Letter: Difference Between Reliability and Safety

by

While I was not able to attend the George Washington University Space Policy Institute’s Sept. 28 workshop on the findings of the Review of U.S. Human Space Flight Plans Committee, I did see the charts presented by Commercial Spaceflight Federation President Brett Alexander, and must take specific objection to the use of the following quote, deliberately taken out of context: “Mike Griffin, in testimony to Congress in 2003: ‘What, precisely, are the precautions that we would take to safeguard a human crew that we would deliberately omit when launching, say, a billion-dollar Mars Exploration Rover (MER) mission? The answer is, of course, ‘none.’”

My intent with this quote was to address the issue of reliability of various launch vehicles, when considering the difference between human and robotic missions. My point, then and now, is that all existing launch vehicles are designed to be as reliable as we can make them.

The stakes are so high — for both human and robotic missions — that there is nothing we would do to increase the reliability of a launch vehicle merely because we intended it to be used for a human mission.

We already do all we can.

This, however, is only part of the picture. For human missions, the issue of safety must also be considered, and is in fact paramount. A safe launch system is, of course, generally reliable; this goes without saying. But to be safe, a launch system must also offer a highly reliable abort mode for those occasions when the statistics on reliability turn up the wrong card. Substantial modifications for any existing non-human-rated launch system will be required to enable a reliable abort mode throughout the ascent profile, and substantial design differences will exist in any clean-sheet design for a human-rated system, as opposed to a system designed only for cargo. These facts are known to all who have worked in the field of manned spaceflight. To gloss over them deliberately in order to make a point with the hope of somehow winning an important argument should be beneath us.

It is of course true — as I have stated many times — that existing launch vehicles such as Atlas and Delta can certainly be modified and/or upgraded to carry crew, and to do so with some measure of improved safety, especially relative to the space shuttle today. However, careful study of this issue while I was at NASA showed that modification of existing vehicles would in the end yield systems that were both more expensive on a lifecycle cost basis and less safe than the selected Ares 1 design. No fact brought forth since the Exploration Systems Architecture Study effort has altered this conclusion.

Finally, I will state for the record, as I have done many times, that the above conclusion — that Ares 1 offers the cheaper path for human access to low Earth orbit (LEO) by NASA — holds true only if one is considering a space launch architecture to include interplanetary human missions, rather than a point-design for LEO access. If one wishes only to replace the shuttle as a means of providing human access to LEO, and to omit consideration of any beyond-LEO future for human spaceflight, then without doubt a cheaper system than Ares 1 and Orion can be found.

However, I continue to be on record with my belief that such a future is unworthy of this nation.

 

Mike Griffin

Professor, University of Alabama, Huntsville

NASA Administrator from April 2005 to January 2009