Thursday, October 29, 2015

The Orbital Post-Mortem

Click the arrow to watch replays of the Orbital Antares ORB-3 launch and failure on October 28, 2014. Video source: SpaceKSCBlog YouTube channel.

A year and a day after an engine failure destroyed a commercial cargo delivery, NASA issued a press release and an executive summary of its investigation findings.

Orbital Sciences was one of two companies delivering cargo to the International Space Station under a Commercial Resupply Services contract. Orbital's Cygnus module launched atop the company's Antares rocket, boosted by Aerojet-Rocketdyne AJ-26 engines. Those engines were originally unused 1970s Soviet engines purchased by Aerojet and modified for Orbital's use.

Orbital merged in February 2015 with ATK to become Orbital ATK.

The press release states:

The team determined the proximate cause of the Antares launch vehicle failure was an explosion within the AJ-26 rocket engine and identified three credible technical root causes, any one or combination of which could have resulted in the engine failure. The team outlined six technical findings and made seven recommendations to address those technical findings. In addition, since Orbital ATK was in the process of procuring and testing new engines to replace the AJ-26 for future Antares flights while the investigation was ongoing, the team provided several recommendations for Orbital ATK and the ISS Program that were used to support those testing activities and to reduce overall risk for Antares return-to-flight and follow-on mission efforts.

Apparently NASA released only the executive summary because their findings refer to proprietary features of the Orbital engine and rocket. The executive summary mentions in several places that it cannot provide details of their findings due to proprietary information.

According to the summary, the NASA Independent Review Team (IRT) “determined that the proximate cause of the Antares launch vehicle failure was an explosion within the AJ26 rocket engine installed in the Main Engine 1 position.”

Specifically, there was an explosion in the E15 Liquid Oxygen (LO2) turbopump, which then damaged the AJ26 rocket engine designated E16 installed in the Main Engine 2 position. The explosion caused the engines to lose thrust, and the launch vehicle fell back to Earth and impacted the ground, resulting in total destruction of the vehicle and its cargo ...

The IRT also developed a detailed system-level fault tree, timeline of events, and failure scenarios, and performed analysis and forensic investigation of the hardware recovered from the accident. The IRT concluded that the cause of the explosion on launch was loss of rotor radial positioning resulting in contact and frictional rubbing between rotating and stationary components within the Engine LO2 turbopump Hydraulic Balance Assembly (HBA) seal package. This frictional rubbing led to ignition and fire involving LO2 within the turbopump HBA. This conclusion is consistent with the proximate cause determination made by the Orbital ATK [Accident Investigation Board] investigation findings.

The report lists three possible “technical root causes” but cannot say for certainty if one or more caused the loss of rotor radial positioning:

  1. “... [T]he HBA and thrust bearing designs have several intricacies and sensitivities that make it difficult to reliably manage bearing loads. As a result, this area of the turbopump is vulnerable to oxygen fire and failures. The AJ26 engines were not subjected to a thorough delta-qualification program to demonstrate their operational capability and margin for use on Antares. Performing a thorough delta-qualification program for Antares would likely have revealed these issues.”
  2. “Forensic investigation identified the presence of both titanium and silica [Foreign Object Debris] within E15 prior to its impact on the beach. However, no firm conclusions can be drawn with respect to the quantity of FOD introduced to or already present within the engine prior to or at the time of the explosion. The lack of significant particle impact damage to the recovered impeller and other components indicates that there were not gross-levels of FOD present within the system. In addition, there is no clear forensic evidence that FOD directly or indirectly led to the E15 failure.”
  3. “Forensic investigation performed by Orbital ATK and NASA discovered the presence of a defect on the turbine housing bearing bore that was not consistent with baseline design requirements. The investigation determined that the defect was introduced during machining of the bearing bore housing and was therefore present prior to the engine [Acceptance Test Program] and Antares launch for Orb-3. Forensic investigation of Engine E17, which failed during ATP in May 2014, discovered the presence of a similar non-conforming defect in the housing bearing bore.”

The report lists seven technical recommendations. The first is that “NASA should not rely on the AJ26 for further missions without undertaking a more thorough inspection ...” That's not a concern, because Orbital ATK has ordered new RD-181 engines from Russian supplier NPO Energomash.

A redesigned Antares with Russian-made RD-181 engines is scheduled for a pad test fire in early 2016. Image source: NASA.

Orbital ATK issued a press release today stating the company plans a “hot fire” test in early 2016 for the redesigned Antares booster with two RD-181 engines.

In order to accommodate the RD-181 engines, several modifications were made to the Antares vehicle design. Upgrades completed and installed over the summer include a newly-developed and qualified thrust adapter structure, modifications to the first stage core propellant tanks, modified engine control avionics and new propellant feedlines.

In addition to the work being conducted for the hot fire test, the team is also receiving hardware for the full Antares rocket that will be used for its return to flight in the spring of 2016. The crew will be assembling that vehicle in parallel with the hot fire test vehicle. The next set of Antares engines recently completed acceptance testing and is scheduled to arrive at Wallops in December.

The summary also recommends cultural changes at both Orbital ATK and NASA suggesting that pre-launch processing was inadequate by both.

Technical recommendations call for Orbital to do a more thorough engine acceptance test, including more sensors, ensuring cleanliness, and “a more robust and verifiable moisture barrier approach.”

Three programmatic findings fault the relationship between Orbital and Aerojet-Rocketdyne. The report cites the “lack of an integrated partnership” between the two companies, which created “a low level of confidence in loss-of-mission predictions.”

In September, Aerojet agreed to pay Orbital $50 million and retake possession of Orbital's remaining ten AJ-26 engines. The agreement effectively ended any claims the two companies had against one another because of the accident.

The executive summary also states that the “management and risk model established for commercial services is well understood and embraced by and at all levels of the NASA team and community.” But the review team cited four “opportunities for improvement,” a polite way of pointing four metaphorical fingers.

The team concluded that “there is clear and consistent understanding and acceptance of the CRS risk approach throughout NASA and by the Service Providers,” but “[t]he perception of acceptable risk for any given CRS mission varied significantly within and between NASA organizations.”

The Launch Vehicle Assessment (LVA) as currently developed may provide a false sense of security in the overall risk associated with the launch vehicle design and a particular mission ... The proprietary nature of launch vehicle information may be serving as an artificial barrier to communications and leading to communication shortfalls.

The report recommends:

The ISS Program should reassign LVA management responsibility to a senior engineer at the Marshall Space Flight Center (MSFC) with significant launch vehicle development and assessment experience, as well as increase the number of individuals from MSFC engineering supporting development of the LVA.

In other words, Big Brother should be watching more closely.

Releasing the report today has a curious timing beyond just the one-year anniversary of the incident.

In April 2014, NASA requested proposals from commercial vendors for a second round of ISS cargo delivery services.

It's believed that incumbents Orbital ATK and SpaceX submitted bids for Commercial Resupply Services Round 2 (CRS-2). Other candidates are Boeing with a cargo version of their CST-100 Starliner commercial crew vehicle, Sierra Nevada Corporation with a cargo version of the Dream Chaser spaceplane, and a Lockheed Martin system called Jupiter.

A promotional video released October 7 of the cargo Dream Chaser. Video source: SNCspacesystems YouTube channel.

According to a CRS-2 procurement web site, NASA will announce the new contract winners on Thursday November 5.

With both Orbital and SpaceX suffering accidents in the last year, it would seem the other candidates have a decent chance of winning a contract.

Even if Orbital ATK fails to win a new contract, they will still fly the remaining flights under their CRS-1 agreement. In August, NASA ordered two more Cygnus deliveries under CRS-1, which would take the existing contract through 2018.

It's strictly my speculation, but I think NASA may issue contracts to as many as three or four vendors. The last year demonstrated that two vendors are inadequate for redundancy. It would be in the agency's interest to grow the market with more delivery options.

The LockMart Jupiter appears to be a paper exercise, so that would seem to be the least likely to win a contract.

Desperately needing a contract is SNC, which is still fulfilling milestones for earlier rounds of a commercial crew contract. Dream Chaser failed to win any crew flights when those contracts were issued in September 2014. The spaceplane offers NASA an option other than SpaceX for returning cargo to Earth, and it's designed to land at any international airport runway. SNC has proposed delivering cargo to runways near NASA space centers in Houston and Huntsville, as well as Kennedy Space Center.

The crew version of the Boeing CST-100 will use parachutes and airbags to land in the U.S. southwest desert, so presumably the cargo version would do the same.

My worthless prediction is a few more cargo flights for SpaceX and Orbital ATK, with some flights circa 2020 going to Boeing and Sierra Nevada.

Check back on November 5.

UPDATE October 29, 2015 9:45 PM EDTJeff Foust of Space News reports he was able to get a copy of Orbital's internal report, and they reached a somewhat different conclusion on the root cause.

The Orbital report, though, identified a single “highly probable” technical root cause of the failure: a machining defect in turbine assembly of the turbopump that dates back to when the engine was manufactured in the former Soviet Union more than 40 years ago. An adjacent piece of the engine, recovered with little damage after the explosion, showed a “clear defect” in its machining, according to the report.

Orbital's report also identified several other technical root causes it deemed “credible” but less likely to have caused the explosion. Those additional causes included other technical problems with the engine and “poor long-term storage” of the engine, causing corrosion of engine components.

At first glance, it sounds like Orbital ATK is trying to put the blame on Aerojet-Rocketdyne.

No comments:

Post a Comment