The Space Reviewin association with SpaceNews

Antares failure
An explosion erupts from the aft end of the Antares rocket about ten seconds after liftoff October 28 from the Mid-Atlantic Regional Spaceport in Virginia. (credit: NASA/Joel Kowsky)

The importance for commercial spaceflight to recover and respond

Bookmark and Share

Every mode of transportation suffers failures. Planes crash, cars collide, trains derail, ships capsize. We even trip over our own two feet when walking. While we can, and should, take measures to reduce the chances of such accidents, and make them more survivable, no system is perfect. How we respond to, and recover from, an accident is critical.

Commercial space transportation is no exception to this. Every year, there are a handful of launch failures worldwide, both government and commercial. Many of them are relatively obscure and not photogenic: an upper stage misfiring, for example, stranding a satellite in a useless or unstable orbit.

The two accidents are linked by proximity in time, but in reality little else. Yet, there’s the danger that they will be tied together by the public, and by policymakers.

Others are not, like a rocket exploding on the launch pad. Still others are even worse, involving injuries and even death. The commercial space industry suffered that double whammy this past week, perhaps one of the worst weeks in recent memory for commercial space endeavors. The failure of an Antares rocket, carrying a Cygnus cargo spacecraft, on October 28 was still making headlines three days later when Virgin Galactic’s SpaceShipTwo crashed during a test flight, killing one pilot and injuring the other.

The two accidents are linked by proximity in time, but in reality little else. Yet, there’s the danger that they will be tied together by the public, and by policymakers, by the fact that they were commercial, despite the otherwise vast differences. Already some people have raised questions about the capabilities of commercial space in general, for example, from these accidents. How the industry responds to and recovers from these accidents will shape those public and policy reactions in the months to come.


On the evening of October 27, many space enthusiasts were mad. Some had tuned in to NASA TV to watch the launch of an Antares rocket carrying a Cygnus cargo spacecraft. Others had gone outside to watch the launch directly: the twilight launch promised to be visible over much of the Eastern US. However, the launch never happened, for reasons having nothing to do with the rocket, its payload, or the weather. Instead, a sailboat had wandered into a restricted zone about 70 kilometers offshore, and did not leave before the ten-minute launch window closed.

Twenty-four hours later, though, that collective animus towards that anonymous wayward boat was forgotten. Once again, people tuned in to watch the launch, again without any rocket, payload, or weather issues, and this time without any range issues as well. At 6:22 pm EDT, the Antares rocket ignited its main engines and lifted off the pad at the Mid-Atlantic Regional Spaceport, the commercial launch site located at NASA’s Wallops Flight Facility in Virginia.

About ten seconds after liftoff, though, things went horribly wrong. Video of the launch showed the rocket’s plume appeared to brighten, followed by an explosion at the base of the vehicle. The rocket fell back to earth, creating an even larger explosion.

The accident destroyed the Cygnus spacecraft, which was carrying 2,290 kilograms of cargo for the station: a mix of food and other supplies for the crew, hardware for the ISS itself, and experiments and other payloads, including 29 small satellites—26 from Earth imaging company Planet Labs—that were to be deployed from the station.

At a briefing immediately after the accident, NASA played down the effect the loss of the vehicle would have on ISS operations. “We’re in good shape from a consumables and supplies standpoint,” William Gerstenmaier, NASA associate administrator for human exploration and operations, said. “There was no cargo that was absolutely critical to us that was lost on this flight. The crew is in no danger.”

The station, NASA officials said, has four to six months of supplies on hand at the time of the launch failure. A Progress spacecraft carrying supplies for the Russian segment of the station launched as scheduled within hours of the failure. The next US cargo supply mission is a Dragon cargo flight scheduled for launch no earlier than December 9.

“It is possible that we may decide to accelerate this change if the AJ26 turns out to be implicated in the failure,” Thompson said.

With the ISS not in immediate danger from the loss of the Cygnus, attention turned instead on the cause of the accident. Speculation focused on the AJ26 engines that power the Antares first stage. Those engines were originally built decades ago in the former Soviet Union for the N-1—the Soviet rival to the Saturn V—and designated NK-33. Aerojet Rocketdyne acquired some of those engines, refurbished them, and now offers them as the AJ26.

The AJ26 has not been without problems, though. One engine failed in a 2011 test firing at NASA’s Stennis Space Center, a problem traced to a fuel leak in that engine. In May, another AJ26 engine failed on the test stand at Stennis. Neither Aerojet Rocketdyne nor Orbital Sciences have disclosed what caused that most recent test failure.

“We have come up with probably two potential root causes, both of which we can screen for,” Orbital executive vice president Frank Culbertson said in a September 30 presentation at the International Astronautical Congress in Toronto. He did not disclose what those root causes were.

The AJ26 engines flown on this Antares had passed a series of tests leading up to the failed launch. “We didn’t see any anomalies or anything that would indicate there were problems with the engine,” Culbertson said at the briefing immediately after the launch failure.

As this mission was being prepared for launch, Orbital Sciences was making plans to change the engine. In a quarterly conference call with financial analysts on October 16, Orbital CEO Dave Thompson said the company had selected a new engine, but declined to name it. Prior to the launch failure, company officials had suggested they would make an announcement after the company submitted a proposal to NASA in mid-November for a new contract to delivery cargo to the ISS. Speculation has ranged from a solid motor from ATK, with whom Orbital is merging, to a derivative of the RD-180 engine from Russian manufacturer NPO Energomash.

The day after the accident, with Orbital’s stock price declining by more than 15 percent, Thompson again talked with financial analysts. “It is possible that we may decide to accelerate this change if the AJ26 turns out to be implicated in the failure,” he said. Prior to the accident, he said the company expected to have a “second-generation” Antares, with that new engine, ready for its first launch in two years. Thompson said it may be possible to accelerate that schedule, but declined to say by how much.

One bit of good news from the accident was that the Antares launch pad at MARS appeared to escape major damage. “The overall findings indicate the major elements of the launch complex infrastructure, such as the pad and fuel tanks, avoided serious damage, although some repairs will be necessary,” Orbital said in an October 29 statement.

There was damage, the company said, to the transporter erector that cradles the rocket in the horizontal position when moved to the pad, then erects it vertically. Photos showed that some of the lightning rods around the pad had fallen. The company said there was damage to pipes that run from fuel tanks near the pad to the launch mount, although the tanks themselves appeared to be unharmed.

As this article was being prepared for publication Monday, Orbital announced the members of its Accident Investigation Board for the Antares failure, led by David Steffy, chief engineer of the company’s advanced programs group. The board consists of Orbital and NASA officials, as well as former Space Shuttle program manager Wayne Hale. The FAA’s Office of Commercial Space Transportation has oversight of that investigation.

Last week, Thompson said he hoped it would be “days, not weeks” until the company narrowed down the most likely causes of the failure, although the full investigation would take longer. Orbital had planned to launch Antares again on another resupply mission next April, but Thompson said that launch would be delayed by at least three months. “It certainly could be considerably longer than that, depending on what we find in the review,” he said. “I hope it would be not more than a year.”

NTSB investigation
NTSB investigators and Virgin Galactic pilot Todd Ericson examine debris from SpaceShipTwo on November 1, a day after the vehicle’s fatal accident. (credit: NTSB)


While Orbital’s plans to launch Antares were known far in advance, Virgin Galactic and Scaled Composites did not broadcast in advance their plans to perform the next test flight of SpaceShipTwo. It was only Friday morning, as the WhiteKnightTwo aircraft, with SpaceShipTwo attached between its twin fuselages, taxied out at the Mojave Air and Space Port did Virgin announce it was conducting another test.

What alarmed Witt was not seeing and hearing what should have been taking place on a normal flight. “It wasn’t because something did happen, it was what I was not hearing and not seeing. If there was a huge explosion, I didn’t see it.”

Many in the industry expected the next SpaceShipTwo test flight to be the first powered flight since early January. In early October, company officials said they were performing the final qualification test firings on the ground of a revised hybrid rocket motor, after Virgin announced in May it was switching from a rubber-based fuel to a plastic-based one. In mid-October, company CEO George Whitesides said at the International Symposium on Personal and Commercial Spaceflight in Las Cruces, New Mexico, that those final tests had been completed.

After what the company said was a weather-based delay, SpaceShipTwo and WhiteKnightTwo were finally airborne and, shortly after 10 am PDT, SpaceShipTwo was released and ignited its rocket motor. Then came this ominous statement from the company, via Twitter: “SpaceShipTwo has experienced an in-flight anomaly. Additional info and statement forthcoming.”

While it was almost instantly clear what happened to Antares, SpaceShipTwo’s failure played out in slow motion over the next several hours, as reports—often conflicting, as is the case in breaking news events—came in of debris falling in the desert north of Mojave, and parachutes spotted. Had the two pilots managed to bail out of the vehicle? And just what happened?

By late in the day, officials said one of the two pilots had died in the accident, while the other had been transported to the hospital with “major” injuries. On Saturday, Scaled Composites announced that Michael Alsbury, the 39-year-old co-pilot of SpaceShipTwo, had died in the accident. Peter Siebold, the 43-year-old director of flight operations for the company and the vehicle’s pilot, was hospitalized, the extent of his injuries unknown.

Immediately after the accident, speculation focused on the vehicle’s rocket motor. It had, after all, been the subject of lengthy development delays, including the decision earlier this year to change the fuel it uses. (Sierra Nevada Corporation, which developed the earlier hybrid motor with the rubber-based fuel, made it clear in a statement issued late Friday that it had no involvement with this flight.) In 2007, three Scaled employees died, and three others were injured, in a ground test of the nitrous oxide injector for a hybrid rocket motor.

However, even in the hours after the accident, it was clear that there had not been a massive explosion in the engine. “I detected nothing that appeared abnormal,” said Stu Witt, chief executive of the Mojave Air and Space Port, at a Friday afternoon press conference. What alarmed him, he said, was not seeing and hearing what should have been taking place on a normal flight. “It wasn’t because something did happen, it was what I was not hearing and not seeing. If there was a huge explosion, I didn’t see it.”

Without hours of the accident, the National Transportation Safety Board (NTSB) announced it was sending investigators to Mojave to study the accident. The NTSB would take the lead in the investigation under an agreement with the FAA, marking the first time the NTSB, best known for investigating airplane crashes and other transportation accidents, was leading the investigation of a human spaceflight accident.

That investigation started Saturday, with investigators examining debris that stretched for about eight kilometers, telemetry from the flight, and performing interviews of various people. (Those interviews have not included, as of late Monday, Siebold, who NTSB officials said was not yet medically able to participate in an interview.) NTSB’s acting chairman, Christopher Hart, said that since this was a heavily-instrumented test flight, they had access to more data than usual for a crash investigation.

“The engine burn was normal up until the extension of the feathers,” Hart said.

That additional data helped NTSB quickly rewrite the narrative of the accident. In a press briefing Sunday night, Hart said investigators were turning their attention to SpaceShipTwo’s feathering system. That system raises the vehicle’s twin tail booms during reentry in order to increase drag and improve stability. To engage that system, pilots must first turn one handle to unlock the booms, then turn another to raise them.

Hart indicated Sunday night that the co-pilot, Alsbury, may have prematurely unlocked the feathering system. “About nine seconds after the engine ignited, telemetry data told us, showed us, that the feather parameters changed from ‘lock’ to ‘unlock,’” he said. There was no indication in the data that the pilots deliberately engaged the feather.

A camera in the cockpit of SpaceShipTwo confirmed the telemetry. “Review of that camera is consistent with that telemetry data, and shows that the feather lock-unlock lever was moved, by the co-pilot, from the lock position to the unlock position,” he said. (In a briefing Monday night, Hart appeared to indicate that they could not confirm that it was indeed co-pilot Alsbury who unlocked the feathering system, but NTSB stated shortly after the briefing that it was the co-pilot who did so.)

Unlocking the feather at that point in time was at least several seconds premature. Hart said normal protocol was to unlock the feather at Mach 1.4, but SpaceShipTwo was travelling at about Mach 1.0 at the time the feather was unlocked.

Once the feather was unlocked, the booms extended even without actions by the crew. “Approximately two seconds after the feathering parameters indicated that the lock-unlock lever was moved from ‘lock’ to ‘unlock,’ the feathers moved towards the extended position, the deployed position, even though the feather handle itself had not been moved,” Hart said.

Hart declined to speculate on whether that caused the loss of the vehicle, saying he was giving “a statement of fact, not a statement of cause.” Many observers, though, noted that extending the booms as the vehicle was accelerating through Mach 1 could create aerodynamic forces strong enough to break the vehicle apart.

As for the much-maligned engine, it appears to be exonerated, at least for now. Hart said that investigators had recovered the engine and propellant tanks, which showed no signs of problems. “All were intact, showed no signs of burn-through, no signs of being breached,” he said. “The engine burn was normal up until the extension of the feathers.”

Respond and recover

In the hours and days since the SpaceShipTwo accident, various individuals and organizations offered a common message of condolences to the families of the pilots, and hopes that Virgin Galactic will be able to recover from the accident.

While an entirely commercial mission, even NASA administrator Charles Bolden weighed in. “While not a NASA mission, the pain of this tragedy will be felt by all the men and women who have devoted their lives to exploration,” he said in a statement issued late Friday. “Space flight is incredibly difficult, and we commend the passion of all in the space community who take on risk to push the boundaries of human achievement.”

There were a number of rapid reactions in the media, in the form of op-eds and interviews, immediately after the accident that spanned the spectrum of thought, from those who argued that Virgin Galactic should proceed full speed ahead to those who concluded the company, and even space tourism in general, was doomed. However, such instant commentaries rarely have any lasting value: those that, for example, blamed the accident on the company’s troubled engine development effort have been forced to backtrack once the NTSB said there appeared to be no issued with the vehicle’s engine on this flight.

The same applies to the commentaries that attempted to link the two accidents, and thus conclude the commercial space transportation was therefore inherently unsafe or otherwise not viable. While that temptation to link the two is understandable, given that the two accidents took place less than 72 hours apart, there’s little else the two events have in common.

“The thing to do is to acknowledge the challenges we’re facing and make sure we’re doing everything possible to mitigate those challenges using the best technologies, the best systems, the best approaches,” said Tumlinson.

Antares was an orbital rocket; SpaceShipTwo a suborbital spaceplane. Antares used refurbished liquid-propellant engines originally manufactured decades ago in the former Soviet Union; SpaceShipTwo used a hybrid rocket engine developed internally. Orbital Sciences has been in business for three decades, with extensive experience in launch vehicles and satellites; Virgin Galactic has been around a decade, with SpaceShipTwo its first project, although its partner Scaled Composites does have its experience with SpaceShipOne a decade ago. The Antares failure appears to be closely linked to its main engine; SpaceShipTwo’s failure no longer appears closely linked to its main engine. And so on.

What the two accidents do demonstrate is that any company or organization in space field—government and commercial—can suffer an accident. Whether it’s an inherent design flaw or an oversight in operations, space systems often have little tolerance for failure. That makes the track records of organizations (including commercial space ventures) with long track records of success, like Arianespace and United Launch Alliance, all the more remarkable.

In the weeks to come, there will be more scrutiny placed on commercial space ventures, including by politicians, who will ask if companies and government regulators are doing enough to prevent accidents like these, and if commercial space ventures can be trusted to carry out key missions for agencies like NASA, or carry people for the sake of experiencing spaceflight. How the industry responds to such scrutiny—how it recovers and responds—may make all the difference in how politicians and regulators react to these accidents.

“The thing to do is to acknowledge the challenges we’re facing and make sure we’re doing everything possible to mitigate those challenges using the best technologies, the best systems, the best approaches,” said Rick Tumlinson, a long-time space advocate who has been involved in a number of commercial space ventures, most recently space resources company Deep Space Industries, in an interview Saturday.

“Let’s not make the same mistakes that led to things like Challenger,” he added. “We need to assume that things can’t work until we can prove they can, especially when humans are in the loop.”

It may be that Orbital Science and Virgin Galactic bounce back quickly, resuming flights of Antares and SpaceShipTwo as fast as the technology permits. Or, these vehicles may never fly again. If the latter, it’s not the end of commercial spaceflight, of course, with other ventures, both suborbital and orbital, also developing or operating similar vehicles.

However, if it’s the former, remember: last week will not be the last time that a commercial space vehicle has an accident, with or without people on board. Just as with cars, planes, ships, and other modes of transportation, there will be launch failures and spaceship accidents in the future. And, once again, how the companies involved, and the industry in general, recovers and responds will be critical.