Tourist submersible visiting the Titanic is missing Part 2 69

[SwinnyGG]

Who said I was defending?

No one said you are defending - it's not implied. It is literally your direct response to a comment which pretty gently questioned the decision to use rhino liner to coat the hull - which against is completely untested solution for a critical aspect of the hull design.

Don't fall for the sensationalism. Rhino liner wasn't a structural component of the hull and PlayStation controllers didn't cause the collapse. Being able to incorporate commercially available components into a design is an engineering talent, not a fault.

 

[Tomfh]

Tug, that connection looks exactly like the “watertight” connection to my pool robot. Hence the advice to never leave the robot in the pool.

 

[Murph 9000]

Rhino liner wasn't a structural component of the hull …

Not directly, but it could possibly be considered holistically critical to the structure. If sealing the outer surface of the CF against the 5500psi seawater is critical to the integrity and longevity of the structure, it does become a critical part of the structural design, despite not adding significant strength or rigidity.

Personally, I think I'd have chosen something like one of the epoxy hull primer/filler products from International, rather than Rhino; for an outer coating to seal and protect the CF. They won't be rated for use at extreme pressure, but likely would form a good bond with the CF.

 

[blakmax]

I have had a week off travelling... a benefit reserved for the retired. I note the simulation provided by IRStuff, and I make the comment that this representation depends yet again on the composite failure theory used in the model. Again, if it is the Tsai-Hill or Tsai-Wu model the results is totally fictitious.

With regard to the ability of the adhesive to transfer sufficient load to enable the titanium to fracture rather than the adhesive, that is totally incredible. The load capacity of an adhesive bond depends upon the relative stiffness of the adherends (EiTi/Eoto) and I can assure readers that a 5 inch thick CF structure will always cause the adhesive to be the critical member.

The other question that must be asked is the design methodology for the adhesive joint. There is a common belief that the "strength" of a bonded joint is measured by a lap-shear test such as ASTM D1002. That test does NOT under any circumstances generate design data. That test typically uses a standard thickness of adherends and a standard overlap length. Change any single parameter and the results will change. Unless the design matches the specimen geometry any prediction of bond strength on the basis of lap-shear strength is totally meaningless, especially using FEM analysis.

The other issue is that adhesives are dominated by elastic-plastic behaviour, not just elastic behaviour. Typically up to 60% of the strain energy to failure is plastic behaviour. So taking a badly derived allowable strength from a lap-shear test and using an elastic analysis (typical of FEM) to predict strength will have little relevance to the actual structural strength.

Next, is the relationship between temperature and adhesive strength.Adhesive properties change significantly with temperature; high temperatures result in low shear modulus, low plastic strength but high strain to failure. Low temperatures cause high shear modulus, high plastic strength but low strain to failure. Unless your FEM actually uses data at the relevant temperature, and unless the analysis has an elastic-plastic capability (rare) then your bond strength estimates have very little relevance to reality.

Next, while Tugboat's suggested modification of the overlap has some merit, in practice unless there is some symmetry in the joint (fir tree joint or extended double overlap design) there will always be a very significant bending moment generated by the mis-alignment of neutral axes and that will always risk failure of the vertical bond face, and a5 inch thick CF structure will almost certainly result in tensile stresses at surface loads and temperatures. I suspect that the only viable configuration would be a fir tree joint, but that would need careful manufacturing procedures to generate the structure in practice.

Blakmax

 

[blakmax]

Wow.

I didn't mean to kill the discussion with my last posting!!! Sorry Guys and Gals.

My background: I have been working in composites and adhesive bonding since 1972. (I think Moses was still in shorts then.) I am well and truely retired but I have extensive experience and expertise in composite and adhesive bond failure forensics, and a different perspective on adhesive bond damage tolerance than the usual accepted mantra that fails to distinguish between analysis of defects that may lead to cohesion failure at high loads and those that result from weaker mixed-mode failure and very weak adhesion failure. I spent years between 2013 and 2020 delivering courses on this technology, until COVID stopped my activities. I am now almost 76, and I would re4ally like to pass on my knowledge to the wider adhesives and composites community.

In the current discussion WRT the submersible, there are many issues yet to be examined. For example,
What was the surface treatment for the titanium surfaces prior to bonding?

Were there appropriate tests to assess hydration resistance provided by those treatments?

Importantly, all fibres are treated with a "size" to prevent damage during manufacture. It is essential that that size material must be compatible with the resin system such that the fibres bond to the resin. In compatibility will result in adhesion failure between the fibre and the resin with a loss of shear, tension and compression strengths. I have seen examples of this problem that resulted in substantial reduction in component service life.

We don't know if or how the finished CF shell was checked with NDT, and if any defects found were "repaired". Those who know me know very well my opinions on "injection repairs". (you only achieve two things by injecting fresh resin or adhesive into a void: You hide the defect so that NDT can no longer detect the air gap, and you get a warm fuzzy feeling that you have fixed the problem. The surfaces of the voids are chemically fully reacted, so you can never achieve a viable bond.)

So my questions to the discussion group are:

Would it be of value if I placed my lecture notes for my course for open access?
If so, where should I post the files?

I have often said that with composites and adhesive bonding, it is a stupid ass that trips on the same stone twice. I reiterate the point that deciding to use cheap inexperienced engineers so that you do not inhibit free thinking really leaves you open to being an ass. EEAAW EEAA Eawlways says that!

Regards

Blakamx

 

[SwinnyGG]

I don't think you killed the discussion... there's just not a ton of new information and everything that's publicly available we've already steamrolled in the discussion. I expect if/when new information is made public this thread will get hot again.

What was the surface treatment for the titanium surfaces prior to bonding?

There are a couple of paths to activating titanium and/or resin surfaces to prepare for adhesive bonding (based on your stated experience, expect you're well aware of this already). At this point, I think the group consensus is that we suspect none was used, but we can't prove that conclusively.

Were there appropriate tests to assess hydration resistance provided by those treatments?

Similar to the above... I think we all suspect the answer is no (I certainly do) but we can't prove it.

Importantly, all fibres are treated with a "size" to prevent damage during manufacture.........

We don't know if or how the finished CF shell was checked with NDT, and if any defects found were "repaired".

Given that the hull was wound by a very reputable and experienced manufacturer of composite structures, I still maintain that failures due to a deviation from standard practices in the carbon fiber section itself are not likely. Certainly possible, but not likely. The most likely failure mode in my opinion is at the adhesive joint, and is due to design errors by the OceanGate team and not by the manufacturer of any one component of the hull.

But, maybe with your experience you can answer a question I've been thinking over... my understanding is that the composite house was given specifications on maximum pressure limits etc to design the composite section against. Also that they are highly experienced and well reputed for filament wound structures, but not for filament-wound structures for high pressure underwater applications. I tend to lay blame at the feet of the OceanGate staff for selecting the wrong material for the application, and then dictating design terms to the composite manufacturer, who essentially just 'followed the drawing'.

Question is, in the world of commercial composite structures, would thorough NDT be expected as part of their in-house quality control process, regardless of whether or not it was specified by the client, or is that the sort of thing that you would expect to only receive if it was directed on plans/by contract?

I have been running on the assumption that any specific porosity/density requirements needed to be directly specified by the OceanGate engineers, and that if the composite hull section was not heavily inspected via ultrasound or whatever, it is simply because the OceanGate team failed to specify quality requirements correctly.

 

[btrueblood]

We don't know if or how the finished CF shell was checked with NDT, and if any defects found were "repaired".

We know that Stockton Rush fired his chief engineer (and threatened him with a lawsuit if he talked) for suggesting the sub was unsafe to dive unless periodic NDT was performed.

Would love to see the course notes, and to have them preserved for future engineer's use. Dunno where they could be kept though.

 

[SwinnyGG]

We know that Stockton Rush fired his chief engineer (and threatened him with a lawsuit if he talked) for suggesting the sub was unsafe to dive unless periodic NDT was performed.

Fair point, I had forgotten about that wrinkle in the chief engineer saga when I typed my long response above.

If I'm remembering correctly, though (feel free to correct me otherwise...) that was in the context of ongoing NDT to continuously check the hull for accumulated damage, and not necessarily a one-time NDT check of the completed part prior to accepting it from the manufacturer. Maybe I'm remembering wrong or maybe it wasn't clear in the history we have available.

 

[IRstuff]

I still maintain that failures due to a deviation from standard practices in the carbon fiber section itself are not likely.

That's a pure speculation since the manufacturer simply wound to Oceangate's (Rush's) specifications; if they were insufficient, which is a highly probable case, then it most definitely could have been simple failure of the composite, even as the manufacturer did everything to industry practice. Oceangate did no testing whatsoever to validate their design concept of using that specific winding approach and thickness of the finished winding.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[JStephen]

I don't think the guy that was fired was the chief engineer, more like the chief pilot.
There are quite a few videos on Youtube relating to this, but mostly all just rehashing the same limited information.
It seems obvious that there are quite few issues that COULD have caused the failure.
And I expect here in a few months or year or whatever, the NTSB or whoever else will opine on what DID cause the failure.

One issue that I've not seen addressed. There was a supposed leak of the transcript between submersible and surface (with some speculation as to whether or not it was genuine). Assuming it is genuine, one item of concern there was that when they tried to surface, and jettisoned everything they could, they were still barely moving upward. So either some kind of miscalculation on the weight had taken place, or some space had flooded that wasn't meant to flood. Or maybe the CF became waterlogged, if that's possible? Anyway, that seems a most curious occurrence and makes me wonder what the heck was going on there.

 

[Brian Malone]

Any continued discussion of the 'leaked' transcript seems to me to be wasted effort and falsely adds validity to specious data. If that transcript had any validity, corroboration to known sources would be available. That whole thing is/was to drive activity to a YouTube channel to generate views and ad dollars. Shameful monetization of a tragedy . . .

 

[SwinnyGG]

That's a pure speculation since the manufacturer simply wound to Oceangate's (Rush's) specifications; if they were insufficient, which is a highly probable case, then it most definitely could have been simple failure of the composite

Yep, agree that I'm speculating. I'm not saying the carbon fiber didn't fail, I'm just saying that I suspect if it did fail it wasn't due to defects outside of what would be expected to occur during production of a filament wound carbon fiber part adhering to industry standard best practices for that construction.

We're saying the same thing though. What I mean is that my suspicion is that the hull was likely correctly manufactured, per the normal best practices of the filament wound carbon fiber structure process - but that it was correctly manufactured according to a bad design.

Part was built correctly, to match a drawing that was wrong. IE it's likely a failure due to OceanGate's design process, and not a manufacturing error during production. That's based on the assumption that OceanGate's contract with the manufacturer was 'build this filament wound carbon thing according to this specification which we came up with'. That's a lot different than 'design and build a carbon fiber filament wound hull section for a 4,000m submarine, and include all engineering to ensure long term durability under exposure to deep sea submerged conditions'. We don't know for sure, but based on the information we have so far it seems that the arrangement was #1 and not #2.

 

[MintJulep]

@blakmax

I personally would very much appreciate being able to read your lecture notes. Your expertise is evident in your posts here, and the few papers that you have linked to in the past.

I commend your willingness to pass your accumulated experience and knowledge along to the rest of us.

As for where:

Did you lecture in association with an association, or a university? If so, do they have a suitable web place for your notes?

You could create your own website, blog etc. and post there, although that might cause more effort and recurring cost then you care to take on.

You might find a good home at one of the places listed here:

https://en.wikipedia.org/wiki/OpenCourseWare

 

[TugboatEng]

I'm interested. I am proposing an adhesive option to mount various markings underwater on steel vessels. It's similar to draft markings. I need something that tolerated permanent immersion with SSPC SP-3 prep. I'm looking at acrylic adhesives due to the large number of installations. I'm curious about primers to prevent hydration. I keep seeing interesting things about siloxanes.

 

[blakmax]

Hi MintJulep

In 2006, I was engaged as a consultant to a helicopter crash in Fiji. It occurred on the same day as there was a military coup and therefore the Australian and New Zealand governments withdrew the usual formal investigative support from government organisations, so CAAFI engaged a consultant who became frustrated with the advice from industry companies involved with helicopter maintenance and through a contact in the NZ Defence Testing Authority, he contacted me while I was working with the Australian Air Force. In 1996, the FAA conducted a review of 36 crashes of this particular type of helicopter, and they could not find a common cause.
The Investigator In Charge had eliminated the usual causes but could not exclude failure of the main rotor blade.

Within one day of examining the wreckage in Fiji, I advised the IIC of what was the probable cause of the blade failure, based on my experience with failure assessment of bond failures on RAAF F-111 structural construction methods and repairs. My findings were supported by the NTSC investigations and my curse work was a result of collaboration between my company Adhesion Associates P/L (now closed) and the FAA. I acknowledge the support of the FAA and EASA in the support of delivery and development of my course.

As background, in 1992 I had undertaken an assessment of adhesive bonded repairs and found that 40% of repairs repeated the same deficient practices (Donkey, same stone...). With regulatory support, we changed the surface preparation process to address hydration resistance. We improved training such that the technicians understood the fundamental mechanisms of adhesion such that they also understood the consequences of shortcuts in process procedures. We also changed the adhesive used for repairs. Now, changing the adhesive does not make the fantastic difference that adhesive salesmen would have you believe, but in our case the colour of the different adhesives enabled us to be sure of the period when repairs were performed. Between 1992 and 2007 when I retired from RAAF employment, the repeat repair rate fell from 40% to 0.06% and because of the quality management systems we had established, we could tell exactly where the technician had taken a short cut.

To be honest, the many years of implementation of bad technology has generated a poor opinion of a really good technology which if implemented correctly could dramatically improve structural integrity. What is needed is education of how adhesive bonds from and in particular, the causes of adhesive bond failures both short term and long term. I'll take your advice about the web site.

Thanks

Max

 

[dik]

It would be appreciated here, too... thanks, Dik

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[btrueblood]

What a crap post, dik. Reported.

 

[dik]

Post deleted, but, pretty much what I expected... and in milliseconds.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[JStephen]

One of the things that makes this story of interest is that it is considerably out of our normal realm of experience. I'm not sure if anyone has actually witnessed, by video or other means, the actual implosion of something this size under this much pressure. And I don't know that anyone has actually examined the aftermath of a similar implosion under this much pressure- other submarine failures have presumably been much shallower, with more ductile hulls. So the question of "what exactly happens to the people" is bound to come up. I think earlier in the discussion was the assertion that there would be no remains recovered, followed some time later by the reported recovery of remains, so apparently there are some unknowns even on that aspect.

 

[LittleInch]

Without getting ghoulish "remains" means some matter which is identifiable at microscopic level as human tissue or bone fragments.

Not like digging up a shallow grave.

Airplane high speed crashes like we've seen recently are exactly the same. Think Of that German Wings plane which deliberately nose dived into the ground at 500 MPH in France. That's what the rescue teams have to deal with. People are essentially some solid and fibrous mass with a lot of water. The water would go away but the other mass remains, just not in any recognisable form. That's my understanding at least.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.