Mike Griffin made a speech this morning at the Space Transportation Association breakfast (something he has been doing annually since he became administrator).
Jeff Foust has some notes from it. I just read it.
The problem, as always, is that NASA never provides any data to support his assertions — we must simply take his word for it. For instance, when he says:
Beyond the costs involved, our probabilistic risk assessment for loss of crew on Ares 1 showed it to be twice as safe – I repeat, twice as safe – as a human-rated EELV-derived vehicle. This figure of merit was a significant factor in our decision to go with the Shuttle-derived Ares 1, yet is ignored by almost everyone suggesting that we make a change. I cannot responsibly ignore it, for reasons having nothing to do with money. But if to someone else it is just about the money, then the cost of unreliability must be considered. Incurring even one additional accident through the use of a less-reliable system wipes out all of the savings of the hypothetically cheaper vehicle. Solely from a fiscal perspective, we should be willing to pay a premium for safety, if necessary.
Who can argue with that? But if it’s true, release the PRA, with its assumptions. Show us how and why it’s “twice as safe.”
There is no discussion, of course, of how this satisfies the Aldridge Commission requirements to be “affordable and sustainable,” and to contribute to national security and promote private enterprise. That’s because it doesn’t.
And this is what I find most annoying about his defense, because it’s a theme that recurs often with him:
But no matter what decisions we make, we at NASA cannot possibly make everyone happy. Most decisions will produce an unhappy outcome for someone. That is not by itself a symptom of incompetence, bad intentions, or a lack of integrity on our part, as some have contended. Allocation of public funds to any particular alternative inevitably leaves aggrieved parties who will not receive those funds.
There is an implicit assumption here that all his critics are craven, and not acting in good faith — that their only reason for criticism is because they have a pecuniary interest in a different solution. I’ve never accused him of incompetence, bad intentions or a lack of integrity — I simply think that he’s mistaken. But he is implicitly accusing me, and every other person who thinks that there are better solutions, of the latter, when he says that we’re just in it for the money. And he does it often enough that one does have to wonder if there is some psychological projection going on.
Newsflash, Mike. I’m not likely to financially benefit from any choice that NASA makes (at least no more likely than I would be with the current architecture). I’ll either get consulting work from a NASA contractor or NASA itself (should I need it) or not, regardless of the vehicle design. I in fact don’t even offer any specific alternative with regard to launch vehicles, because I think that issue is beside the point of the much broader one — how to make it affordable for many people to go beyond earth orbit, and not just a few NASA astronauts. And I argue for that not because I think I’ll get rich if he chooses an alternative, but because I think that the nation will.
I want to see a different approach because I care about our future in space, and I find the current one a waste of taxpayers’ money. That doesn’t mean that I think that Dr. Griffin doesn’t care about our space future, or the taxpayers’ dollars — clearly he is not indifferent to either. But we fundamentally disagree about the best means to achieve the goal. And this attitude of his that anyone who disagrees with him is doing it for the money is just one more reason that I won’t miss him.
[Early evening update]
Clark Lindsey has further thoughts.
As he notes, the crux of the issue is whether or not we need a heavy lifter. Again, people look to the success of Apollo, and assume that it was successful because we did it in a single launch. And it was. But Apollo had different goals than we do (or at least we should) now. Apollo was a race, and it had unlimited funds, and a limited goal — to land a man on the moon and return him safely to the earth before the Soviets could do so. Now, budgets are tight, and the goal is to build a sustainable infrastructure.
Any new rethinking of NASA plans have to be held up to the template laid out by the Aldridge Commission, something that the Sixty-Day Study clearly never did. And we have to do that fundamental trade that Mike Griffin was unwilling to do. Do we want to get economies of scale through activity levels, or through vehicle size? The former is much more likely to give us true economy, and a lot more bang for the buck, than the latter. Apollo on Steroids isn’t even close to the right approach.
[Update a few minutes later]
With regard to the issue of whether or not Ares is safer than Atlas (and ignoring the fact that, as Clark points out obliquely, a paper rocket is always safer than a real one), why the emphasis on ascent safety? As Jon Goff noted a while back (link not handy), most of the risk to crew in a lunar mission happens after they get into orbit, so focusing on launch safety isn’t necessarily a smart use of funds if you’re worried about safety overall.
And you know what else? Despite what Mike said this morning, I’ll bet they didn’t even include costs of unreliability in their overall trade, because the flight rate is so low, and the assumed reliability is so high for such a low rate, that the expected value of mission loss is probably in the noise. He could prove me wrong, though, by just releasing the data…
*Sigh*…I really don’t need another writing project right now, but this really is begging to be rebutted.
~Jon
I don’t even know how to rebut it, because I don’t know what his assumptions are.
Two thoughts…
If one system has a likelihood of failure of .004, then twice as safe is .002. The first system still has a pretty small chance of failure.
More importantly, even if he were to release the data, it is laughable to think one can determine the likelihood of failure without a large volume of real data to work with. Probabilistic assessments are only valid if you can apply the Law of Large Numbers.
Incurring even one additional accident through the use of a less-reliable system wipes out all of the savings of the hypothetically cheaper vehicle. Solely from a fiscal perspective, we should be willing to pay a premium for safety, if necessary.
In that case, how can Mike “responsibly ignore” the potential safety benefits of reusable vehicles?
If one system has a likelihood of failure of .004, then twice as safe is .002. The first system still has a pretty small chance of failure.
Only if your flight rate is small. If military fighters had that sort of failure rate, the Air Force would have no airplanes left. NASA’s still operating on the assumption that human spaceflight will always be very rare, because it always has been. Which becomes a self-fulfilling prophecy.
He’s still using old ESAS numbers for some reason. For those, Atlas V Heavy had LOC of 1 in 900, Delta IV Heavy LOC of 1 in 1100, and Ares I (the five segment version) LOC of 1 in 2021. Also what could an LOC possibly cost? $5 billion? $10 billion? Think about it. You lose maybe six people, delay the program for two years. I think $10 billion is reasonable estimate under those circumstances.
Somehow, I doubt he’s advocating a mere $10 million more per launch premium for that additional 1 in 1000 safety improvement that Ares I supposedly has over the EELVs.
The “PRA” that ESAS used can be found here:
Ummmmm….
Where?
It isn’t loss of crew , its loss of mission. Even in 1970, when we were a lot less risk averse, we held further Apollo missions until we found out what went wrong, even though nobody died.
And do you really think an EELV will become four times more reliable than the the best existing launchers because somebody sprinkled man-rating pixie dust on it?
Or that Ares I will be eight times more reliable?
…do you really think an EELV will become four times more reliable than the the best existing launchers because somebody sprinkled man-rating pixie dust on it?
Or that Ares I will be eight times more reliable?
Of whom are you asking that question? If it’s me, the answer is certainly not, and I have no idea why you ask it. What is your point, and how does it relate in any way to this post, or comments so far?
Rand,
Your comment on the risk of launch safety vs. safety during operations in the lunar vicinity was one of the key points I was going to make in my rebuttal. Using the old BS ESAS numbers for LOC probabilities, a 1:1000 number for EELV systems means that if you fly 500 missions, you have a 61% chance of having no LOC events. That’s over 80-120 years at the planned ESAS rate. At 4 flights per year you would theoretically have an 89% chance of not having any failures in 30 years.
OTOH, ESAS was claiming an overall LOC probability of about 2% per missions (about 1:60 IIRC). At two manned lunar missions per year, that means that you have a higher chance of losing a crew in 3.5 years of lunar operations than you would in 30 years of flying 4 manned EELV missions per year.
This argument doesn’t hunt. Especially since in order to fly on der Griffenshaft, they’ve stripped out a lot of the redundancy, MMOD protection and radiation shielding that directly impacts the LOC numbers during the riskiest parts of a lunar mission. I wouldn’t be surprised if Ares-I isn’t making the net lunar mission more likely to kill crews than if they went with an EELV architecture. Once again, his argument is BS, plain and simple.
Also noticed that nowhere in there was there any discussion of depots, in spite of the fact that Griffin is fully aware of them (and has even supported them in recent history).
~Jon
Rand:
I was responding to Karl, who seems to be arguing that the relevant metric is loss of crew, and then calculating the appropriate reliability premium for Ares I if the ESAS LOC numbers are correct.
My argument is first, that any loss of mission will be expensive. The mission will be scrubbed, so you will lose the replacement cost of a Lunar mission. Second, there will be a standown. Even if only for three months, you will incur the fixed costs of the entire program during this period. So the cost of each loss of mission could easily exceed a billion dollars, and might well be twice that.
Second, that the ESAS numbers are optimistic. The most reliable expendable launcher will have a LOM rate closer to 1% than .25%. The second best launcher will be closer to 2% than .5%. So the penalty for a less reliable launcher will be greater.
“My argument is first, that any loss of mission will be expensive. The mission will be scrubbed, so you will lose the replacement cost of a Lunar mission.”
Um, even if the hardware is in orbit, why do you have to replace it if you have a couple of month standdown? It’s not like the vehicles are going to disappear or anything. At worst, you’re talking about the replacement cost of some of the propellants.
~Jon
My argument is first, that any loss of mission will be expensive. The mission will be scrubbed, so you will lose the replacement cost of a Lunar mission.
I don’t know what “scrubbed” means in this context. In a sane world, the mission would be delayed, with a new crew. The only cost would be the time to get the new crew and capsule up, and the replacement of the hardware (only the capsule and service module, or whatever was to go up on the Ares 1, and the launcher itself).
Jon:
When the reliability and cost of cryogenic propellant transfer in orbit are demonstrated facts, then you can say that all you have to do is replace some propellants.
We don’t have that capability, and don’t know how cost effective it will be if we get it.
And in any case, the cost of a standown will remain.
I was going to say, “Prepare for escalation of commitment arguments to keep constellation”, but then I read Griffins speech and hes already making them.
“We must not allow indecision and uncertainty to cause, again, the waste of billions of dollars already invested”
Escalation of commitment is a fallacy and constellation should be objectively assessed based on current technical merit. It is not enough to continue a program because so much as already been put into it.
Forget about “human rating the avionics”, I want to know the nuts and bolts of thrust and payload and could-they-do-the-job.
Can anyone tell me what “is” the EELV architecture? Which EELV and what is the payload? If the EELV is one of the “heavy” variants with parallel core stages, do those parallel core stages don’t appear to have the same kind of “engine out” redundancy as the old Saturns — is this a problem?
I also thought that EELV’s had low-thrust upper stages meant for taking lighter payloads into geosynchronous orbit and that for heavier payloads into low Earth orbit, they required “lofted trajectories” that had “abort deadzones.” Is this FUD or is there some proper background on this?
If the EELV is one of the “heavy” variants with parallel core stages, do those parallel core stages don’t appear to have the same kind of “engine out” redundancy as the old Saturns — is this a problem?
Not for NASA, obviously, since Ares 1 has no engine-out capability.
I also thought that EELV’s had low-thrust upper stages meant for taking lighter payloads into geosynchronous orbit and that for heavier payloads into low Earth orbit, they required “lofted trajectories” that had “abort deadzones.” Is this FUD or is there some proper background on this?
That didn’t stop Lockheed from proposing Ariane V as a launcher for space tourists.
If you have to violently “abort,” there’s a good chance you’re going to die no matter what “zone” you’re in. There are no safe zones where ejection is a 100% slam dunk, no matter what Mike Griffin thinks.
If you really care about safety, what you should be looking for is *intact* abort, where the vehicle returns in one piece. No expendable can do that.
When you decide to go with an expendable rather than a reusable, you’re already throwing away most possible safety features and committing to a circus stunt that’s more dangerous than professional bullriding. At that point, it makes little sense to spend an additional $5+ billion to develop a slightly safer bull.
(To put costs into perspective, NASA’s 2nd Generation RLV, while oversized and overengineered, had an estimated development of only $5-6 billion.)
Paul,
I think that one of the keys is not tying the architecture to any one vehicle. There are several solutions, and I think one of the keys needs to be making things more launcher independent so an issue with one launcher doesn’t stand down the whole program. As it is, there are several existing and near-term commercial launch options whose reliability is good enough that the odds of losing a crew over the course of 20-30 years is pretty darned low. Making Orion launcher independent and then focusing on the actual lunar development would make more sense.
~Jon
Will, I was discussing Griffin’s claim, clearly I might add. He said:
Beyond the costs involved, our probabilistic risk assessment for loss of crew on Ares 1 showed it to be twice as safe – I repeat, twice as safe – as a human-rated EELV-derived vehicle. This figure of merit was a significant factor in our decision to go with the Shuttle-derived Ares 1, yet is ignored by almost everyone suggesting that we make a change.
When one looks at the numbers that “probabilistic risk asssessment” claims the Ares I is safer (in terms of LOC) than EELVs by 1 in 2000 chance roughly (I was wrong in my prior post). It doesnt’ take much cost savings per launch to justify not having that extra bit of safety.
With both EELV launchers active, there need not be a standdown. Delta IV and Atlas V have little in common (payload integration and launch control probably is it).
Karl:
Delta IV and Atls V both use the RL-10 engine.