February 05, 2003
Aircraft and Space Shuttle Accident Rates

We have now lost both the Challenger and the Columbia. That's 40% of the Shuttle fleet. It's time to seriously reexamine the US space program. Should the Shuttle continue to be operated? Should a new kind of shuttle be designed? What should be the criteria used to answer these questions? The debate about the future of the US Space Shuttle should be a debate about how we can make space travel much safer, more reliable, and lower in cost. These are interrelated goals. Unreliable launchers and passenger carrier spacecraft are more likely to be lost. Loss of a launcher is both fatal for the crew and incredibly costly. Higher reliability equipment is safer and less costly to maintain.

Let's compare aircraft safety to Space Shuttle safety.

The 1995 fatal accident rate per million miles flown for these large scheduled airlines declined to 0.0004 from 0.0008 the year before. Based on 100,000 departures, the fatal rate was 0.024, down from 0.050 in 1994.

Scheduled commuter or regional airline fatalities dropped to 9 persons from 25 in 1994 for the lowest level since 1990. The fatal accident rate fell both in terms of million miles flown to 0.003 from 0.005 in 1994, and from 0.083 to 0.057 in terms of 100,000 departures. It was the fourth consecutive annual decline in the fatal accident rates.

In 1995 the fatal accident rate per 100,000 departures for the airlines flying the smaller aircraft was 0.057 and impressively was less than half that for the big jets operated by the majors. Assuming they are talking about number of accidents and not how many people died in each accident (anyone know?) then to compare that to the Space Shuttle record we compare the 2 fatal Space Shuttle accidents out of 113 flights. That works out to 1770 fatal accidents per 100,000 flights. 1770 for the Shuttle divided by 0.057 for smaller commercial aircraft works out to an accident rate that is over 31,000 times greater for the shuttle than for small craft commercial aviation. When compared to large craft commercial aviation using 1995 again as a comparison point (note that there is fluctuation from year to year because there can be clusters of accidents in a year just by chance - but the trend in commercial aviation is toward ever lower accident rates) we see that the Shuttle is over 73,000 times more dangerous.

James Dunnigan has a table showing failure rates of launchers that have been used more than 100 times each. The US Space Shuttle has a lower failure rate than the other launchers. The failure rates range from 5% for the Russian R-7 Soyuz and European Ariane 1-4 to 14% for the US Atlas. In his article Dunnigan argues that the International Space Station (ISS) is the major justification for the US Space Shuttle. But before we get to that let's think about what these failure rates mean.

If the best space launch vehicle in existence has an failure rate of 2% and the rest are worse this argues that achieving an acceptable level of spacecraft passenger safety can not be done by developing small incremental improvements to current launch vehicle technology. One option is to design passenger carrying spacecraft to integrate with the launchers in a way that allows the spacecraft passengers can survive launch failures. Such a technology was built into Apollo (the Apollo Escape Tower for pulling the CM away from a failing rocket). Another, and not mutually exclusive approach, is to develop a launcher technology that is inherently more reliable than current technology. We can't expect to implement either of these approaches with the Space Shuttle. Technologies that would offer greater than order-of-magntitude improvements in safety and reliability can not be retrofitted into existing designs. Limitations inherent in the original design of the Space Shuttle makes it totally inappropriate as a target for attempts to make big strides in the reliability and safety of human space travel. We need to start over from scratch.

In 1950 there were 2,482 thousand aircraft departures, 19,102,905 passengers carried, and 6 fatal accidents. In 1997 there were 8,157 thousand departures, 598,895,000 passengers carried, and 3 fatal accidents. Fatal accidents per million aircraft miles flown dropped from 0.0126 to 0.0005. The number of fatal accidents per million miles flown was about 25 times greater in 1950 than in 1997. This is the standard against which spacecraft should be compared. The Space Shuttle is at least 3 orders of magnitude more dangerous than passenger aircraft from 1950. Could the aircraft in the fleet of 1950 have been continually modified to make them as safe as a passenger aircraft manufactured 20 or 30 years later? Of course not. Better design and fabrication techniques produced later designs that were inherently more reliable.

But let's go back even further to look at aircraft safety in 1938. That's when some US government agency was created that started tracking aircraft safety. It is not clear from the table what kind of fatal accident rate measure they were using. But compare the 1938 rate of 11.9 to the 1950 rate of 5.0 and the year 2000 rate of 1.1. The 1938 rate of fatal accidents was about an order of magnitude higher than it is now. But its still more than two orders of magnitude lower than the fatal accident rate of the Space Shuttle.

1938 was 35 years after the first aircraft flight of Orville and Wilbur Wright on December 17, 1903 at Kitty Hawk North Carolina. Manned space travel began on April 12, 1961 when a Soviet air force pilot, Major Yuri A. Gagarin, made an orbit of the Earth. So manned space travel is over 40 years old. Space travel into Earth's orbit is orders of magnitude more dangerous after 40 years than aircraft travel was when it was only 35 years old.

Aside: If anyone has aircraft safety data that goes back to the era of biplanes in WWI (leaving aside casualties from war) then please pass it along. It seems quite possible that aircraft safety was never as dangerous as spacecraft safety is now or it was only that dangerous for a relatively short period of time.

Is the safety of spacecraft travel going to improve? Don't look to NASA Space Shuttle contractor Boeing for leadership in spacecraft safety improvements.

"I expect the shuttle will fly another 20 years," said Rick Stephens, vice president and general manager of Boeing's Homeland Security and Services and Integrated Defense Systems.

Imagine another 20 years of space travel that is 4 orders of magnitude more dangerous than air travel. Boeing would be happy to keep getting paid to maintain the dangerous Space Shuttle for that long. Any reason to develop a more technologically advanced, cheaper, more reliable, and safer alternative? Why do that as long as the current dangerous unreliable obsolete system is generating a large revenue stream?

But Stephens, who has headed up operations in Boeing's Space and Communications Services and Reusable Space Systems, said he did not think the tragedy would speed up the search for alternatives to the shuttles.

How about at least saying that the latest tragedy should be a wake-up call that we should start working on a better design? The Shuttle should be treated as a means to an end rather than a glorious end in itself. That end should be the a continuing improvement in the ability to move humans into space. The Space Shuttle is irrelevant to that goal. An extremely dangerous, unreliable, expensive (all by the standards of the commercial aircraft industry of 65 years ago) launch system built with early 1970s technology that costs $500 million per trip is going to ensure that human presence in space remains a rarity.

Why have the Space Shuttle? What do we need it for? NASA says we need it for the International Space Station. But the International Space Station has been so scaled back in capabilities that it can do very little science. Without the ISS to give the Shuttle a purpose is the Shuttle worth operating?

The Shuttle has been used for upgrades to Hubble. But absent the Shuttle as a repair device a replacement for the Hubble could have been built. That would have cost more than than sending up the Shuttle to do a repair. But at $500 million per Shuttle mission its not cheap. The money saved by occasionally doing Hubble repairs and upgrades is hardly justification for keeping the Shuttle around. If no money was spent on the Shuttle at all then the amount saved could pay for many Hubbles.

There is also the age of the Shuttles from the standpoint of on-going maintenance. The fleet is way older than its designers expected it to get. Parts are hard to find.

The fleet - 22 years old - has now been flying for twice as long as its builders first envisioned. Some parts were made so long ago that they are no longer available. Shuttle engineers have had to turn to Internet auction site eBay for desperately needed hardware and electronics.

Where should we go from here? Science fiction writer Jerry Pournelle has a discussion going about Single Stage To Orbit (SSTO) vehicles. Jerry's message is pretty simple. Instead of operating old technology we should build lots of experimental designs to test out various concepts and see what works, what doesn't, and why.

Two stages to orbit, or one stage and a flyable zero which may well be a ring of jet engines, is another possibility: again the operations penalties are not insignificant. The operational penalties are not small: imagine if every time you wanted to fly across the Atlantic, you had to have a second airplane that did nothing but get your plane aloft. It may be required, but it's not desirable.

So: let me sum it up. We need to build more rocket ships. We need to fly more rocket ships. We need better data. These were conclusions we sent to the President in 1983, and repeated to a different President in 1989. They haven't changed. We need X programs. Real ones, not corporate welfare programs like the "X"-33.

What is more important? Is it more important to use expensive old technology so that we can have humans in space in the short to medium term? Or is it more important to experiment and try out lots of engineering experiments for different approaches for spacecraft designs? Do we want to innovate? Do we want to advance? For the amount of money that is going into keeping the old tech Space Shuttle going and to build an incredibly expensive low scientific value space station we could be designing and trying out many innovative spacecraft designs.

Our old space launch technology is woefully inadequate. Compared to aircraft technology it becomes clear just how unsafe, unreliable, and costly our space launch technology really is. Lots of incremental improvements to an old design will not get us very far. If we want to go into space in a serious manner then we need to admit our mistake in funding the old technology for as long as we have. It's time to move on. Its time to let go of the past. Start concentrating on finding the technologies we need for the future.

Update: Jay Manifold has a bit more info. Also, an earlier post of his points out the low science value of both the Space Shuttle and the ISS. This is something I intend to explore at greater length in future posts. Look at the NSF budget and what it buys in the way of scientific advance. The Jerry Pournelle link above has a quote that the NSF budget at 5 billion dollars per year is half what the Shuttle plus ISS are going to cost per year to operate (supposedly $10 billion - need to search out some details on this). The ISS is going to have one guy on it doing science part time. The amount of science that could be accomplished if that money was spent in other ways (e.g. just give it to the NSF and thereby triple the NSF's budget) could be enormous.

The ISS is not for science. Its purpose is to make people feel good that there are people up there working in space. It accomplishes very little beyond that except for giving aerospace companies multi-billion dollar contracts that stretch out for decades. If we want to do science in space then the money would be far better spent on unmanned probes and satellites. If we want to do human exploration of space then the money would be far better spent on developing next generation launch vehicles and also nuclear thermal propulsion systems. NASA's manned space budget of billions of dollars per year is yielding precious little in science and precious little in technological advances that could lower the cost and increase the safety of human space travel.

If it hasn't yet become clear that I think that the International Space Station and Space Shuttle are monumentally stupid wastes of money then let me make it clear: the International Space Station and Space Shuttle are monumentally stupid wastes of money.

Share |      Randall Parker, 2003 February 05 11:16 PM  Space Exploration

Bob said at February 6, 2003 10:53 PM:

After reading your post, I was going to suggest we dub the ISS the "Irrational Space Station". After thinking about it for a moment, though, I realised the substitution works as a broad generalization.

So far, every common phrase I can think of starting with "International" seems to work better with "Irrational" instead.

Alex Sheppard said at June 3, 2003 10:29 PM:

You're quite right about the structure of NASA and the corruption of the arrangements nowadays. The contractors have no interest in innovation and simply using the agency as a tit on which to suck. The politicians have not the will to change this arrangement for fear of upsetting powerful aerospace/defense/other interests.

Grant said at June 5, 2003 8:59 PM:

Mmmh - I agreew with the conclusions, but he mangles the facts. Like why compare unmanned launch failures when looking at fatal accidents (AFAIK, no one has died on Ariane ).

Also, when talking about old aircraft, he should remember that B52's and Hercules are 50 years old and still safely flying (and being improved in the case of a Herc).

I still don't think it is fair to compare a Shuttle safety record to something like a 737 though - the Shuttle has a far more difficult job to do. Its more fair to compare with a military combat aircraft from the same era - like a F-14. Even though no F-14's have been lost in combat, I think a lot have crashed.


Randall Parker said at June 5, 2003 9:34 PM:


The reliability records of the rockets are useful points of comparison because they demonstrate that none of the many rocket producers in the world has managed to make a rocket any more reliable than the Shuttle. The kinds of technology used are inherently unreliable. We need new kinds of technologies to make travel into space a lower risk proposition.

Also, just because the Ariane is not used for human launch does not mean it can't be. A small human carrier shuttle could be designed to fit on the top of it.

50 year old bombers: I'd still be willing to bet that their accident rates are higher than that of a 767 or other more recent designs. However, even 50 year old aircraft are orders of magnitude safer than any space launch vehicle.

Yes, the Shuttle has a more difficult job to do. But do you want space travel to become commonplace? If you do then you should want a launch vehicle that is a few orders of magnitude more reliable than the Shuttle.

My most important argument here is that we need to make much larger strides. We can only do that by designing lots of new experimental vehicles. Jerry Pournelle has been making this argument for at least 2 decades and probably much longer. He's right. We are not learning much from operating the Shuttle. We could learn much more by spending the money on a series of X vehicles instead.

Paul Dietz said at June 12, 2003 4:42 AM:

The reason the shuttle (or, indeed, all launch vehicles) are so comparatively unreliable is that they have not been flown very much. Aircraft became reliable by redesign, yes, but that redesign was in response to experience gained in accidents. We can't fully debug machinery before use any more than we can fully debug programs before use. A great many classes of aircraft design bugs have been found over the last century as we went down the experience curve.

This suggest that to get reliable launchers, we should have a *small* launch vehicle that is launched a great deal. And we should expect to lose many of them as we find the unpredicted design flaws.

Joe said at March 30, 2004 10:14 AM:

Aliens shot down the space crafts, they did not break up.

Kaf said at April 10, 2005 10:44 PM:

I'd like to see a comparison between the Shuttle's 2% success rate and that of R-7 (Russian Soyuz) for the last 25 years, not since the R-7 started flying in the 60s. Or at least with that of manned Soyuz launches. I find it really hard to believe that the Shuttle is safer than modern days Soyus'...



Mike said at July 16, 2005 5:36 AM:


I hope Discovery technicians can smoke out the electrical problem in the malfunctioning fuel gauge scenario and then find the right part on eBay! Sheeeesh!

Randall, you could probably update your article with SpaceShipOne and the X-Prize as emerging models of innovation. What we need is someone like Maj. General Billy Mitchell (http://www.christopherlong.co.uk/per.mitchell.html) to shake up the status quo. Maybe an astronaut riding SpaceShipOne would be spectacular enough to get people to start thinking differently (sorry, I'm a Mac design & software fan, not their marketing fan for mangling grammar - i.e. "different"), and this before we have another Shenandoah-type disaster.

My prayers are with the Discovery crew and their families along with future crews.


Mike Corbin said at July 2, 2006 7:05 PM:

The space program presently depends upon the few space shuttles for all of our manned flights. The next scheduled launch is for the 4th of July, 2006. It is important for the American people to have something to pull themselves together as a symbol of national pride and accomplishment. That is the one singleminded reason for the space program in the first place. The Chinese just returned from orbiting the Earth for 5 days, and that was their 2nd successful flight. Our Hubble space telescope was recently repaired and is now working again just fine. Our interplanetary space probes seem to be working and we have learned more about the structure of our universe with these tools than we ever knew before. How much peaceful cooperation with the Russians did our space program bring? We have just passed a House bill to fund more unmanned missions to Mars. Did you enjoy the pictures from Mars? How about the pictures from Saturn, did you like them? Do you think our recent probe that examined a passing comet was worth the effort? It cost plenty, but we learned we could hit the target if we ever had to. And we learned more about comets than we ever knew. We are scheduled to return to the Moon. We are still heartbroken about the tragic accident of 2003, but wer'e not ready to quit.

Phil Steinmeyer said at July 26, 2009 2:37 PM:

I basically agree with Paul Dietz (posting over 6 years ago). You cannot get expect to get substantially better safety records without traversing the experience curve, and you can't do that without launches - many launches.

The problem is that launches into space are so much expensive and complex than airplane launches.

Wilbur and Orville Wright were able to build and launch their airplane as private citizens - basically hobbyists. Within a decade, MANY others were able to imitate them, and airplane design advanced significantly.

Beyond the hobbyist stage, commercial (and military) applications were reasonably apparent and quickly propelled development forward.

Space flight is extraordinarily expensive. I'm not aware of any easy way to dramatically reduce the cost in the reasonably near future.

Applications for manned space flight are limited. National prestige is nice, but you can only be the first nation to the moon once. Yes, there are more distant objectives (Mars), but these would be extremely costly, and we don't have the cold war as impetus for hugely expensive national prestige programs anymore. (And that's a good thing.)

Aside from prestige, there is a bit of value in tourism, but at current price/cost points, this value is limited. There is a bit of value in the repair business and other odd jobs. But in general, I'm not aware of many high-value applications for manned space flight (as opposed to unmanned) that will drive things forward in the near future.

I don't know exact budgets and plans at the moment, but it seems we've allowed our manned space program to shrink in relative-budget terms, and this is probably a good thing. If/when launches get cheaper/easier/safer, and/or applications for manned space flights become more compelling, we can ramp up the manned program, just as we ramped it up quickly in the first place, decades ago.

Marcel F. Williams said at August 15, 2009 2:39 AM:

It should be noted that there have been over 100 flights of the Space Shuttle since the fatal Challenger accident involving a malfunction of one of the SRBs. And there have been no SRB malfunctions in over 20 years. During the 127 flights of the Space Shuttle there have been only on fatal accident involving damage to the shuttle's thermal tiles. And there has never been any fatal accidents resulting from malfunctions from the SSMEs.

This suggest that a shuttle derived vehicle like the Sidemount should be exceptionally safe since there will be no thermal tiles to damage and since problems with the SRBs were successful resolved more than 20 years ago. Additionally,the manned side-mount should be even safer with a LAS which it would hopefully never have to use.

Post a comment
Name (not anon or anonymous):
Email Address:
Remember info?

Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright