It’s been a tragic week for the commercial space industry. With the media primed from the spectacular fireball of the unmanned space station resupply mission ORB-3, tragedy struck in Mojave with the in-flight breakup of Virgin Galactic’s SpaceShipTwo on its third powered flight, killing one test pilot. This marks the first in-flight fatality of the burgeoning space tourism industry, though not the only fatality in the SpaceShipTwo program – an explosion during engine testing on the ground killed three test engineers in 2007.
Initial reports from the NTSB indicate the feather locking mechanism was moved by the pilot from “lock” to “unlock” prematurely. The mechanism was unlocked in a region of high aerodynamic loading and two seconds later the vehicle was destroyed by aerodynamic forces. By my back-of-the-envelope calculation, the operator unlocked the feathering mechanism about five seconds before called in the procedure. The unlock mechanism, pilot, and procedures are now the focus of the NTSB investigation.
The NTSB findings were a shock to the amateur space community who assumed the accident was caused by the rocket engine – the long awaited first test flight of the vehicle with an upgraded rocket engine. The engine has had a dubious development history and has been a constant source of external ridicule for Virgin. Pilot error championed the news outlets for obvious reasons – a simple, superficial explanation for the public and partial exoneration of design flaws in Virgin’s vehicle. Richard Branson, the figurehead and bankroller, spouted the all-too-common space industry platitude: “Space is Hard”.
Well, space IS hard. Development and operations is hugely capital intensive, the operation environments are extreme, and things generally have to work right the first time. But these are financial and engineering challenges bounded by interest rates and physics – comforting to an engineer.
Virgin states “safety is our guiding principle and the North Star for all programmatic decisions,” and while this sounds reassuring, it is not a tenable requirement to design a vehicle. This “North Star” needs to be assessed as an engineering risk, and the results fed back to design. The risk assessment process has once (Challenger) again (Colombia) proven to be broken. Perhaps it’s because the answers don’t lie in integrals and probability distributions, but in the ego of the human mind.