Metallurgist faked test results for US Sub steel

Nick

ACCESS: Confidential
Joined
24 September 2009
Messages
105
Reaction score
111
Elaine Thomas was the chief tester at Bradken Inc, a firm that supplied high grade metals to firms building submarines for the US Navy. She has now admitted that since 1985 she faked many of the test results on the products they made - apparently to keep the company in business and herself in a job.
Makes you wonder how many subs and ships are still out there with far weaker hulls than expected. How badly does this affect the current USN fleet?

 
From reading about the trial one gets the impression that this was more hubris and stupidity then malice.
One of the stupidest people I have ever had the misfortune to have met had a PHD after his name.
And for the most public example of that lately I could mention a former Governor General of Canada.
 
Elaine Thomas was the chief tester at Bradken Inc, a firm that supplied high grade metals to firms building submarines for the US Navy. She has now admitted that since 1985 she faked many of the test results on the products they made - apparently to keep the company in business and herself in a job.
Makes you wonder how many subs and ships are still out there with far weaker hulls than expected. How badly does this affect the current USN fleet?

They don't have "far weaker hulls." They have cast components which didn't pass their tests. These castings have been getting special attention ince the Navy became aware of this, and at considerable expense the Navy believes they can mitigate/are mitigating problem castings. The hull rings and domes aren't, to my knowledge, of concern.
 
She could have killed hundreds of sailors, as well as the loss of billion-dollar boats, and the possible environmental damage caused by the reactor core in the event of the sinking and total loss of a submarine. The penalties she faces are ridiculous compared to the damage her criminal negligence could have caused.
 
The penalties she faces are ridiculous compared to the damage her criminal negligence could have caused.
There is a reason for this thru...

She is 67 years old.

In ten years she be 77. How much longer is she going to live is the big question.

While I do agree she should have gotten life....

Well actually she probably did. Prison life is rough for young people, let around a granny of questionable health. She at the ages where it cost the state more to keep her in prison then it is worth sadly. They really dont like sendign old folk to the prison cause well, it not worse the hassle.

But she did lose her company retirement and pension apperantly so that may be worse then prison. She going to have use her own savings plus what ever she can scrap up... Thats... Be can be a fate worse then death honestly...
 
It is not just sub steel. It is also ship steel. She should spends the rest of her life in prison!
 
“Ms. Thomas never intended to compromise the integrity of any material and is gratified that the government’s testing does not suggest that the structural integrity of any submarine was in fact compromised,” Thomas’ attorney said in a statement filed on her behalf in district court.

Exactly what drove Thomas to falsify the results of the strength tests is still unclear, but according to the Justice Department, she thought it was “stupid” that the Navy demanded the tests be carried out at -100° Fahrenheit. As a result, the department contends, Thomas changed the results to false positives in some cases.

This was stupid behavior on a couple of levels, the main one being that she was paid to do a job that she just didn't do.

That said, I'm not a metallurgist and demanding that some tests be carried out at -100 degrees Fahrenheit might actually be unnecessary and 'stupid' if you have a better understanding of the material being tested. To fully understand I think you'd have to see the reasoning behind the USN's requirement that this particular test be performed.

I also think an important factor here is that, as far as we know, not a single boat has been compromised or had a problem despite multiple (likely thousands of) depth test dives. Obviously a new risk category may have been added to lifetime and performance considerations for the boats affected, but again, you'd need to see the reasoning behind the test to make a judgement on how bad this actually is.

Last year, however, the Navy declared that it was confident that measures taken to address any possible hazard related to the falsification had been successful. “We have done the work to understand any potential risk, and believe we have mitigated any potential risk for our in-service submarines,” James Geurts, Assistant Secretary of the Navy for Research, Development, and Acquisition said the Navy, told USNI News. “It did cost us some time to go do the exploration to make sure that we were comfortable with the safety of our sailors,” he added.


Sorry, bit of a devil's advocate post here.
 
Another great article from the drive about this
Wish there was more info on this: “according to the Justice Department, she thought it was “stupid” that the Navy demanded the tests be carried out at -100° Fahrenheit. As a result, the department contends, Thomas changed the results to false positives in some cases.”

Were the -100° F test failures, or did she perform less strenuous tests?

And this from The Drive is just ridiculous:

“What seems especially concerning in this instance is that it took the Navy such a long time to realize something was amiss, considering that Thomas began to tamper with test results as early as 1985.”

Unless the USN has someone peeking over the shoulder of all testers and validators they have no way of catching poor work like this.

Final point, is -100° F testing extreme? I’d like to know the rationale behind it.
 


Surface temps in Greenland measured at around -90° F recently. One could conceivably find lower temps closer to the pole, if someone breached the ice cap, for example. So an increased margin of less than seven degrees than the lowest temp (recorded) in the Arctic Circle does not seem unreasonable in the slightest. If I've determined I need X strength at -100° to penetrate the ice with my hull, it seems pretty important to exceed that.

Edit to add: If the Navy did find an issue that effected several of the active submarines built since 1985, does anyone think they'd come out and say that this affects their performance or safety margins? Because I doubt they would be keen to make information of an issue of that scope public. "Oh, by the way, everybody, two-thirds of our submarine fleet have reduced performance envelopes and potentially increased vulnerability caused by manufacturing fraud."
(This is obviously not an allegation this has occurred)
 
Last edited:
Another great article from the drive about this
Wish there was more info on this: “according to the Justice Department, she thought it was “stupid” that the Navy demanded the tests be carried out at -100° Fahrenheit. As a result, the department contends, Thomas changed the results to false positives in some cases.”

Were the -100° F test failures, or did she perform less strenuous tests?

And this from The Drive is just ridiculous:

“What seems especially concerning in this instance is that it took the Navy such a long time to realize something was amiss, considering that Thomas began to tamper with test results as early as 1985.”

Unless the USN has someone peeking over the shoulder of all testers and validators they have no way of catching poor work like this.

Final point, is -100° F testing extreme? I’d like to know the rationale behind it.
We should go to a two facility testing program. No communication between facilities, or just put 3 people testing each sample independently.
 
We should go to a two facility testing program. No communication between facilities, or just put 3 people testing each sample independently.
I have been involved in similar testing, though not completely blind. We tested our own batch samples for production contracts, and also a few of a batch get sent to an accredited third-party lab for the customer's own testing. I don't know what the arrangements are/were for these contracts, but I would be surprised if something similar does not get done here.

If, in this example, there was a batch of steel that was on the edge of spec, then it's entirely possible to get random samples from that batch that all passed, even if there were multiple samples that are sub-spec. Or vice versa. It's a very small sample being tested. You run the chance of getting an above or below average test population.
In the same vein, even in a single plate of high-hard, the metal of the plate is not uniform across the plate. There can be impurities and areas where the concentrations of the alloy are slightly different. If you're riding the edge of the spec, that can get you seemingly incoherent results out of the same batch.

We never used Chinese steel for this reason (but it can be just of true of any steel to a degree). The steel delivered supposedly from the same batch would have wildly different properties from eachother. It was not reliable enough for our purposes.

One other rare issue in our own experience was that our own lab and the other were using the exact same test specs, but there were variables beyond the specification (which one would assume to be nigh impossible given the stringency of the spec, but was not).
 
We should go to a two facility testing program. No communication between facilities, or just put 3 people testing each sample independently.
I have been involved in similar testing, though not completely blind. We tested our own batch samples for production contracts, and also a few of a batch get sent to an accredited third-party lab for the customer's own testing. I don't know what the arrangements are/were for these contracts, but I would be surprised if something similar does not get done here.

If, in this example, there was a batch of steel that was on the edge of spec, then it's entirely possible to get random samples from that batch that all passed, even if there were multiple samples that are sub-spec. Or vice versa. It's a very small sample being tested. You run the chance of getting an above or below average test population.
In the same vein, even in a single plate of high-hard, the metal of the plate is not uniform across the plate. There can be impurities and areas where the concentrations of the alloy are slightly different. If you're riding the edge of the spec, that can get you seemingly incoherent results out of the same batch.

We never used Chinese steel for this reason (but it can be just of true of any steel to a degree). The steel delivered supposedly from the same batch would have wildly different properties from eachother. It was not reliable enough for our purposes.

One other rare issue in our own experience was that our own lab and the other were using the exact same test specs, but there were variables beyond the specification (which one would assume to be nigh impossible given the stringency of the spec, but was not).
I was wondering what would happen if last sentence were to happen in multiple independent tests.
 
An instance of slightly different (and I mean slight!) test methods getting completely different results? It caused a lot of heartburn and back and forth between labs and the customer until they figured out what the missing variable was. Lots and lots of testing and retesting, cross-observation between the labs themselves, people with eagles and stars on their shoulders observing tests.
An epidemic of head-scratching (and an occasional bit of frustrated sniping falling just short of finger-pointing) from everyone because mostly everyone involved just wanted to figure out what exactly was going on and why, even more than trying to determine whether millions of dollars of product produced actually met the spec, because if the testing was somehow broken, ALL the testing to spec performed on a wide range of items might be worthless.
 
The penalties she faces are ridiculous compared to the damage her criminal negligence could have caused.
There is a reason for this thru...

She is 67 years old.

In ten years she be 77. How much longer is she going to live is the big question.

While I do agree she should have gotten life....

Well actually she probably did. Prison life is rough for young people, let around a granny of questionable health. She at the ages where it cost the state more to keep her in prison then it is worth sadly. They really dont like sendign old folk to the prison cause well, it not worse the hassle.

But she did lose her company retirement and pension apperantly so that may be worse then prison. She going to have use her own savings plus what ever she can scrap up... Thats... Be can be a fate worse then death honestly...
Let's remember that, reprehensible as such conduct is in a professional, she is probably a scapegoat for others higher up the food chain. The managers and investors who create environments that encourages corner-cutting and dishonesty--even when they do not apply active pressure--are seldom punished. No one asks what sort of budget/equipment/time did management allow for these tests? How were bonus payments structured? Etc.
 
...it is hard to fathom. You are given certain criteria. You decide it is silly so you fudge the numbers to clear steel that isnt yo specifications. This kind of behavior is incredibly poisonous. It must be punished as it must be made example of. I'm usually lenient and dont want unnecessary suffering but this crap can get people killed. What if I cut corners of certain alloys for cost in constructing a bridge because, well, I thought the specifications were too strict. Fu*$ em this is ridiculous. Cheaper stuff will do.

Poisonous kind of stuff has led to so many engineering disasters.
 
The penalties she faces are ridiculous compared to the damage her criminal negligence could have caused.
There is a reason for this thru...

She is 67 years old.

In ten years she be 77. How much longer is she going to live is the big question.

While I do agree she should have gotten life....

Well actually she probably did. Prison life is rough for young people, let around a granny of questionable health. She at the ages where it cost the state more to keep her in prison then it is worth sadly. They really dont like sendign old folk to the prison cause well, it not worse the hassle.

But she did lose her company retirement and pension apperantly so that may be worse then prison. She going to have use her own savings plus what ever she can scrap up... Thats... Be can be a fate worse then death honestly...
Good for her then that she won't need a dime while she's in prison. Funny how things can work out...
 
Nuytco recently was allowed to use recently declassified submarine steel for their EXOSUIT 2000. This could set them back.
 
...it is hard to fathom. You are given certain criteria. You decide it is silly so you fudge the numbers
Not really, it happens all the time. If the spec check is genuinely safety-important then it needs to be thought through properly so it can’t be fudged by accident or deliberately, otherwise it’s just marketing.

Another example is the mixed Uranium/Plutonium reactor fuel the Brits were selling. First stage QA, fully automated size measurement. Second stage, manual measurement of random samples with manual data entry. Third stage, automated gauge check.
After many hours of repetitively typing in the exact same measurements the robots had already taken, some of the workers decided this was silly and started faking the measurements. After all, they basically never found anything that wasn’t caught by the first check and anything really dangerous they missed would be caught by the third check.

Fortunately nothing actually seems to have gone wrong, but this sort of thing shouldn’t really surprise anyone with a basic understanding of human nature.

 
My background is 'Big Pharma QA/QC', where we had a running battle with target-obsessed production managers...

The most memorable 'gotcha' came from a supplier of aerosol 'dosing' valves. These had a secondary mechanism for back-flow, to pressurise can with propellant after can with product was 'capped'. It also allowed 'continuous discharge' if over-driven by accident or unfortunate user. And provided 'fail-soft' path if over-heated...

Happens we hand-primed cans prior to dosage and 'Twin-Impinger' (TI) testing. That happy day, one of my cans turned 'fire extinguisher' when squeezed in actuator. Then another, then another, another, another...
WTF ??
Re-samples did the same, so wasn't due to box of samples having been, um, dropped and their valve-guts deranged.
Then another batch failed thus. Then another...
We notified our QA/QC inspectors, who echoed our horror.
Documentation found the valves were from same batch, supplier got a 'WTF ?' heads-up.

Anyhow, a day or two later, a call came in from supplier on our lab's central phone. I was nearest, but working a TI rig, with cadence as strict as musket drill.

"Continuous discharge ?" whoever answered it echoed. "Yeah, we're getting a lot of that. Who ? Oh, Nik got the first, then dozens more. Helluva problem. 'Does he Work Out ?' What do you mean ? 'Does he Pump Iron ?' {giggles} Nah..."

The whole lab heard that caller's hapless cry of, "Oh, Sh****t !!!'

Investigation found the supplier had, after decades of testing the 'fill' pressure on aerosol valves, stopped doing it. They'd never, ever had a failure. Not one. That was less than the one (1) IATA fail we'd had in that time, where you dunked filled cans into boiling water per 'boil an egg', for "Shall Not Be Rent Asunder". Whoomph !! Yeah, those old side-seam cans' "unpeeling at altitude on a really hot day" failure mode. Ours were 'deep drawn', but there'd been a tiny flaw in one place in that reel. A microscopic fleck of rust...

The valves ?? Seems supplier had not thought to mention change to their customers. So, no CYA when Implacable Murphy came calling. The 'consequentials' were above our pay-scale, but they certainly resumed testing the 'Secondary Discharge Pressure'. And, we later heard, they quietly picked up the considerable disposal / salvage costs of affected batches. And they replaced every batch of untested valves. And they gave us a very good price, almost at-cost, for next four or five years...

Apparently, the alternative to such 'humble pie' had been prompt bankruptcy, then us buying their assets, minus perps, for 'cents on dollar'...
 
Back
Top Bottom