“Rocketman” has thoughts related to yesterday’s post, here, and more details on last week’s Ames meeting:

First up for review was the preferred solution of a single plane attenuation system. It was believed to be the lightest weight band-aid available for the already overweight and underperforming stick. Architectural changes made to the segmented spaghetti-like stack have made it stiffer. That had the unfortunate side-effect of sending increased loads up to the crew compartment. Those loads are even higher than the seat-of-the-pants requirements loosened by fiat (and loosey-goosey analysis) from the accepted Gemini era 0.25 g peak to almost three times that value (0.7 g).

Scratch the goateed one’s favorite solution.

I know it’s popular and easy for some who don’t understand physics and engineering to say, “well we had problems on program X, and we managed to solve them,” without understanding that there are some technical problems that simply have no solution, regardless of how much time and money is put into them. Sometimes, all the potential solutions simply introduce new problems, or make the old ones worse, and you end up chasing your tail and failing to get the design to close. I don’t know if the Ares thrust-oscillation problem falls in that category, but the possibility cannot be excluded, and so far, it’s looking that way. And when that happens, the only solution is to go back to the drawing board, and start with a different concept.

A good engineer will recognize such problems early, and not waste too much time and money trying to solve them. As Einstein once said, a clever man will solve a problem — a wise man will avoid it. We had, and still have, at least for now, people running NASA’s human spaceflight development program who (as the Brits would say) are too clever by half.

Rocketman also thinks that the writing is on the wall:

All hailed and praised the rocket scientist with more degrees than fingers on our right hands. As the Kool-Aid flowed, we reveled in the plans to renew the minions’ skills for building cathedrals to the sky. We marveled at the safety numbers that flew out of the supercomputers. And we ignored the mounting number of little things that were “normal for a development program.” The steroids flowed, the oscillations increased, and the dollars disappeared.

Fast forward five years with nothing to show for the investments made so far. “Wait a little longer,” they say. “We’re getting our arms around this.” And the congressfolk from Florida push for more money to shower over the falls. And the Senator from Alabama holds up any investments in commercial opportunities so that his precious voters will hold on to their jobs.

But it is already too late. Indeed, the seal on the codex has been broken. All the President’s men do not carry rose colored glasses around with them. They see the forest for the trees.

Clark Lindsey has a summary of a Space News editorial by Bob Bigelow, that would appear to be explaining what should have been obvious to the Senator from Huntsville (who is also the Senator from Decatur, so I continue to be baffled why he doesn’t want to ramp up EELV production).

/– Talks about the overruns in Constellation
/– “Constellation appears to be yet another ill-conceived NASA boondoggle suffering from all too familiar runaway costs.”
/– Talks about the serious technical problems with Ares/Orion
/– Criticizes the reduction of Orion from 6 to 4 passengers.
/– Gives a vigorous defense of the accomplishments of SpaceX and its potential to provide low cost launch access.
/– Says Shelby made an error in ignoring the commercial access capabilities of the Atlas V, which is produced in Alabama.
/– “Bigelow Aerospace has studied human-rating of the Atlas 5 and found the concept to be both viable and economically attractive.”
/– The high reliability of the Atlas V undercuts Shelby’s comments about commercial launchers.
/– “Commercial crew transport, as demonstrated by SpaceX’s dramatic progress and the existing Atlas 5 launcher, represents a viable, affordable, and robust path forward.”
/– Talks about the many launch vehicle project failures at NASA.
/– “Moreover, to hear a Republican senator espouse the virtues of a bloated, costly government program over innovative commercial concepts is so paradoxical that it requires no further comment from me.”

Hey, since when did Republicans care about business, or markets? Not when it doesn’t suit their parochial political interests, for sure.

Norm Augustine and his panel have quite the challenge ahead of them.

No one who has been following the program will be shocked to learn that the major, fundamental design issues continue, and that they aren’t just “teething pains” of a new program. Despite a lot of happy talk from Griffin and Cook and Hanley over the past few months, thrust oscillation remains a serious problem for the Ares I first stage:

According to a NASA blog, the engineers are still looking at putting a series of passive dampers at the bottom of the rocket and a series of spring-like brackets in the middle to soak up the vibrations like shock absorbers.

Originally the brackets, called a dual plane C-spring isolator system, were too heavy to incorporate into the overall design. An updated version uses titanium, which is as strong as steel but lighter.

However, the fixes are not easy and engineers have yet to settle on a solution. According to NASA officials who attended the meeting, the shaking problem is more difficult to combat than originally thought as each solid rocket burns slightly differently.

You don’t say. That means that a passive solution won’t work, unless they can predict prior to flight exactly what the characteristics will be for each SRB (a longer way of saying…it won’t work). They’ll have to have an active approach that can actually measure the vibrations in real time and try to compensate for them. My solution? Bag the solid first stage. Here’s one that will save even more money. Bag Ares I.

And all is not well at the pointy end of the rocket, either:

An Air Force memo obtained by Todd Halvorson of Florida Today indicates that military safety officials are worried that NASA’s Orion capsule and its crew might not survive an emergency escape off an exploding Ares I rocket.

As I understand it, the concern is that the launch abort system is sized to accelerate away from an exploding upper stage, and to outrun an out-of-control first stage, but not from the flack created by the massive explosion of an SRB. Parenthetically (without the parentheses) it should be noted that one of the ways that NASA put its thumb on the scales when it compared Ares to EELV was to assume that the same LAS would be used in both cases, but the latter has a much more benign failure environment, and could get by with a much lighter LAS, so dinging the EELV for lacking the performance to lift an unnecessary weight was stacking the deck against it.

Anyway, how likely is it that the first stage will explode? Well, I find this sadly amusing:

…the article also has Hanley pooh-poohing the Air Force’s concerns, saying that “supercomputer analyses” will prove that the Ares I rocket is a fine vehicle and Orion’s launch abort system will be able to save the crew in the event disaster strikes.

They have top men looking at it. Who?

Top. Men.

Here’s my question. If they know the results of the “supercomputer analyses” before they have performed them, why are they bothering to perform them? Couldn’t they save some money and just skip them?

Florida Today quotes Hanley saying that the statistical probability of an Ares I first-stage failure is remote. He pinpointed it at 1 in 3,000 to 1 in 3,500.

Gotta love that verb, “pinpointed.”

You know, those were the kinds of numbers that they were claiming for Shuttle, right up until around January 27th, 1986. They got some new data the next day, though, that significantly altered the estimates going forward…

So, once again, show us the numbers, Jeff. Show us your work.

It’s hard to know from this brief news story, so I don’t even know what he means by “failure.” Does he mean spontaneously explode without warning? Well, it’s not unheard of for solid rocket motors to do just that, though it has never occurred in the Shuttle program. But I suspect that what the Air Force is concerned about is a different kind of failure — a guidance failure that requires the Range Safety Officer to destroy the vehicle so that it doesn’t hit any uninvolved areas (e.g., Daytona). And considering that an SRB has never had to do guidance without help from a partner on the other side of the tanks and the SSME gimbals, that’s a non-trivial concern. And when the stage is destroyed (by setting off a linear charge along its length) it could create explosive debris that the LAS may not outrun. I assume that’s the Air Force’s (probably supported by an analysis from Aerospace) concern.

Of course, this all raises the question of whether or not we should even have a launch abort system, as I’ve discussed previously, with further thoughts here. Of course, the whole problem goes back to NASA’s “cargo-cult” engineering approach to Constellation, in which they think that if they just go back and do things the way the Apollo gods did, except “on steroids,” they’ll once again have a successful program.

Frank Sietzen has kicked off a discussion over at NASA Watch:

While the present Orion-Ares 1 architecture may well be the “safe, simple, soonest” launch solution promised by ESMD, notice nobody is claiming an Orion-Ares 1 stack will be cheaper than a Shuttle flight. My question to readers: what is the government’s role and responsibility in reducing the cost of access to space? Would you bring back NGLT-or a revamped version of the SLI minus specific vehicle test beds such as the X-33/X-34? How would you revitalize spaceplane research? And would any of you remove funding from existing NASA programs such as exploration to fund research in advanced launch technologies? Or has that ship sailed?

I would certainly remove funding from Ares development for it (because I’d do that on general principles). But NASA wasn’t particularly good at funding launch R&D, as exemplified by the X-33 fiasco. What I would do is get NASA out of the launch development business entirely, and back into the R&T business, and start to view industry as the customer for it, as NACA did. If NASA really wanted to support commercial industry with VSE (as recommended by the Aldridge Commission), it would be doing two fundamental things, neither of which it’s doing much of right now. And no, COTS doesn’t count — it has nothing to do with the VSE.

First, it would be purchasing services, including launch services, from the commercial sector, as it does for unmanned exploration, and stop trying to develop and operate its own dedicated vehicles. Second, it would be canvassing those providers for input as to what high-risk technologies could reduce future costs and increase reliability, and start investing in those. That could include developing X-Vehicles, but they should be true X-Vehicles, each one focused on proving out one or (at most) two key technologies, and not relying on those technologies to be able to fly at all (the grand failing of X-33). They would also be much less risk averse for X programs, and not idiotically shut them down when almost complete out of fear of failure (e.g., X-34). Not to mention demanding that they incorporate some pet NASA project, like a Marshall-developed engine (X-34 again).

There are lots of lessons to be learned from space history, but unfortunately, the space policy establishment seems determined to learn the stupendously wrong ones (e.g., Shuttle proved reusables don’t work, so let’s do Apollo again), and ignore the sensible ones.

Jon Goff has an instructive post on NASA’s supposed risk aversion, and points out that had they taken the same attitude half a century ago, Apollo would likely have failed (as Constellation seems likely to fail today, ironically, because it’s too much like Apollo, but without either the requirements or the management talent of that project).

But this opens up a much broader discussion of risk. There are multiple kinds of risk for a space project (and technology projects in general). There is technical risk — the risk that what you are trying to do may not be achievable within the schedule or budget because the technologies haven’t been sufficiently demonstrated and there are too many unknowns (both “known” and “unknown”). There is program risk — the risk that you may not manage the various aspects (including risk) of the project adequately, also resulting in cost increases and/or schedule slips (it is clear that this has happened to Constellation, as many have been pointing out for years). There is market risk — the risk that the thing that you’re building won’t actually satisfy the need. And for government programs, there is political risk — the risk that your project will lose political support prior to completion (actually sort of the government version of market risk, except that you’ll often find out that the market has disappeared prior to project completion, which is, I suppose, a feature rather than a bug).

A key element of being a successful project manager is managing and mitigating these risks. A lot of it happens at the very beginning of the project, when it’s a lot cheaper to do risk mitigation, and decisions taken have long-term consequences.

In that context. Mike Griffin failed the day that he selected the current architecture, because it had so much risk (of all varieties described above) cooked in right from the get go.

Not having access to the probabilistic risk analysis (PRA) that is a standard management tool for such decisions (I’m being generous in the assumption that one was actually performed, and the decision based on it) it appears to an outsider that no risk was considered other than purely technical. NASA (in defiance of one of the roles stated in its charter) chose the path that was deemed to have the best chance of success because it broke no new technological soil. In its own parlance, it chose to develop launch systems built from components of high TRLs (Technology Readiness Level). That is, they had been demonstrated operationally in the operational environment, in previous programs. Of course, because they didn’t really understand the operational environments (it’s a lot more than just flying through the atmosphere and going into space), it bit them on the ass, and took out a lot of meat (almost doubling the initial estimates of development cost, and slipping the schedule one year per year since it started). “Demonstrated in the environment” includes the environment of an integrated launch system. For instance, the fact that the SRBs only killed one Shuttle crew didn’t make them safe or, in isolation, “human rated.” And the fact that the vibration environment when they were structurally buffered from the crew system by a tank full of a million and a half pounds of propellant was minimally acceptable didn’t mean that it was a good idea to put a much smaller stage directly on top of them.

They probably considered program risk, but assumed that it was non-existent, because they were running the program, and who was better at managing programs than them? They apparently completely ignored political risk, or misassessed it. They seemed to think that the way to maintain program support was to completely ignore the recommendations of the Aldridge Commission — to make it affordable, sustainable, support commercial activities and national security — and instead to cater to the parochial demands of a few Senators on the Hill, particularly Senator Shelby. That one hasn’t bitten them on the ass yet, but it probably will when the Augustine recommendations come out in the fall, removing whatever glutiginous meat remained.

The tragic thing, as Jon points out, is that in avoiding the narrow technical risk of delivering, storing, manipulating propellants on orbit, something that is absolutely essential for future space exploration and space development (because they’ll never be able to come up with a launch system that can do a Mars mission in a single launch), and focusing all their efforts on a perceived “low-risk” but unnecessary new launch system, and in ignoring the systems necessary to get beyond LEO (as opposed to simply getting to orbit, which the private sector has had down for years) they have wiped out more billions in taxpayer dollars, and allowed the day that we would once again go beyond earth orbit, to recede far into the future.

At least, that is, if the lunar-bound vehicle has the word “NASA” on its side.

Fortunately, some of us, more attuned to the real risks, have other ideas.