The NASA inspector general has concerns about the independence of contractor-led launch mishap investigations as demonstrated in the aftermath of recent mishaps by Orbital ATK [OA] and Space Exploration Technologies Corp. (SpaceX).
In a report released late Tuesday (IG-16-025), NASA IG said the space agency deviated from official policy for mishap investigations for the Commercial Resupply Services (CRS) program, of which SpaceX and Orbital ATK are contractors. NASA’s official policy requires all mishap investigation boards to be independent while the NTSB and the Air Force have similar requirements.
In contrast, the CRS program and FAA license requires SpaceX to conduct its own investigation, but does not require company investigation boards to screen for conflicts of interest or maintain independence. NASA IG said FAA officials stated NASA can implement additional independence requirements for contractor-led investigations through its contracts as long as they do not conflict with FAA regulations. For its mid-2015 launch failure, NASA IG said SpaceX’s investigation board featured employees as 11 of 12 voting members and that only the 11 SpaceX board signed the final accident investigation report.
NASA and SpaceX officials told NASA IG that specific expertise in the failed launch vehicle is important for an accident investigation board and that this factor should be taken into consideration, even though it may impact the board’s independence. NASA IG agreed, but said other processes such as NASA’s anomaly engineering board or the Air Force’s engineering analysis group are available to obtain this expertise without compromising the independence of an investigation board.
NASA’s official investigation policy allows for the creation of a non-voting anomaly engineering board to examine technical and factual issues. The Air Force also creates a non-voting engineering analysis group to assist its safety investigation boards. NASA’s Launch Services Program (LSP) told NASA IG it had access to all required technical expertise for its analysis of the 2015 SpaceX failure.
NASA created its own investigation boards for both the Orbital ATK and SpaceX launch failures to independently verify and review the contractor investigations. While the three stakeholders have similar incentives to safely and quickly return to flight, NASA IG believes the structure of the contractor-led investigations may not result in a full review of all programmatic and organizational contributing factors and, consequently, these factors may not be fully addressed to prevent future failures. NASA IG said NASA lost $118 million in cargo in SpaceX’s 2015 failure.
Commercial space advocate and industry consultant Rand Simberg said Wednesday there are two ways to look at NASA IG’s concerns. He said, in theory, contractor-led investigations could be a conflict of interest. But Simberg said, in reality, the contractor has a lot at stake in making sure it gets the mishap issued fixed correctly. Simberg also said it is worth considering that NASA is a government bureaucracy with its own institutional biases and reasons to do things.
“NASA is not going out of business if it screws up,” Simberg said. “If SpaceX messes up, it could go out of business.”
Simberg said there is a difference between SpaceX, Orbital ATK and the accident investigation boards from the Challenger and Columbia disasters. NASA managed Challenger and Columbia and had a lot of input into those investigations. But NASA isn’t managing CRS, he said, they’re purchasing services from contractors.
NASA IG said in 2015 (IG-15-023) that four of seven members on Orbital ATK’s mishap investigation board were company employees, including the chairperson, who was a former manager of the Taurus II and Antares programs and was also involved in risk analysis and testing decisions for the Antares and AJ-26 engine that caused the failure. Of the other three board members, two were current NASA employees, one who had responsibility for the risk analysis used to approve the 2014 CRS launch and for reviewing the company’s return to flight plan for technical feasibility.