There’s a lot to chew on in this Av Week article about costs of EELV versus Ares I:

A NASA-funded study found that a human-rated Delta IV heavy rocket could be a cheaper route to the International Space Station than NASA’s Ares I crew launch vehicle.

But the human-rated United Launch Alliance rocket would be less expensive only if the Ares V heavy-lift moon rocket development is deferred, the Aerospace Corp. study reports. And the Delta IV alternative could add two years or more to the “gap” in U.S. human access to orbit if it starts this year, according to the unreleased study obtained by Aviation Week.

I understand that Ares I was supposed to be a “down payment” on Ares V, and that if Ares I isn’t developed, but Ares V is on schedule, then Ares V has to absorb the development costs of those Ares I components that would have been book kept under it, thus increasing Ares V development costs. So it sort of makes sense that it will cost more overall if the money that would have gone into Ares I instead goes into “human rating” Delta IV, while much of the Ares I development costs would have to be paid in addition, with no Ares I for the money. I understand the theory. I think it’s bogus, of course, but it’s hard to know without (as usual) seeing the study assumptions.

Many observers, including me, find ludicrous the notion that a new rocket development could be ready sooner than “human rating” (I’m going to continue to use quotes on the phrase, because it’s meaningless, or means whatever NASA wants it to mean to justify whatever they want to do on a given day) an existing, reliable, insurable vehicle that already carries payloads worth hundreds of millions of dollars.

Ordered by Richard Gilbrech, the former associate administrator for exploration, the $500,000 study evaluated six different versions of the Delta IV heavy as an alternate to the Ares I, which NASA is developing in-house based on the solid-fuel first stage space shuttle boosters, the shuttle external tank, and the J-2 engine used in the upper stages of the Saturn V.

So, they spent about two and a half man years on the study. I wonder where they got the data for it? Presumably, they got Ares I data from NASA (wonder how much that was fudged?) and went to Boeing for the Delta data, as well as Rocketdyne for the engine numbers.

The study did not address the other U.S. Evolved Expendable Launch Vehicle (EELV) – a heavy-lift version of the Atlas V – because of “no clear advantages and several disadvantages,” including the difficulty in obtaining human-rating data on its Russian RD-180 engines.

Again with the “human-rating” mysticism. There is no such thing as a “human-rated” rocket engine. To the degree that “human rating” exists at all, it is a function of integrated system design, not individual systems.

Are they saying that they don’t understand the engine reliability? Or what? Again, this is a rocket that has an excellent record, with no in-flight engine failures of which I’m aware, and certainly none of a nature that would have prevented a safe abort. And as for whether or not they are “Russian,” Aerojet could establish domestic production fairly quickly any time the Air Force (or NASA) decided that was necessary, and were willing to pay for it.

And Atlas has better upper-stage performance than Delta (though probably one of the reasons that they looked at various new Delta configurations was to address the issue of the underpowered upper stage that was a concern during OSP). But if we want to have a robust transportation system for humans as well as cargo, it would make a lot of sense to invest enough to make Orion (or whatever) capable of launching on either vehicle. And I continue to believe, until I see a credible study with assumptions that indicates otherwise, that it would be cheaper to modify both EELVs to carry a capsule than to continue to move forward with Ares.

Now here’s where the numbers start to make no sense:

For the Delta IV heavy using a modified J-2X upper stage engine, the study found no real cost savings over the Ares I. But for the version using four RL-10s, Aerospace Corp. estimated design, development, test and evaluation would come in at about $2 billion less in fiscal 2009 dollars than the Ares I projections. Life-cycle costs for 14 flights of that version of the Delta to low Earth orbit would be about $16 billion, $3 billion less than Ares I in today’s dollars, the study found.

Obviously, since the J-2X is (despite its name) essentially a new engine, it will cost more than using RL-10s. But I don’t understand what they could be doing to the Delta that would cost as much as testing and resolving vibration and other issues with what’s essentially a new SRB first-stage for the Ares. It also seems to me that the Delta upper stage is simpler, because it doesn’t have to provide roll control for the entire stack, as the Ares upper stage does. Furthermore, I wonder if they assumed the same Launch Abort System for both vehicles, when the requirements for that of Ares are much more stringent, resulting in a much heavier and more complex system?

But the most amazing thing in this paragraph is an estimate of the “life cycle costs.”

In March, the GAO had an estimate of development costs alone for Ares I of $20B. So how can the total life cycle costs of fourteen flights be less than that? LCC would include not only development costs, but the fixed costs of operating over however many years those flights occur, and the marginal cost of the hardware for fourteen ships. If we assume four flights a year, that’s three and a half years of operations, for a dedicated pad that can’t be charged off to anything else. If the fixed costs are only a billion dollars a year, that adds three and half billion to the twenty. If the hardware costs only a hundred million per flight (I’m being very generous here, I suspect it will be higher) that adds another billion and a half, so we have an LCC up to twenty-five billion now (and a cost per flight of over a billion and a half each, or more than half as much again as a Shuttle flight — what a bargain). A lower flight rate would, of course, cost even more due to the extension of fixed costs into the out years.

On the other hand, we aren’t told what the “development costs” are for the Delta mods, but I would presume that it would continue to use the same pad that it currently does, which is shared with cargo flights, so fixed costs attributable to the crewed flights would be much less (though there might be some costs assigned from VAB ops in both cases).

Bottom line is that I find it hard to believe that the “life cycle cost” numbers are anywhere as close as indicated here.

Now, the last question is how it could possibly take longer to modify a Delta than to develop a new launcher. “Rocket Man” has a theory:

It was not surprising today, then, that Av Week would be handed a still unreleased report on the viability of EELVs to replace the mis-begotten Stick. Sure an EELV can do the job. And it can do the job less expensively. But guess what? It would take seven years to remake the rocket in human-rated form. Come again?

An already 8 for 9 rocket that carries multi-billion dollar satellites to orbit, has no problem getting insurance today. The design, development, and test phase has been paid for. Walk into your favorite EELV store and offer to buy several dozen and watch the prices fall.

But where to set the bar for human rating? If you are trying to get a new rocket built, and you’re paying for the study, you set the bar obnoxiously high. Only then can a real rocket fall behind in the race with a paper rocket. Without getting into the tit-for-tat arguments that the Italian Waiter’s minions are well studied for, there is only one more thing that our “risk is our middle name” astronauts would like to see over their heads at lift-off.

A reliable launch abort system.

And if a new one of those is going to take seven years to develop, shouldn’t we get someone else to take over the job from the current contractors?

(For the uninitiated, this blogger uses a lot of metaphors — Italian Waiter, for reasons unclear to me, is MSFC’s Ares manager, Steve Cook.)

Sounds right to me (and of course, as Bobby Block points out over at the Orlando Sentinel, this report is already obsolete, since it’s becoming accepted that Ares I won’t make its 2015 date anyway, which means that even if it really takes seven years to get a Delta ready to fly, it still doesn’t really increase the gap).

And this also ignores, as noted above the fact that a Delta LAS might be quite a bit less of a development risk than the one for Ares. It all depends on the requirements. And the assumptions. Something that we continue to not be allowed to see.

[Update a couple minutes later]

Oh, and as for Ares V being deferred? It’s already happening. Hopefully forever.

[Update a few minutes later]

Clark Lindsey wonders what ULA will do. I agree that now that Griffin is gone, the risk of Lockmart losing Orion out of political spite is much reduced. Boeing and ATK stand to lose the most if Ares goes down, but if there are Delta upgrades to be done, Boeing would benefit (and the jobs would be mostly in Huntsville, I’d guess).

[Update mid afternoon]

Clark Lindsey has further thoughts.