To learn how Aerojet Rocketdyne went from exploring AM has a manufacturing technology to its increased reliance on 3D printing for part production, ENGINEERING.com spoke with Jay Littles, director of advanced launch vehicle propulsion at Aerojet Rocketdyne.
The RL10
The RL10 rocket has been in service since 1963, sending numerous spacecraft into orbit, including New Horizons and Voyager 1, the first spacecraft to reach interstellar space. As an important legacy product, the company sought to introduce AM to the manufacturing equation in order to reduce costs.
The current model for the RL10C-1 features a complex array of drawn, hydroformed stainless-steel tubes that have been brazed together to create the thrust chamber. Aerojet Rocketdyne was able to improve on this design by consolidating the parts into two copper components that were then 3D printed via selective laser melting (SLM).
These tubes were essentially replaced with a series of channels designed into the larger copper parts. As a result, the part count was reduced by over 90 percent. The entire system was printed in just under a month’s time, cutting the lead time by several months. Upon completion, the thrust chamber underwent a successful hot fire testing. As far as Aerojet Rocketdyne is aware, this is the largest 3D-printed copper part to undergo a hot fire testing successfully.
Littles explained that the copper alloy played a role in the overall design efficiency of the system. The thrust chamber is a part of an engine that performs an expander cycle, which sees the liquid fuel flow through the chamber’s channels—previously made up of steel tubes¬—picks up heat as the combustion takes place within the chamber and becomes a gas to power the engine’s turbine. This process is made that much more efficient when using a highly conductive material like copper.
“We went from something that was steel tubes and a much larger chamber to a shorter chamber made from copper,” Littles said. “We were able to implement some interesting design features using AM to allow us to get enough energy out of the chamber to run the cycle the way we did.” Printing with copper involves some specific challenges related to the conductivity of the material during the printing process. To ensure that the end part maintained the proper physical properties necessary for the chamber’s design, Littles’ team relied on smaller-scale versions of the part before fabricating the full-scale component, perfecting the process as they went along.
“The RL10 is a product that’s in service and has a tremendously long and successful history, so when we’re implementing something that is a new technology into a legacy product like that, there’s a tremendous pressure to make sure that we don’t do anything that affects the reliability of the system,” Littles said. “We’ve done extensive process optimization and materials characterization, developing new design curves frankly associated with these additive materials and subcomponent testing and ensuring that we have the right inspection processes. There’s a tremendous emphasis to make sure that we do not affect reliability.”
This work actually builds off of a large body of background work that the company has already performed with AM. In this way, the RL10 was a culmination of its 3D printing experience thus far.
Exploring AM
Littles relayed that the company’s work into AM actually predates Aerojet Rocketdyne in its current form, before Pratt & Whitney Rocketdyne was acquired. About six or seven years ago, in the early days of SLM, the team at Pratt & Whitney Rocketdyne was exploring the use of technologies to make rocket construction more affordable as it applied to two different types of engines: very high-performance engines, like the AR1 booster engine currently being developed, and more affordable engines, like the gas-generator cycle F1 rocket used in the Apollo program.
To determine how AM might affect the cost of an engine, the company designed a 3D-printable version of its F1 gas-generator injection, which historically required tens of parts and would have taken a year or two to make.
“We did it just as a geometry feasibility to see if we could build something like this,” Littles said, explaining that the design was printed by an external supplier. “We got a lot of people internally very excited because we were able to make something that would have been multiple pieces and taken a long time to make into something that looked very much like a single piece.”
Littles pointed out that, though they were able to create such a part, there were some very important pieces missing. “When we made that part, we realized we had done ourselves somewhat of a disservice because we made something that looked like an F1 gas-generator injector, a highly complex part that operates in extreme environments. But we really had no idea how the thing performed. We knew nothing about the material properties, densification or surface effects or how the part would actually perform in this environment.”
It was at that point that the team began to actually fill in a lot of the missing pieces, performing materials characterization and process optimization. Over the years, the company has been able to nail down these details in such a way as to formulate standards for approaching new AM projects. For this reason, Aerojet Rocketdyne has recently begun reporting numerous test fires for 3D-printed rocket components.
“We have design rules and design handbooks now that dictate what we can and cannot do with the SLM process,” Littles said. “We’ve done component geometry demonstrations, but we’re also really at the point now where a number of those material systems we were working with are sufficiently far along in our technology readiness process such that we are going into system validation and even production with some parts.”
3D Printing Rocket Engine Parts
Whereas the RL10 is for a legacy product, and, thus, demonstrates the quality and reliability of the AM process, Aerojet Rocketdyne is also 3D printing parts that prove other important benefits of AM for rocket engine production. In 2013, the company performed successful tests on a liquidoxygen/gaseous hydrogen rocket injector assembly, which was followed by successful tests on the Bantam demonstration engine in 2014 and the AR1 engine in 2015.
For the Bantam, the engine was composed of three 3D-printed parts: the entire injector and dome assembly, the combustion chamber, and a throat and nozzle section. Using AM in this way enabled a cost reduction of about 65 percent. At the same time, total design and manufacturing time was cut from over a year to just a couple of months.
About the approach to the Bantam family of engines, Littles said, “Compared to the RL10, the Bantam family is a little bit more open. It’s a broader family, and we’re doing more things to change the design. There’s a little bit more flexibility in what we can do from a creativity standpoint within the Bantam family development that we’re working through now.”
While the Bantam family has allowed Littles and his team to be more creative in the design process, the AR10 will enable Aerojet Rocketdyne to shorten its design cycle. This is key, as the company has a target delivery date of 2019. The successful delivery of the engine will make it possible to meet the requirements of the 2015 National Defense Authorization Act, which calls for the replacement of the Russian-built RD-180 by an American-made alternative for national-security space launches by 2019.
“With AM, we’re able to make components and get them into tests very quickly,” Littles said of the AR10. “Instead of just performing analysis on a number of different designs to try to optimize performance, we can actually get things to test and have empirical data such that we can anchor our models and have a much shorter design cycle than we would have historically.”
Since the company’s initial research into AM, Aerojet Rocketdyne has been able to set up a program in such a way that it’s become easier to translate knowledge of the technology from one engine to the next. “It’s somewhat of an art to begin with,” Littles explained. “We did do some engineering analysis and calculations to guide that early work. We started using things such as calculations related to energy types to make predictions. There are a lot of variables in the process. It’s not just the laser power, but it’s scan strategy, how you raster the laser around and powder requirements. There’s a tremendous number of those variables.”
More recently, the company has been relying on software through a separate project that allows Littles’ team to optimize 3D printing parameters analytically in order to fabricate parts in a much more efficient and repeatable way. At the same time, the company is developing a domain expertise that makes it possible to apply AM to a part design when it appears to be advantageous.
“We have a good understanding of the types of components and the systems or subsystems that we get the best benefit for,” Littles said, “taking complex assemblies of parts and making them into one or two pieces. It’s part count reduction and touch labor reduction, so we know sort of the families of components that offer the big benefits.”
Though none of the rocket parts have yet flown, Aerojet Rocketdyne has production components that are in the qualification stages that will be on flight systems. The qualification process for such parts is obviously a long but critical one for the systems necessary for sending rockets into space. When it comes to making those rocket engines, every opportunity to reduce costs, production schedules and manual labor helps.