The standard manufacturing and assembly processes for heat exchangers can be expensive and time-consuming. They also limit the geometric configurations that can be considered during the design process.
Additive manufacturing (AM), when viable, can be a more efficient manufacturing process and gives engineers much more design freedom, often allowing them to create features that standard machining cannot.
Researchers are looking for ways to build high-performance heat exchangers using AM. However, most metal-based AM processes are so expensive that commercial production is not viable.
Looking to see if another approach might work, a research team led by Gregory F. Nellis, a professor at the University of Wisconsin (UW), designed and built an efficient and cost-effective heat exchanger using lower-cost, deposition-based AM processes. The team successfully printed a heat exchanger using fused deposition modeling that has sufficient thermal conductivity to outperform standard heat exchangers, at a more affordable cost.
Higher performance, less cost
Nellis’s plan was to use AM to create a two-fluid heat exchanger in which air exchanges heat with a liquid. Performance results from testing with this prototype would then be assessed against an industry-standard copper/aluminum heat exchanger manufactured with conventional equipment.
Nellis and his team—UW engineering professors Tim Osswald and Natalie Rudolph, Jake Boxleitner (a graduate student at the time, now employed at Brayton Energy), and Tom Mulholland (a doctoral student who now works with Samsung) — employed fused deposition modeling (FDM), a much less-expensive AM process, to manufacture several heat exchanger prototypes. The AM material developed by the team consisted of a composite polymer filled with conductive flakes to provide high conductivity in the direction of heat flow.
First attempts using these AM methods focused on fabricating extended surfaces with aerodynamic shapes that bridge the air-side gap formed between two water channel walls. That build did not meet the performance targets.
One of the key challenges was maintaining a continuous tool path within each build layer to allow uninterrupted extrusion and therefore defect-free water channel walls.
“It was tough coming up with a heat exchanger that was reliably leak-tight while still having high performance,” Nellis said. “It took us a while to understand that we needed designs where the nozzle extruding the filament never stopped. Any time the nozzle stopped—for example, to add a fin or some other feature—we would inevitably get a leak. Once we established this as a constraint, the design evolved quickly toward a final version that could achieve commercially relevant performance.”
Using a Design-of-Experiment approach, the team created simulations for unique geometric configurations, using three different frontal air velocities per geometry. The design variables were constrained based on several factors, including manufacturing limits.
The team relied on UW’s Center for High Throughput Computing to carry out many parallel simulations of heat exchanger designs and moved quickly toward an optimal solution. The team also developed a rapid method for screening various polymer/filler combinations to ensure the right combination of properties and printability (both the design and material are now patented).
The final result of the process was the development of a simple heat exchanger that performed on par with the computer models. The 3D-printed model can be easily modified as manufacturing capabilities improve, resulting in higher performance.
Test results were especially encouraging: The measured heat transfer rate for the model exceeded that of the industry-standard copper/aluminum heat exchanger of the same size and at the same operating conditions by 12 percent and exceeded the heat transfer rate of the conventional industry-standard stainless-steel heat exchanger by 285 percent.
“One of the nicest surprises was how quickly we could iterate on a design using additive manufacturing,” said Nellis. “It was sometimes possible to modify a design one day, print it overnight, test it the next day, and modify it again. The design/build/test cycle was greatly accelerated compared to other manufacturing technologies.”
Beating the heat
Nellis used the FDM process to create an air-cooled, two-fluid heat exchanger technology that is cost-effective to manufacture and competes well with standard commercial HVAC.
“Our work has led to an understanding of the fundamental design constraints that must be respected for this combination of manufacturing technique and application,” he said.
Nellis has been contacted by several companies about commercialization of his process, which has potential impacts for other fields outside of HVAC.
“I think any multifluid device that must be low cost could be manufactured using deposition-based AM processes and therefore benefit from our research,” he said.
From Nellis’s perspective, the best aspect of the project was how the most interesting and important work happened at the intersection of different fields.
“It was a lot of fun for Jake and I, with expertise in heat exchanger design, to meet regularly with the manufacturing team—Tom Mulholland, Natalie Rudolph, and Tim Osswald,” he said. “Often, we would together find a solution that was better than either team could have found in isolation, because together we could see the bigger picture.”
Mark Crawford is a technology writer in Corrales, N.M.
