24.10.2017
Blue Origin’s BE-4 engine during its first hotfire test, which the company said October 19 was a success. (credit: Blue Origin)
-
It was a brief statement, but a significant one. On Thursday afternoon, Blue Origin tweeted out a six-second video with the caption, “First hotfire of our BE-4 engine is a success #GradatimFerociter.” At long last, the BE-4 engine that the company had been developing for several years had come to life.
The company said little else about the test, declining to state the test’s duration, thrust level, or even when the test itself took place. Nonetheless, the test is a milestone not just for the company but the American space industry. It will force decisions by other companies in the months to come about the development and use of other engines.
“Congratulations to the entire Blue Origin team on the successful hotfire of a full-scale BE-4 engine,” ULA said, without discussing its engine decision plans. |
BE-4, capable of producing nearly 2.45 million newtons of thrust, uses liquid oxygen and methane propellants. The engine has one certain application, on the company’s own New Glenn launch vehicle. That vehicle will use seven BE-4 engines in its first stage and one in its second, giving the vehicle the ability to place up to 45 metric tons into low Earth orbit and 13 metric tons into geostationary transfer orbit, with a first flight expected around 2020.
Blue Origin, though, is not the only potential user of the BE-4. United Launch Alliance has expressed its interest in using the BE-4 for its next-generation launch vehicle, the Vulcan. The two companies formally announced their cooperation in using the engine on the vehicle back in 2014, and as recently as earlier this year, ULA leadership emphasized that the BE-4 remained the frontrunner for Vulcan—provided the engine passed hotfire tests.
“We look first to the combustion instability as the chief technical risk that must be retired before we’d be able to pick an engine,” ULA president and CEO Tory Bruno said in an interview six months ago. At that time, hotfire tests were scheduled to begin in a matter of weeks.
Those tests, though, were apparently delayed by a mishap on a Blue Origin test stand in May that destroyed a set of “powerpack” hardware, a key component of the engine. The company acknowledged the failure, saying it was “not unusual during development” and that it would be “back into testing soon,” but otherwise remained quiet about the incident.
“I get to make this decision, like, once. This is a big decision and if you don’t get it right, it’s very hard to come back from that,” Bruno said in April. |
ULA acknowledged last week’s test, but didn’t provide any details about when, and how, it will make a decision on the engine for Vulcan. “Congratulations to the entire Blue Origin team on the successful hotfire of a full-scale BE-4 engine,” the company said in a statement. “This is a tremendous accomplishment in the development of a new engine.”
Bruno, in April, said he would take his time making a decision on the engine for Vulcan, even if BE-4 appeared to be in the lead. “I get to make this decision, like, once. This is a big decision and if you don’t get it right, it’s very hard to come back from that,” he said then. “So I’m going to take my time and listen to all these experts and stakeholders and then do it.”
It’s a choice since ULA has another option for Vulcan: the AR1 engine under development by Aerojet Rocketdyne. It uses a more conventional combination of liquid oxygen and kerosene—the same used by the Russian-built RD-180 that powers ULA’s Atlas V—and is being designed by a company with a long heritage of engine projects.
It’s, though, well behind the BE-4. The company announced in May it had performed a set of hotfire tests in May, but only of the preburner portion of the engine. The engine also completed a critical design review (CDR) around the same time.
“Completing the CDR is a significant milestone for the AR1 program. It means that we have finalized our design and confirmed that it meets the diverse set of operational requirements necessary for national security missions,” said Eileen Drake, president of Aerojet Rocketdyne, in a statement. “This important milestone keeps AR1 squarely on track for flight readiness in 2019.”
The problem for the AR1, though, is that it may be an engine without a customer should ULA go with the BE-4. SpaceX develops its own engines in-house, while Orbital ATK, which also has plans to develop a large launch vehicle in the same class as the Atlas and Delta, is leveraging its expertise in solid motors.
With funding from the Air Force supplementing internal funds, as well as a small amount from ULA, work on AR1 will continue, at least through ULA’s engine decision and quite possibly some time beyond.
And what about SpaceX? The company has been working on a large “methalox” engine of its own, called Raptor, intended for use on the company’s BFR reusable launch system. SpaceX announced last September the first test of Raptor, and those tests have continued in the year since.
“We already have now 1,200 seconds of firing across 42 main engine tests,” Musk said in a speech at the International Astronautical Congress (IAC) in Adelaide, Australia, last month. “We’ve fired it for 100 seconds. It could fire for much longer than 100 seconds. That’s just the size of the test tanks.”
However, the Raptor the company is developing is different from what he announced at last year’s IAC. Then, Raptor was designed to generate up to 3 million newtons of thrust. The revised engine, as described in his Australia talk last month, now maxes out at about 1.7 million newtons, making it less powerful than Blue Origin’s BE-4.
Musk addressed this earlier this month when he participated in an Ask Me Anything session on Reddit. “The engine thrust dropped roughly in proportion to the vehicle mass reduction from the first IAC talk,” he said, referring to the scaled-down design of the overall BFR system compared to the Interplanetary Transport System he discussed in 2016.
That change, he said, is linked to the need for engine-out capability when landing the BFR spaceship. “The difficulty of deep throttling an engine increases in a non-linear way, so 2:1 is fairly easy, but a deep 5:1 is very hard,” he wrote. “Granularity is also a big factor. If you just have two engines that do everything, the engine complexity is much higher and, if one fails, you've lost half your power.”
“The advantage of getting somewhere in 30 mins by rocket instead of 15 hours by plane will be negatively affected if ‘but also, you might die’ is on the ticket,” Musk wrote. |
He was optimistic in that discussion about scaling up the current version of Raptor being tested to the revised design. “Very simple to scale the dev Raptor to 170 tons,” he wrote. “The flight engine design is much lighter and tighter, and is extremely focused on reliability.”
That emphasis on reliability is particularly essential, he added, for SpaceX’s plans to use the BFR system for point-to-point transportation. “The advantage of getting somewhere in 30 mins by rocket instead of 15 hours by plane will be negatively affected if ‘but also, you might die’ is on the ticket.”
SpaceX, like Aerojet Rocketdyne and other companies, is getting support from the Air Force to develop Raptor. Last week, the Air Force announced it was modifying an existing agreement with SpaceX, adding nearly $41 million to an award made last January to support Raptor development. Neither SpaceX nor the Air Force elaborated on what the additional funding would support.
It is, however it shakes out, a dynamic time for the domestic launch industry. Three new large engines are under development, with two already undergoing full-scale testing and the third set to begin such tests likely some time next year. (There’s also the qualification of the shuttle-era RS-25 engines ongoing by NASA and Aerojet Rocketdyne for use on the Space Launch System; while initial SLS flights will use shuttle-heritage engines, new RS-25 engines will eventually be manufactured for later SLS missions—assuming there will be SLS missions beyond the few that will use existing engines.) There’s no guarantee that all the engines will be successful, or that all will find a vehicle, but these efforts have made this part of the industry fascinating to watch.
Quelle: The Space Review