aerospace couldn’t be better, as larger and more efficient engines, heads-up displays and electronics push thermal management requirements to new extremes. Conformal cooling offers a solution to this challenge, something that wouldn’t be possible without 3D printing. Heat exchangers are the key technology in this space, and so using additive manufacturing to create conformal heat exchangers is a logical move.
United Technologies has done just that in a joint project between the UTC Additive Manufacturing Center of Expertise (AMCOE) and America Makes. UTC engineers focused on developing a rigorous method of dividing heat exchanger geometry into key design features in order to optimize the powder-bed fusion process.
Engineering.com had the opportunity to sit down with Venkat Vedula, Executive Director for Additive Manufacturing at UTC Technologies, to find out more.
How are thermal management requirements for aircraft changing?
Overall, whether it's about the on-board advanced electronics or more efficient jet engines, one of the key challenges is managing the heat that's coming out of them. You have to remove the heat that's being generated from source like electronic boards and chips or combustor and turbine components. So that is significantly driving the need for higher and better performance thermal management systems. That’s been one of our key technology focus areas.
I understand that heat exchangers are being required to conform to more constrained spaces. Why is that?
One of the challenges when you look at an engine cross section, as an example, is that the space that's available for any thermal management systems is fairly limited. A lot of times, we are constrained just in terms of the rest of the structure that is in place. So, the challenge for the past several decades has been that when you look at the heat exchangers, they're typically similar to what you have in your car. Those radiators are primarily rectilinear and when you’re limited to that sort of shape, you need to add ducts to bridge the space in between, so that creates efficiency losses.
The UTC case study mentions a lack of prescriptive processes for additive manufacturing of heat exchangers. Is this a problem of standardization, certification, or something else?
Certification is certainly one element of it. But one of the key challenges that we’ve found working with the powder-bed fusion process is really understanding the process itself. As you’re probably aware, these machines typically have about 150 process parameters. So, when we are developing the process, we’re looking at the key process variables, like the power, scan speed, different scan strategies and how they impact the quality of the material. What is a good operating zone versus what happens when we are outside of the zone? Just fundamentally understanding about the process itself, and how do you tie the process to a particular application?
There’s notorious variation between additive machines. Are you working on multiple models or is there a single machine dedicated to the process?
Yeah, fun question! We have found that machine-to-machine variation in two different sets of machines from the same manufacturer and even on the same machine from one location to another.
So, that's a key part of our focus within the organization: to develop that understanding and look at, what is that variation? But we also have different OEM machines, including the single laser and multi laser machines.
Does using additive manufacturing affect your material selection?
It does, because the alloys that are available today for metal additives are primarily nickel based alloys, titanium alloys, and then of course, higher-strength aluminum alloys are coming along. For this particular application we focus on nickel based alloys, but we’re also developing aluminum alloy heat exchangers as well.
Can you explain how designing with AM in mind affected your approach to the design process?
Yeah! One of the phrases I always hear is, "With additive, the complexity comes for free." That statement is true, but one of our key challenges is that we don't have design tools that give us what we call end-to-end tool chains. That is, you can start with a concept, go to the design inspiration to reduce the weight, structural optimization, that sort of thing—and ultimately, be able to build that part. So, in parallel, as we were applying this methodology for the heat exchangers, we also looked at, from a design perspective, what are the manufacturing constraints?
As I mentioned before about the thin fins with the different orientations, what kind of supports do you need? Or not? Incorporate all of that, and we can build heat exchangers successfully.
Was there any post-processing involved in making the heat exchangers?
The heat exchangers in the case study had zero support structures internally, and so the only post-processing required is just trimming off the supports from the outside. Significant effort went into figuring that out: How do you orient the heat exchanger, to be able to achieve that?
What about a heat treatment for stress relief?
Yes. Right now, for the nickel base, we are going through the HIP process.
How does the size of the AM heat exchanger compare to one made with conventional methods?
On a back-to-back comparison, we're able to achieve about twenty percent size reduction for the same performance.
The case study mentions engineered laser scan strategies. Can you tell us more about that?
Yeah! The individual scan strategy goes to what I mentioned about understanding the process. All of the machines that we buy come with default parameters—a certain, let’s say, laser scan power or speed—and we’re able to tailor that as part of this program. So, when we’re doing thin fins, we can tailor the process parameters a certain way, as well as the scan vectors themselves, to achieve fully dense defect free fins in the entire heat exchanger core.
For more news from United Technologies, check out Advanced Robotics for Manufacturing (ARM) Announces Selection of Eleven Technology Projects.