Solid oxide fuel cells have many attractive features but are prone to cost and materials degradation problems.
To help overcome such challenges, Mohawk Innovative Technology, Inc (MITI) has designed some of the critical parts for longer life and greater efficiency.
One example is the anode offgas recycle blower (AORB), part of the balance-of-plant, essentially a low-pressure fan/ compressor system that recycles the exhaust and returns it to the front of the fuel cell.
“SOFC balance-of-plant designers were thinking that this blower would be an off-the- shelf unit,” says Dr Jose Luis Cordova, VP of engineering at MITI. “But due to the process gases in the system, traditional blowers tend to corrode and degrade; the hydrogen in the mixture attacks the alloys the blowers are made of and also damages the magnets and electrical components of the motors that power the blowers. Most blowers also contain lubricants, like oil, that degrade as well. So you end up with very low-reliability blowers — representing a significant portion of the balance-of-plant cost — and your SOFC plant needs an overhaul every two to four thousand hours.”
The USDOE has been investing in SOFCs for years (to the tune of $750 million since 1995), and this performance falls far short of the USDOE’s goal of an operating lifetime of 40 000 hours for a typical SOFC — as well as an installation-cost reduction from an average of $12 000/kWe to $900/kWe.
Time to rethink the blowers. “We realised that Mohawk’s proprietary, oil-free, compliant foil bearing (CFB) technology, specialised coatings, and extensive turbomachinery expertise were a good fit for this challenge,” says Cordova.
USDOE funding provided the means for Mohawk to design and test AORB prototypes in a demonstrator SOFC power plant run by FuelCell Energy. Rigorous testing under realistic operating conditions measured durability and performance, with the latest versions demonstrating no significant degradation of components or output and complete elimination of any performance or reliability issues.
Yet the cost of an AORB remained prohibitively high — in large part due to its high-speed centrifugal impeller, which operates continuously under extreme mechanical and thermal stress. For longest life, this part must be made from expensive, high-strength, nickel-base, corrosion-resistant superalloy materials like Inconel 718 or Haynes 282 that are difficult to machine or cast. In addition, achieving optimal aerodynamic efficiency in an impeller requires complex three-dimensional geometries that are a challenge to manufacture. And because of the incipient nature of the current SOFC market, impellers are produced in relatively small batches, and economies of scale are difficult to realise.
Enter AM
How to bring costs down? Additive manufacturing (AM, aka 3D printing) provided a compelling answer. While the original project with FuelCell Energy was evolving, Mohawk was also getting calls from R&D groups looking for help with their own fuel-cell component designs. “Because many of these manufacturers and integrators were still at the research stage, each one had a different operating condition in mind,” says Cordova. “Using traditional manufacturing to make just the handful of the custom impeller wheels or volutes they wanted would have been extremely expensive. So that’s where we started looking at AM; we did our own research into AM system makers and connected with laser-powder-bed-fusion (LPBF) provider Velo3D.”
With its goal of reducing costs and improving performance of SOFCs, the DOE is enthusiastic about innovative manufacturing methods such as AM, says Cordova. “Their funding [through The Small Business Industrial Research Project] supports our current partnership with Velo3D as well as our previous one with FuelCell Energy.”
The switch to AM was an eye-opener: “Our traditional, subtractively manufactured impeller wheels were running up to $15 000 to $19 000 apiece,” says Cordova. “When we 3D printed them, in small batches of around eight units rather than one at a time, this dropped to $500 to $600.
“As well as cutting manufacturing costs, LPBF is the one technology that could provide us with the design flexibility we were looking for. AM is indifferent to the number of impeller blades, their angles, or spacing—all of which have a direct impact on aerodynamic efficiency. We now have the geometric precision needed to achieve both higher-performance rotating turbomachinery designs and reduce associated manufacturing costs.”
For 3D printing impellers on a Velo3D Sapphire system (at Duncan Machine, a contract manufacturer in Velo3D’s global network), the choice was made to use Inconel 718—one of the nickel-based alloys with a strong temperature tolerance that withstand the stress of rotation best.
“Inconel was very attractive to us because it’s chemically inert enough and retains its mechanical properties at pretty high temperatures that definitely surpass aluminium or titanium,” says Mohawk mechanical engineer Hannah Lea.
Although Velo3D had already certified Inconel 718 for their machines, Mohawk did additional material studies to add to the body of knowledge about the 3D-printed version of the superalloy. “Our tests demonstrated that LPBF 3D-printed Inconel 718 had mechanical properties, like yield stress and creep tolerance, that were higher than those of cast material,” Lea says. “This was more than adequate for high-stress centrifugal blower and compressor applications within the operational temperature range.”
Iteration made easy
As their impeller work progressed, Mohawk’s engineers collaborated with Velo3D experts on design iterations, modifications and printing strategies. “It was really interesting because we didn’t have to make any major design changes to the original impeller we were working with — with Velo3D’s Sapphire system we could just print what we wanted,” says Cordova. “We did do some process adjustments and tweaking in terms of support-structure considerations and surface-finish modifications.”
Of course tweaking is just another day in the office for design engineers. As the impeller project progressed, AM provided much faster turnaround times than casting or milling would have allowed, since parts could be printed, evaluated, iterated and printed again quickly. In subsequent 3D printing runs, multiple examples of old and new impeller designs could be simultaneously made on the same build plate to compare results.
The relatively small size of the impellers (60 mm in diameter) necessitated the team’s development of a “sacrificial shroud” — a temporary printed enclosure that held the blades true during manufacturing.
“What was really interesting about this approach is that shrouded impellers are, for most current additive technology, basically untouchable because of all the traditional support structures they require,” says Velo3D’s Mohawk-project leader Matt Karesh. “We used a, not support-free, but reduced-support approach. Mohawk was saying, ‘we don’t need the shroud in the end, but the shroud makes our part better, so we’ll attach this thing that’s typically extremely hard to print—and just cut it off after.’ Using Velo3d’s technology, they were able to build that disposable shroud onto their impeller, get the airfoil and flow-path shapes they wanted, and then it was a very simple machining operation to remove the shroud.”
Surface finish was another focus. “The surface was a bit rough in our early iterations,” says Mohawk engineer Rochelle Wooding. “What was interesting about the sacrificial shroud was that it gave us a flow path through the blades that we could use to correct for roughness using extrusion honing; it took some further iteration to determine how much material to add to the blades to achieve the required blade thickness that we wanted. The final surface finish we achieved is comparable to that of a cast part, and suits our purposes aerodynamically.” What’s more, all critical design dimensions enabling proper impeller operation were within tolerances.
Next steps
Next steps are retrofitting AORBs with the new impellers and testing them in field conditions. “We expect that successful execution of these two tasks will fully demonstrate that 3D-printed Inconel parts delivered by LPBF technology are a viable and reliable alternative for manufacturing turbomachinery components,” says Cordova. Work is already underway using AM for other blower parts like housings and volutes.
