Stern Measures Keep NASA's Kepler Mission on Track

Hugh Pickens writes "NASA's new Space Science Division Director, Dr. S. Alan Stern, appears to be making headway in keeping in space projects like the Kepler Mission at their original budgeted costs. The New York Times reports that Stern's plan is to hold projects responsible for overruns, forcing mission leaders to trim parts of their projects, streamline procedures or find other sources of financing. 'The mission that makes the mess is responsible for cleaning it up,' Stern says. Because of management problems, technical issues and other difficulties on the Kepler Mission, the price tag for Kepler went up 20% to $550 million and the launch slipped from the original 2006 target date to 2008. When the Kepler team asked for another $42 million, Stern's team threatened to open the project to new bids so other researchers could take it over using the equipment that had already been built."

Comments

They need some corporate sponsorship

[spun]

Put some big old advertising on it, call it Verizon Awesome Space Planet Finder. Offer to let sponsoring corporations name the first earth-like planet found. You'd have funding coming out your black hole, I tell ya'.

Please, for the love of science, don't anyone take this seriously, m'kay?

Re:lowest bidder mentality

[evanbd]

Historically, some of the stuff NASA was trying to get bids on was so far outside the realm of expertise of any possible bidder that no one would have been willing to just eat the cost overruns. So cost-plus contracts were awarded. It's become ingrained, and contractors have realized that they can simply threaten to not deliver if cost overruns won't be payed for.

There are some small aerospace companies that place fixed-price bids on NASA contracts, but none of the major ones do. Many of these companies are of the opinion that taking cost-plus work is severely damaging to the company mindset and correspondingly to its ability to function.

In general, I think fixed-price contracts would be a good idea. If you're worried about paying for a large project along the way, write the contract with intermediate deliverables, like test results from subsystems or prototype versions.

Re:Staying within budget?

[ScrewMaster]

Well, keep in mind that a there is a difference between a commercial part and a part certified to mil spec, and military grade parts often cost a lot more. But yeah, there's a lot of profiteering going on amongst military suppliers, has been for decades. There are various C.O.T.S. (Commercial Off-The-Shelf) programs in our militaries. Their purpose is to seek out non-mil-spec commercially available hardware that can either be used in a military application as-is, or can be brought up to spec relatively cheaply.

Re:lowest bidder mentality

[Anonymous Coward]

Does it truly work to threaten non-delivery -- could NASA not just sue to recoup all development costs for breach of contract? I agree that if there truly is no historical data to estimate costs for a project that cost+ contracts seem reasonable, but I'd imagine that would be the minority of contracts at this point. Aren't NASA's requests still large enough to entice bidders on a fixed-cost contract?

Re:Staying within budget?

[Carnildo]

Didn't they spend some ridiculous sum (> $100) on a nut and bolt once?

For most applications, parts can be qualified for use in batches: take a few parts from the batch, test them to destruction, and if they meet spec, the whole batch is qualified. When I worked in a mechanical testing lab, strength-testing a bolt to destruction would cost (equipment + labor + overhead) $1.50. At the typical ratio of one part tested out of every ten thousand, that's a tiny fraction of a penny per bolt.

Military hardware requirements generally state that each individual part meet spec. This requires non-destructive testing. The company I worked for never did non-destructive testing, but the one time we were asked for a quote, it was $30 per part. If that's typical for the industry, it's obvious why the military was spending $100 per bolt.

Re:Staying within budget?

[ScrewMaster]

For most applications, parts can be qualified for use in batches

Statistical Process Control, yes. You'll still do both destructive and non-destructive testing especially for a military application, but yeah, hundred percent testing is expensive as hell. I did a number of SPC data acquisition systems for fastener manufacturers (self-tapping screws, mostly) and they would typically test 20 parts from a barrel of screws. That was sufficient for commercial use but would hardly be acceptable in a military program. And load-testing a bolt? There you'll have to use a high-powered tension machine (Tinius-Olsen or something on that order, if it's a large part) and those things aren't exactly fast.

There are many other failure modes that a threaded fastener can suffer as well, and depending upon the specifics you might have to test for those as well. That's not including performance testing and design verification either.

Re:Preflight testing was scaled back

[earlymon]

Way to go guys ! You saved $42 million but increased the chance of the entire $500 million project failing due to not enough preflight tests! Good choice there ! Nice one ! Give that man a cigar! Here's a money quote from one of the first sites I'd googled for the initial Hubble failure:

The initial failure of the Hubble Space Telescope is an example of problems caused by relying on computer simulations. In 1990, when the orbiting telescope sent its first photographs back to Earth, the images were unexpectedly fuzzy and out of focus. NASA determined that the problem was the result of a human error made years before the launch: the telescope's mirror had been ground into the wrong shape. The mirror, tested prior to launch like the telescope's other separate components, functioned properly on its own. However, the manufacturers did not actually test the mirror in conjunction with the other components. The manufacturers relied on computer simulations to determine that the separate components would work together. The simulation didn't take into account the possibility of a misshapen mirror.

Because of the Hubble problems, NASA learned "a great lesson" about "the merits of actually testing a system rather than depending upon theory and simulation," explains Doran Baker, founder and vice-president of Utah State University's Space Dynamics Laboratory. From - http://findarticles.com/p/articles/mi_qa3797/is_199810/ai_n8814801

That's just one slice, but not at all the whole story. I get too pissed off even thinking of the early Hubble days to grope further to substantiate, but NASA blew it on many, many levels of saving a buck and avoiding common-sense operational tests - and I say this as an ex-advisor for the Army and Air Force operational test communiities.

NASA learned their lesson, indeed! They say to never credit to conspiracy what can be explained by incompetence - but the NASA corrolary to that is to never underestimate that a great crowd of incompetents can indeed conspire to hide their incompetency to the point of fostering even more of it.

And I've worked on joint Air Force / NASA projects, so no lectures please about how little I know or that I'm flamebaiting - I'm qualified to speak.

I'm always loyal enough to praise NASA after they get something right - but most of the time, I'm happy to grouse in the hopes of educating voting taxpayers. Today, people still don't essentially get that the shuttle was sold to Nixon by the NASA second-stringers - an outgroup - and he bought it because he was being looked down upon by the Kennedy-Johnson NASA crowd. Sorry - this stuff really chaffs my ass.