Space Tourism Still Not Ready for Flight

Ten years ago when Burt Rutan was predicting 100,000 space tourists in ten years I wrote a widely debated article, Is Space Tourism Ready for Takeoff? My answer then, and my answer now, is no:

The vision is enticing but the facts suggest that space tourism is not ready for market. The problem is not the monetary expense, there are enough millionaires with a yearning for adventure to support an industry. The problem is safety. Simply put, rockets remain among the least safe means of transportation ever invented. Since 1980 the United States has launched some 440 orbital launch rockets (not including the Space Shuttle). Nearly five percent of those rockets have experienced total failure, either blowing up or wandering so far from course as to be useless. The space shuttle has a slightly better record of safety — it was destroyed in two of 113 flights. There are lots of millionaires willing to spend one or two million dollars for a flight into space but how many will risk a two to five percent chance of death?

It is true that we have been “learning by doing” or in this case by learning by exploding. In the 1960s the risk of failure was a stunning 12%. As in other industries, learning by doing reduced the failure rate dramatically over the first units but more slowly thereafter. In the 1970s the failure rate dropped to 5.2% but nearly thirty years later the failure rate for rockets still hovers between four and five percent. We can expect similar slow and steady improvements in the future but there is little reason to expect dramatic improvements in rocket technology

Unfortunately with two disasters this week, one of them sadly involving the loss of life, the safety of rockets continues to be far too low to support significant tourism. Virgin Galactic’s VSS Enterprise, which crashed yesterday, was just on its 23rd powered flight suggesting a failure rate of perhaps 5%, in line with expected values. An earlier tragedy involving tests of the rocket motor killed 3 people.

As I said ten years ago, even a failure rate of 1 in 10,000 is far too high to support space tourism of the “fat guys with camera” variety and we are not yet close to a failure rate of 1 in 10,000.


Rockets are unsafe or NASA is unsafe? How many people have the Russians lost in the last 30 years? I mean, I wouldn't buy a can opener if it had a "Made in Russia" label, but somehow they seem to have a better record than the US in space safety.

Here's a video of a Russian rocket launch in 2013:

Stick around until 0:31.

In Soviet Russia rocket explodes you!

Neither of the recent failures were NASA rockets. Antares is built by Orbital Sciences Corporation, and Space Ship Two was built by Virgin Galactic. The Russian Soyuz rocket has had a failure rate of about 3%, and the Proton-M of about 11%. So Alex's point holds whether you are in the U.S. or Russia.

Scaled Composites built SpaceShipTwo not VG.

The Antares Rocket that blew up (actually blown up by the launch team) had a Russian engine as its main propulsion.

Soyuz has been pretty safe. They've only lost the crew in 2 of 66 launches - and all of those deaths were in the first generation of Soyuz spacecraft back in the late 1960s/early 1970s. Since then, Soyuz missions haven't lost a single astronaut.

According to Wikipedia, the Soyuz-FG -- which Russia uses for manned spaceflight -- has a 46/46 success record. That is entirely consistent with a failure rate in the low single digits.

If this article is to be believed, the rocket that crashed was particularly unsafe.

“They operated in secrecy, which is difficult to understand,” said Mr Sgobba. “They don’t use modern techniques in putting safety into the design."

"They use outdated methods like testing and then seeing what happens. There has been no independent oversight."

"There is no peer review. I have been saying for some years now this was an accident waiting to happen.”

Both of these appear to be hatchet jobs built around cherrypicked or solicited quotes; I do safety and mission assurance for rocket propulsion systems for a living, and I don't recognize most of the people or even organizations they are citing.

What they fail to mention, or mention only in passing, is that Virgin et al responded to the various warnings about their chosen propulsion system by scrapping that propulsion system and standing down from flight testing for a year while ground-testing an alternate propulsion system they had been pursuing from a different manufacturer. This was the first flight test of the new system, and obviously there was a problem that didn't show up on the ground but did manifest in flight. That's what a comprehensive test program, ground and flight, is for. And if the new system turns out to be as unworkable as the last, Virgin has yet another in the pipeline that might be adapted.

As for the departure of some high-ranking people: that might indicate a problem, but I think you'd want to hear that from them before drawing any conclusions. Few technical people want to devote their entire career to a single entrepreneurial company, so turnover is to be expected. If, ten years into the program, you have a twelve-month down time when you're developing a new propulsion system and not flying anything, that's not an unreasonable time to bring a new propulsion chief on board and for the aerodynamics and safety heads to move on to new pastures.

I am, personally and professionally, one of the people who has said from the start that Virgin made a poor choice in their original propulsion system (and elsewhere), but I can't agree with the stand these articles are taking.

Thanks for your perspective. Your credentials are indeed impressive.

By the way, I'd change "Techical" to "Technical". If it's an important document, I always spell-check. (I do now, anyway, after an embarrassing incident I'd like to forget.)

"By the way, I’d change “Techical” to “Technical”."

Nah, he's just burnishing his engineering street creds.

According to Tomasso Sgobba, actual science (testing stuff to see what happens) is "outdated".

"Rockets are unsafe or NASA is unsafe?"

The orbital science and Virgin Galactic failures were private sector, and only one of the two under contract to NASA, but not under the control of NASA for all launch decisions.

Scaled Composites were responsible for several more flight tests BEFORE the handover to Virgin Galatic.

I like the failure rate as a rule of thumb for whether an industry (essentially an adventuring industry) is viable - but I am not convinced. Tourism can be anything. There are tours of long dead east-european factories, Chernobyl, swamps, etc. - My point is that create the right price point or the right 'high-risk job/ pay ratio' (say asteroid miner/ satellite repair-person) and you have people that will make themselves available even if just means an orbiting brothel for after-work 'entertainment' when you have just given the Hubble a tune-up. With orbiting industry, orbital clean-up, inflatable habitats, even one-way trips to wherever, you will find people to work/ play/ visit. The idea is that it will be several industries co-evolving with the side-dish of entertainment and adventure on the side, further broadened by having a wide (even controversial) range of options. Billionaire adventurers is not the future of tourism nor even its vanguard point - it is the frontier town, the oil rig, the arctic installation, king-crab fishers-type work-environment - taken up into space that will facilitate the grand exodus of entertainment-seekers for the middle part of this century.

But you also have to factor in the economic costs. Even if the tourists were willing to accept a 4-5% failure rate, and agree not to sue for damages, you still have a rather large and on-going capital loss. if you plan to use a given launch system 20 times a year, then you have to recover it's entire cost in less than a year.

It's a real example of hubris that NASA at one point thought it was a good idea to send civilians (basically, tourists) into space.

NASA sent one "civilian", a teacher, into space while Reagan was president. That was to inspire students to education in general and STEM in particular.

All other "civilians" were "sent" by Russia to help fund its space industry. Those were real tourists, paying I believe a hundred million for their rides on Russian rockets.

Technically, all persons in space are civilians because space is by treaty non-military and NASA is a civilian agency.

Google has cars that drive themselves but they're still putting human pilots in rocketcraft? Having someone at the controls is simply a guarantee that at least one person will die if there is a serious accident. If I had somehow managed to win a ticket for a trip on the space shuttle back when they were still flying I would have been very tempted to go, but not if I had a spouse or children. Not unless I really hated them and wanted to subject them to a one in fifty chance of daddy go boom.

There's still some stuff in orbit that only humans can do, like repairing the Hubble Telescope. Now, once we have dexterous robots in space that are remotely operated by people on the ground . . .

I'm not sure it's true that there are only things humans can do in orbit. One figure I have is the marginal cost of a space shuttle launch was $450 million. The robonaut that is being trialed on the ISS apparently cost about $25 million. I'm thinking the cost of one space shuttle mission could buy a whole lot of robot.

And then there's the question of whether it's worth sending humans up to service things in orbit or if it's cheaper just to shrug and send up something new. The Hubble telescope apparently cost $2.5 billion. Five missions were sent up to service the Hubble and they apparently didn't do much else, so maybe that cost about $2.25 billion. It may have been more cost effective to simply send up a new, more advanced telescope, on a much cheaper uncrewed rocket launch.

To be blunt, a human life isn't worth very much. It might cost NASA $20 million to train an astronaut and there doesn't appear to be a shortage of volunteers. From an accounting perspective they should spend much less on keeping astronauts alive and just accept they are going to lose people. But that sort of thinking doesn't appear to be appreciated. Human space flight is caught in a bind where it looks bad if people die and so spends a huge amount of money trying to keep people alive, which makes it look bad compared to the cheaper and safer option of just not using humans. As technology continues to improve I'm sure crewed space travel will continue to lose out and before long people will only go into space for the purpose of going into space and won't pretend they have anything worthwile to contribute.

The Space Shuttle was a particularly costly means of getting things and people to orbit. Which is really saying something, given that it is inherently an expensive (energetically) endevour to begin with. I wouldn't use Shuttle costs as a baseline for anything except how political bargaining can increase engineering costs (part of the problem was the way things had to be spread around to many districts to spread the pork around, and a large part was the Air Force wanting a do-everything platform).

Russia seems to be the queen of low cost reliable space launches. (Reliable given the fact that a rocket is a flying stack of explosion juice.) China and India may be strong competitors once they improve their track record. India has a lauch site in an excellent location for low cost launches just eight degrees from the equator. But it is on the west side of the peninsular rather than the east side which means disasters can fall on Indian land rather than out to sea. But perhaps they thought that was better than hitting Sri Lanka. Having a launch site near the equator only gives a payload advantage of perhaps less than 0.25% over Russia with regards to simply getting into orbit, but it gives a huge advantage for getting into an equatorial orbit which is where many satellites want to go. The International Space Station is located in an inconvenient orbit for the United States in order to reduce the cost of getting to it from Russian launch sites. Which has turned out to be fair enough since they're doing most of the launches.

Google cars will simply stop dead if a problem is encountered, which is anything that is outside the programmed operating parameters.

In the case of satellites, they just orbit and humans have had months and years to figure out the problem and solution. But the probes trying to land on Mars crashed before humans knew anything was wrong.

No, Alex is wrong. Of the 135 Space Shuttle launches, only two were lost, that's 1.5%.

You may notice that his statistic was within the quotation from 2004 when the shuttle program was still active.

1.5% is much better than 5%. For these things, you aim at 2 or 3 orders of magnitude less.

Given NASA has done zero tourism, the only space tourism is by Russia. Russia has a 100% success rate for its space tourism trade.

The Russian problem is cost; the $100 million they charged exceeded marginal cost on launches with most fixed costs paid for by NASA.

Russia charged 20 million to Dennis Tito for the first commercial flight. The last flight they charged almost 40 million. Russia charged the U.S., in the the last NASA contract for spaceflight services, was 71 million per seat. Since they are a monopoly prices would naturally go up, as they always tend to do under monopoly conditions.

SpaceX is still scheduled to charge 20 million per seat for the Falcon 9v1.1 and Dragon V2.0 flight system.

"No, Alex is wrong. Of the 135 Space Shuttle launches, only two were lost, that’s 1.5%"

He specifically said, "but how many will risk a two to five percent chance of death?"

If you compare fatality rates with high altitude mountaineering expeditions (pretty similar), something thousands of people a year pay USD 20k-60k for, it wouldn't seem this is a fatal flaw in itself. Risk perception and perceived control, coupled with liability issues will be radically different though.

Are you suggesting those K2/Everest climbs have a 3-5% fatality rate?

Because, um, I don't think that's correct.

... and in fact, they do not:

I presume Bonsai was comparing it to the Space Shuttle which killed people around 1.5% of the times it was launched or about 2% of the time when Alex wrote his old piece. So it is not that far from the fatality rate for climbing Everest. But average human death rate per shuttle flight was a little above 1.5% on account of it being filled to capacity both times it disastrously failed.

I think Bonsai makes a good point about perceived risk. I'm sure climbers feel much more in control of whether they live or die than tourists being launched into space would. If a family member died climbing everest I would probably think they shouldn't have tried to climb it. If they died in a commercial airliner accident I probably wouldn't think they shouldn't have tried air travel.

Would anyone here know: are the $250K Virgin Galactic ticket prices (as advertised on NPR), which scores or hundreds have already paid, non-refundable?

It's been announced refunds are available.

Ticket #2 holder (Branson has #1) has refused to sell his ticket even at higher prices, and said he still would not sell his spot after the crash.

"learning by exploding"


Come on people, this isn't brain surgery!

There are some useful distinctions to made here.

First, there is Loss of Mission (LOM) vs Loss of Crew (LOC). Every rocket that fails is an LOM, but would not neccessarily be an LOC in every case. For example, there are Launch Escape Systems (LES) that dramitically reduce the chance that the most likely types of LOM will result in LOC.

Second, there are 2 ways crews go into space, via capsules or via spaceplanes. Apollo and Soyuz were/are capsules, the Shuttle was a spaceplane. Safety options for spaceplanes are severely restricted due to the nature of their design. Safety options for capsules are greater. For an example of a rocket LOM that did not result in LOC due to the capsule's LES see

We all know about the 2 Space shuttle failures that resulted in LOC, but there hasn't been a capsule LOC since 1970.

Capsule reliabilty may already adequate for commercial travel.

The Virgin Galactic craft that was lost was a sub-orbital rocket-powered glider. Capsules may be reliable (at the ~1% level) but you can't control where they land so you need a search-and-rescue team to go in after the capsule lands. I don't think that is a scalable model if you want to have several launches per day.

None of the crewed Apollo capsules landed more than 5 miles from its it's aim point, most missed by less than 2 miles.

Dragon capsule returns have been even more accurate.

Space launch emergency return and abort scenarios require large scale deployment of rescue/recovery assets, but they would be only a little less for spaceplanes.

Capusle landing accuracy is not a safety or economic issue.

Soyuz capsules have landed hundreds of miles off-course, resulting in things like wolves arriving at the landing site before the rescue team. This isn't a reflection of Russian inadequacy, but of probability. One advantage of a capsule design is that you don't necessarily need the guidance system to work perfectly (at least for reentry from low orbit) as a purely ballistic reentry is a rough but survivable ride. Fly enough capsules, and a few - maybe one in fifty - will wind up far off course.

Winged systems, the guidance system has to work, but the configuration Rutan worked out for SS1 and SS2 is stable and manually controllable throughout the flight envelope, so the guidance system can come down to one guy with stick and rudder. That has traditionally worked much better than 98% of the time, so it may be a net win here.

The Shuttle losses were purely the result of Congress being penny wise-pound foolish - Nixon and Congress would not kill manned space travel in 1970, but would not fund manned space travel. And every year after, Congress would not fund manned space, but would not kill manned space.

Bush would not fund manned space travel but would not kill it.

Obama knew that Congress would never fund manned space travel, so he cancelled manned space travel unless NASA could buy it COTS.

Obama did not do what Nixon, Carter, Reagan, Bush, Clinton, Bush did and keep kicking the can down the road because they didn't want to be the one killing manned space travel.

Obama killed NASA's manned space program. Pissed off a bunch of anti-Federal government, anti tax, anti big government Republicans:

Gov Perry circa 2011: "Unfortunately, with the final landing of the Shuttle Atlantis and no indication of plans for future missions, this administration has set a significantly different milestone by shutting down our nation's legacy of leadership in human spaceflight and exploration, leaving American astronauts with no alternative but to hitchhike into space."

Why didn't Perry praise Obama for ending wasteful pork barrel spending?

For turning wasteful government programs to the private sector?

Because the spending can't be funnelled into Texas.

That's some grade A revisionist history.

"The Space Shuttle program was extended several times beyond its originally envisioned 15-year life span because of the delays in building the United States space station in low Earth orbit — a project which eventually evolved into the International Space Station. It was formally scheduled for mandatory retirement in 2010 in accord with the directives President George W. Bush issued on January 14, 2004 in his Vision for Space Exploration."

Sure, but there was supposed to a replacement in place by 2008 (Orinon), and it was convenient that the retirement would happen on someone else's watch (easy enough to reverse with a new appropriations bill). It was a retirement of the shuttle, but a replacement with a new craft that would then build a lunar base.
Of course, that didn't get funded either. But both presidents get some credit; Bush for calling it a day on the shuttle and Obama for going through with it even though the replacements are still in the works.

President Obama did not end the Constellation program.. A little CIVICS 101. The President presents to congress a NON BINDING budget PROPOSAL. The congress has traditionally ignored THIS President's budgets. A BI PARTISAN congress REFUSED to fund the CONstellation program. Both houses of congress agreed and voted to NOT FUND it.

President Obama called for a 6 billion dollar increase in the NASA budget over 5 years to fully fund commercial crew to close "the gap" the time the U.S. would have to buy seats from the russians.

The Republicans in the house REFUSED to fund commercial crew at that level, they appropriated 270 million for one year and on a year by year basis.

Somebody needs to ask Sarah Brightman about her upcoming "concert from space".

Doesn't the very large number of people who were/are astronauts, and the even larger number of people who want to be astronauts, discredit the notion that there is too small of a number of people willing to risk their respective lives going into space?

Or how about the many rich nuts who try and sometimes succeed to ascend Mount Everest, K2, and other very high mountains which require oodles of money to summit?

Or the occasional loon who delights in trying to fly hot air balloons around the world?

This blog consistently lauds technological progress in the abstract while pooh-poohing individual instances of new technologies and first popular applications of technologies.

Everest climbing has a death rate in the single digit percentage. We lost the Intrade founder to Everest a few years back.

Climbing Mount Everest is very dangerous, maybe a 3% risk of death, so the rockets might be safe enough for a niche market of rich adventurers.

At a conservative estimate, being obese has a 30% chance of significantly shortening your lifespan. So the “fat guys with camera” are already living pretty dangerously.

2%? That's a feature not a bug.

On the positive side, a total rocket failure and the craft was still flyable. It looks like the craft reached the ground in one piece. I'm going to guess the one pilot that died survived the explosion but not the very rough landing in the middle of a rocky desert.

When I climbed Mt McKinley in 1984, the death rate was one out of every 300 attempts. Of course the odds were better on the standard route. (We had good weather and succeeded.) When I was kayaking in the late 70's and occasionally doing first descents, we would stand on the bank scouting a hard rapid and trying to calculate the difficulty and wondering about the odds of death. We figured 1 in 1000 was ok if we didn't run too many such rapids. And, at that time, no serious kayaker had died in a California river. Soon after they began to die. I never thought of myself as doing anything to dangerous, nothing like what is done routinely today in many sports. And we used to say that the dangerous part was driving to the river or mountains.

I kayaked the American river in California and some people died when it was in flood stage, much more experienced kayakers than me (who also did it during a flood). So, like you say, life is full of risks. Even here in the PH you can have a mosquito bite you and you may die of malaria or dengue fever. Actually the biggest risk may be staying indoors and doing nothing...from radon.

Your house is made of brick and you tightly seal it against the subzero manila winter? Otherwise not so much.

At Waterloo: "Captain Mercer of G Troop Royal Horse Artillery described the use of Congreve rockets on 17 June 1815 during the retreat from Quatre Bras as follows:

"The rocketeers had placed a little iron triangle in the road with a rocket lying on it. The order to fire is given - port-fire applied - the fidgety missile begins to sputter out sparks and wriggle its tail for a second or so, and then darts forth straight up the chaussée. A gun stands right in its way, between the wheels of which the shell in the head of the rocket bursts, the gunners fall right and left… our rocketeers kept shooting off rockets, none of which ever followed the course of the first; most of them, on arriving about the middle of the ascent, took a vertical direction, whilst some actually turned back upon ourselves - and one of these, following me like a squib until its shell exploded, actually put me in more danger than all the fire of the enemy throughout the day."

That's why you can't really rely on the wealthy. They're generally motivated by social status and are risk averse. They already have their high social status, and social status is a zero sum game. They can only go down. Hence they mostly focus on preserving their wealth i.e. relative social status and thus most of their "investment" involves buying mansions, property, art, mega yachts, etc.

If you try to focus on making it perfectly safe before doing anything, you'll never get anything done. The focus should be on making it as cheap as possible. There are a lot more desperate people willing to take the risk if it's cheap enough, than there are wealthy people willing to take the risk.

Test failure rates are not the same as operational failure rates. There should always be more failures during testing than operation. And not all test flights are comparable. It is notable that the SpaceShipTwo just switched to a new rocket engine this year.

"was just on its 23rd powered flight suggesting a failure rate of perhaps 5%"

SpaceShipTwo had performed only three previous rocket-powered flights. All of its other test flights were captive carry (carrier plane never released the spaceship) or unpowered glide (carrier plane released the spaceship which then glided unpowered to the ground).

The accident was also the first test of a new rocket motor design. It's difficult to extrapolate a failure rate from early testing when the design is still undergoing significant changes.

I just looked it up and only two out of 79 uncrewed resupply and module launch missions to the Internetional Space Station have failed for a failure rate of only about 2.5% which is excellent. Of those rockets 61 came from Russia.

Alex, people who hype driverless cars should not be bragging about how they once shot down the idea that we'd have 100,000 (!!) space tourists in 2014. Because danger.

>even a failure rate of 1 in 10,000 is far too high

Do tell.

Mature space transportation systems, in the US and Russia and elsewhere, military and civil and commercial, all seem to have failure rates of about 2%/flight. This will probably come down quite a bit as we transition to reusable systems; the Shuttle never really got a chance to explore that as there were too few flights and too much rework/repair needed between each flight (IIRC, it wasn't until late in the program that they ever managed to fly the same orbiter twice in a row without replacing at least one complete engine). I would expect Virgin, XCOR, Blue Origin, and the rest to get down at least below a 0.2%/flight failure rate before taking up paying customers.

As others have noted, there's precedent for high-end adventure tourists tolerating even a 2% death rate, so these outfits aren't doomed from the outset. But there's a limited supply of adrenaline junkies who can write six-figure checks for fun, so the risk will have to come down for the market to grow. And the risk will come down with more operational experience and improved designs, but it isn't clear which curve is steeper.

The other dimension that hasn't been addressed is survivability. Virgin was apparently flying without pressure suits or ejection seats, which makes it almost miraculous that Pete Siebold survived. XCOR I know does have pressure suits and an assisted-escape system, and Blue Origin is being inscrutable as usual.

"would expect Virgin, XCOR, Blue Origin, and the rest to get down at least below a 0.2%/flight failure rate before taking up paying customers."

How could you afford to prove that you got it down to a 0.2%/flight per failure? Or even a less than 1%/flight failure?

This isn't meant to be snarkly. I'm more interested in how many successful flights you'd need for a probabilistic determination of failure risk?

Hard to say. 500 flights with no failures would at least strongly suggest a 0.2%/flight failure rate or less, but it wouldn't be a hard proof and of course you'd like to go into service without paying for 500 flight tests.

The magic word here is FMECA, Failure Modes, Effects, and Criticality Analysis. Roughly speaking, you put together a list of everything that can go wrong, and how it can go wrong. Next you determine the consequences of each failure mode, including multiple simultaneous failures. If the feather mode is manually unlocked before Mach 1, nothing bad happens, and if there's a short in the wire that commands feather mode deployment nothing bad happens, but if both of those happen the spaceship tumbles and disintegrates, that sort of thing. You won't get them all, but you should get the ones that, collectively, make up most of the risk.

Then measure or estimate the probability of each of these failure modes. A lot of this can be done by ground testing. And much can be done by similarity to other systems - the propellant valves on SS2 may not be that different than those on a dozen other spacecraft, and if their collective flight heritage suggests a 0.01%/valve/flight failure rate, you go pencil that in. And there are various empirical formulas based on collective industry experience.

Flight testing is highly useful in this process, but not just simplistically flying around a lot and seeing how often things break. Partly, you're looking for unusual behavior that suggests new failure modes you didn't recognize. Partly, you're pushing things closer to the limit than you would in an operational flight and seeing if they get closer to breaking. Probably the biggest part is taking measurements that go into those industry-wide empirical formulas - just how much vibration does the vehicle experience, and so forth. Might only take a few dozen test flights to get a pretty good estimate.

But, for a new vehicle especially, "pretty good estimate" is not the same as proof. There were FMECAs done for the Space Shuttle early in the program that mathematically determined the fatal crash rate would be somewhere in the range of one per hundred thousand flights. On close and unfortunately retrospective examination, there were a lot of wishful-thinking make-believe numbers penciled into those, and what should have been majpr branches of the FMECA were left out because nobody wanted to deal with them and everybody imagined they were populated only by impossible failures. Basically, lying. Certainly people will be suspecting Virgin Galactic, and to a lesser extent the others, of doing that sort of thing. The FAA will be looking fairly closely at the parts that are covered by their regulatory mandate, i.e. the ones that endanger third parties, but as they are civil servants I'm certain they will be accused of being lazy, incompetent, and/or corrupt.

And these new vehicles will all be doing new things that aren't accounted for in the standard methodologies. Spaceships, for example, almost never reuse the same hardware. Airplanes do, but don't usually push it as close to the ultimate performance limits. So they will have to patch together new rules for things like SS2, and we won't know for sure that SS2 or Lynx or Blue Shephard are below a 0.2%/flight failure rate until they have racked up 500+ flights.

Thanks for your post.

This is an excellent post and highlights the importance of full-up system-level testing. FAMECA and the like are critical steps to building reliable systems, and the only way you get to a reliable system in the end is to go through a process like the one described above.

At the same time, it's not sufficient for the reasons outlined in the last two paragraphs.

Ok seeing a lot of commentary about space travel but I was hoping to see a discussion of what causes rocket failure rates of over 1%?

Keep this in mind, a quick google search seems to indicate something like 29,000 commercial flights in the US every single day. A 1 in 10,000 catastrophic failure rate would amount to about 3 crashes every single day. In reality even 3 a year is pushing it at this point.

Can we get rocket failure rates that low or is there something fundamentally different about rocket engineering that will forever doom it to a high failure rate? If so what options are there for serious commercial traffic in space? Unmanned crafts avoid the loss of human life but still a 1%+ failure rate means a lot of expensive cargo is going to get destroyed too.

How does a 1-5% failure rate for space travel today compare with the failure rate of lond distance ocean travel in centuries past?

Airliners have the advantage of having flown before. Even if you are the first paying customer on a brand-new 787, that airplane will have undergone several test flights by both the original manufacturer and new owner. These will be conducted close to an airport with the facilities for a major overhaul, will gradually expand the tested flight envelope of that particular aircraft, and if they turn up any glitches the pilot will land and have them fixed. And if any serious problems show up after the plane is in operational service, you still usually land and have them fixed before continuing. Well, unless the pilot is a complete idiot that's what you do.

The way things work with space launch vehicles, you get the wings and fuselage of your new 787 delivered separately to the airport by truck, and bolt them together on-site. Then you load paying customers and/or cargo onto an airplane that has never flown before, do some basic ground systems checks, maybe run up the engines for a minute or two, and take off on a transatlantic flight. If anything goes wrong, you either keep trying to cross the Atlantic or ditch wherever you happen to be.

The space shuttle, as mentioned above, was only a partial exception to this. Most space launch vehicles, there simply is no provision to land them intact, so you bet everything on that one first flight. This fundamentally can't be as reliable as an incremental test program followed by ongoing use of proven hardware with a regular maintenance schedule and intact abort capability.

Good reply but are rockets really as reusable as jet engines? Yes I know you can have a refuelable rocket but it seems to me every major launch (by that I mean getting a payload from earth into orbit, not simply accelerating while in orbit) represents a huge amount of stress that simply doesn't compare to what a jet engine goes through on a regular flight.

If that is true then it seems the optimal model is the disposable rocket that you bolt onto your space draft and toss off when it is done. Reusable crafts are fine once they are in orbit and their safety record seems very good (the shuttle never had a problem in orbit, it was the rockets getting it there that were risky...getting down was a problem too but that was again because of the damage done by the rockets that took the shuttle up). Once you are in space it seems gentle acceleration is quite doable and probably very safe. However to get from earth's surface to orbit requires an huge explosive release of energy barring massive engineering projects like a space teather or elevator. Huge explosive release of energy does not sound like something that works well for a reusable craft and I wonder if it would ever be possible to get to risk factors of 1 in 100,000 or more rather than 1 in 100.

Reusable rocket engines aren't actually that big a deal. On another subthread I described the process by which reliability is established for a spacecraft, and among other things noted that it requires a dozen or more flight tests. The same is true of the rockets in isolation - you need a dozen or more ground tests before you can trust the thing to fly once. And nobody wants to throw away dozens of very expensive rocket engines before they can even start the flight test program that will eventually lead to their first check for a paying customer.

So even engines for expendable launch vehicles are designed to function reliably for at least a few dozen ground test runs, and they are usually in good shape at the end of qualification testing. It may be that space-launch rocket engines can't or shouldn't be pushed beyond a hundred flights whereas aircraft jet engine can last for a thousand or so, but that's not going to hugely change the economics and it isn't an argument for expendable systems.

Also, regarding the "huge amount of stress": Note that the energy required to lift a pound of payload to Low Earth Orbit, is about the same as the energy required to fly that same pound of payload from New York City to Sydney, Australia on a jet plane. There's a rate difference, in that the rocket trip takes less than an hour whereas the jet takes more than ten, but there's no reason for the accumulated wear and tear - or the cost in general - to be more than an order of magnitude greater than that of an NYC-Sydney airline ticket. Except, of course, that we've been doing the airline thing longer, particularly if we are focusing on the cost-sensitive commercial users.

Perhaps space rockets are inherently dangerous due to the high velocity and all that explosive fuel. But if so, it would seem that jetliners should be as well. Although a couple of early jetliners (DeHavilland Comets) did break up in flight.

But it's still difficult to imagine there would be many takers for a $250,000. thrill ride that lasts all of 15 minutes. What, exactly, is the experience: 3Gs, and a view out the porthole?

Perhaps there's a market for week-long orbital vacations, but, how many will pay for a 15-minute bang-your-done joyride?

I think they'd do better building a simulator and selling rides in that. That way they could package it with ground school, and let the customer "fly" it. If they could offer a unique experience, perhaps the profits from that could finance an orbital vehicle, and then a vacation satellite?

I could cetainly have a lot more fun for a lot longer time with quarter of a million dollars. A full orbital trip seems much more worth it than a suborbital trip where 6 minutes of weightlessness is likely to simply cause nausea. There is a reason why reduced gravity aircraft used to train astronauts get nicknamed vomit comets. Vomit comet flights show that apparently one third of pasengers become violently ill, another third are ill, and the remaining third are okay. Note that okay does not mean they feel good. People generally recover from this nausea in time, but six minutes is not enough.

And extended stays in orbit run into another problem and that is radiation exposure. While low earth orbit isn't so bad any radiation exposure at all is not good for good for business and it is a serious issue for the paid crew. NASA aims to keep astronauts' chance of getting cancer from their jobs to under 3% which is much higher than what is acceptable in the nuclear power industry. If they are kept to the same standards as nuclear power workers pilots and space hotel staff may not have very long careers.

Comments for this post are closed