Av Week has a fairly detailed technical description of the thrust oscillation problem:
“Conservative” calculations of the potential frequency and amplitude of a thrust oscillation that could occur in the first stage as it nears burnout, and of the way that vibration links to the rest of the vehicle, suggest that it could set up a resonance that would damage critical components and harm the crew (AW&ST Dec. 10, 2007, p. 60).
A thrust-oscillation “focus team,” convened in November 2007, has since calculated that the problem may not be as severe as it appeared earlier in the fall. But the work continues under a looming March deadline, set so designers on both the launch vehicle and Orion can start work in earnest on mitigating the effect, if necessary, before preliminary design review (PDR) at the end of the summer.
“That gives us a good view of the problem with what we see as how big the risk is, [along with] what are the right mitigation strategies for any residual risk left, so that going into PDR we have a good handle on it and we’re designing for it,” says Garry Lyles, an experienced launch vehicle engineer at Marshall who heads the focus team. “You’re not waiting downstream of the [PDR] to start designing your system to accommodate the oscillation.”
Emphasis mine. If it “may not be,” it also “may be.” This goes beyond risk (which is quantifiable), into uncertainty, which by definition is not, and that’s an unhappy place for an engineer to be. They continue with the “may not be” language.
…the focus team has since calculated that the problem may not be as severe as originally feared. Nominally the oscillation frequency of a five-segment booster is 12 Hz. (compared with 15 Hz. for the four-segment version). But after that it gets complicated. Translating RSRM ground-test data into accurate forcing function figures and the stack’s response to that force is extremely difficult, particularly since the upper-stage and Orion designs remain immature and oscillation data are based on ground tests.
They can do flight tests on a Shuttle SRB, but that still won’t tell them how a five-segment motor will behave (though it will give them better data with which to model it). But as it notes, there’s no way to model the dynamic structural behavior of the stack, because they don’t have enough fidelity in the design. They are risking going into a program, spending billions more, without certain knowledge that they’ll have a viable system until they’re well along in the development, at which point they might find out that they have to essentially start over from scratch.
…if the problem doesn’t go away with more data and more refined calculations, or can’t be fixed with propellant redesign, then isolation pads and other mechanical fixes probably will add weight to the overall vehicle. Making it work could eat into the weight margins held at various levels of the Ares I and Orion programs (AW&ST Dec. 10, 2007, p. 52).
Although the problem isn’t fully understood, none of the NASA engineers involved in solving it sees it as a show-stopper.
“I hope this is the worst we’ve got to deal with,” says NASA Administrator Michael Griffin.
Well, apparently, they’re not allowed to see it as a show stopper. People get fired for pointing out that the emperor is naked.
As Dr. Laura says, hope has no power, Mike. It is not a plan. And there are numerous other solutions.