"Phil, to your point about "fixing" objects to CF structures, I do recall when working with CF masts for some sail vessels having to pay particular attention to how fittings were fastened. Example, we had designed a radar mount for these masts that used compression rings to fix them to the mast, so no holes. Great article, and thanks for bringing science [back] to the table."
.
My reply:
Brian, thanks for reading and commenting. And for the kind words. I am going to copy this to the comments thread on the correct post. Cheers!
The quote "I think that hull has a defect near that flange, that will only get worse. The only question in my mind is will it fail catastrophically or not," Stanley wrote.”
Most epoxy carbon fiber for critically-engineered applications comes as high-temp cure prepreg, which is stored in freezers or walk-in coolers until used in a layup and cured in an autoclave. If mishandled or stored too long the epoxy can take an initial partial cure that will interfere later with a full cure and failure to achieve fully the engineered mechanical properties. So, that makes it a bad idea to use surplus from another manufacturing operation. I repeat my doubt that it was the laminate itself which first gave way. The probabilities are much higher that the proximate cause was failure in or at the attachment of a metal part to the carbon fiber epoxy laminated. Thanks for reading and commenting. -- PLF
Using old 'stage B' (prepreg) is definitely a red flag. I used to build composite stages of Rolls Royce and other large aircraft engines. Even when the stage-B wasn't expired you could tell when it was getting close; the adhesion and overall working qualities would measurably decrease, whereas using fresh material was the chef's kiss for assembly. Also, people whining about compressive strength (in this context) are clueless. If the load is axial on a carbon filament the performance is quite excellent. But old stage-b construction and those Ti caps pushing against the tube that allegedly did not have filament reinforcement in that direction seems very sketchy. On paper, I have a hard time believing that tube simply crushed because 'muh carbon fiber.' Back of napkin looks like it could handle 2x all the way up to 4x the pressure at depth. How one cures a 5" thick laminate though....
Chuck, I agree in the main. The epoxy matrix is going to restrain the long slender fibers from deflecting under compressive load when that load is along their axis , and the fibers are going to take the load because they are stiffer than the plastic. My instinct is that, for a pressure vessel such as a submersible, the layup would have to be 0-90-45 repeated. But we don’t know what, in fact, it was. I’d also expect to use enough CF to a structural weight almost that of steel, which I’m guessing would be about 8” - 9” thick. That thickness would give the skin an enormous moment of inertia (I). But, as you say, achieving a full cure might be problematic, certainly without an electronically controlled autoclave to handle a gradual ramp up and ramp down of temperature. My point is, however, like yours — namely, there would not be anything inherently wrong with using CF provided it was done right. Except, again for the problem of making the connections of metal parts to the CF laminate. Cheers!
Exactly. The tube could/should be sound if properly constructed....BUT, is there really a way to couple it!? On a hunch, the safer strategy might have been to construct the entire vessel from a single wind, as seen on, well, pressure tanks lol. They could have saved big bucks by using a much smaller Ti/acrylic to Ti/carbon hatch with viewport. The money saved on Ti could have gone toward a properly rated window and the carbon could have been wound directly onto the Ti flange as a sort of bottleneck arrangement. The only detractor here is that the passengers would have to climb in through the small hatch; I'd gladly make that tradeoff for a proper hatch that can be opened from within. An adhesive bond-line coupling was a pretty insane gamble, whether it failed or not. I'd really like to know the exact failure mode, if only to better understand proper structural design.
Yep, Chuck, in fact, I also agree with your guess that the best shape would be a sphere. It seems clear to me that, since for a given enclosed volume, a sphere has the least surface area, it would be the strongest shape for any given skin thickness. It would be a neat manufacturing trick to arrange for pretty much continuous layers at angles to one another, but it could be done. a hatch and a window could be let it with a bonded joint that sees the external pressure pressing it into place, working with the seal. Just thinking out loud. Thanks for reading and commenting. Cheers!
Peter, for the record, this is from Robert Tagg, Senior Principal Naval Architect, Hebert Engineering, in the SNAME group:
.
“ Phil, thanks for adding some common sense to the discussion. I'm not a CF guy, but the blanket statements about CF not being suitable for compression loads seemed like nonsense to me. CF-metal transitions and portlight design seem much more likely”
From reader Brian Vlad:
14 hr ago
"Phil, to your point about "fixing" objects to CF structures, I do recall when working with CF masts for some sail vessels having to pay particular attention to how fittings were fastened. Example, we had designed a radar mount for these masts that used compression rings to fix them to the mast, so no holes. Great article, and thanks for bringing science [back] to the table."
.
My reply:
Brian, thanks for reading and commenting. And for the kind words. I am going to copy this to the comments thread on the correct post. Cheers!
From reader Robert Tagg who is the Senior Principal Naval Architect for Hebert Engineering:
.
“Have you seen:
https://www.insider.com/submersible-expert-oceangate-ceo-hull-defect-will-only-get-worse-2023-6
.
The quote "I think that hull has a defect near that flange, that will only get worse. The only question in my mind is will it fail catastrophically or not," Stanley wrote.”
.
Seems to support your material transition theory.
Most epoxy carbon fiber for critically-engineered applications comes as high-temp cure prepreg, which is stored in freezers or walk-in coolers until used in a layup and cured in an autoclave. If mishandled or stored too long the epoxy can take an initial partial cure that will interfere later with a full cure and failure to achieve fully the engineered mechanical properties. So, that makes it a bad idea to use surplus from another manufacturing operation. I repeat my doubt that it was the laminate itself which first gave way. The probabilities are much higher that the proximate cause was failure in or at the attachment of a metal part to the carbon fiber epoxy laminated. Thanks for reading and commenting. -- PLF
Using old 'stage B' (prepreg) is definitely a red flag. I used to build composite stages of Rolls Royce and other large aircraft engines. Even when the stage-B wasn't expired you could tell when it was getting close; the adhesion and overall working qualities would measurably decrease, whereas using fresh material was the chef's kiss for assembly. Also, people whining about compressive strength (in this context) are clueless. If the load is axial on a carbon filament the performance is quite excellent. But old stage-b construction and those Ti caps pushing against the tube that allegedly did not have filament reinforcement in that direction seems very sketchy. On paper, I have a hard time believing that tube simply crushed because 'muh carbon fiber.' Back of napkin looks like it could handle 2x all the way up to 4x the pressure at depth. How one cures a 5" thick laminate though....
Chuck, I agree in the main. The epoxy matrix is going to restrain the long slender fibers from deflecting under compressive load when that load is along their axis , and the fibers are going to take the load because they are stiffer than the plastic. My instinct is that, for a pressure vessel such as a submersible, the layup would have to be 0-90-45 repeated. But we don’t know what, in fact, it was. I’d also expect to use enough CF to a structural weight almost that of steel, which I’m guessing would be about 8” - 9” thick. That thickness would give the skin an enormous moment of inertia (I). But, as you say, achieving a full cure might be problematic, certainly without an electronically controlled autoclave to handle a gradual ramp up and ramp down of temperature. My point is, however, like yours — namely, there would not be anything inherently wrong with using CF provided it was done right. Except, again for the problem of making the connections of metal parts to the CF laminate. Cheers!
Exactly. The tube could/should be sound if properly constructed....BUT, is there really a way to couple it!? On a hunch, the safer strategy might have been to construct the entire vessel from a single wind, as seen on, well, pressure tanks lol. They could have saved big bucks by using a much smaller Ti/acrylic to Ti/carbon hatch with viewport. The money saved on Ti could have gone toward a properly rated window and the carbon could have been wound directly onto the Ti flange as a sort of bottleneck arrangement. The only detractor here is that the passengers would have to climb in through the small hatch; I'd gladly make that tradeoff for a proper hatch that can be opened from within. An adhesive bond-line coupling was a pretty insane gamble, whether it failed or not. I'd really like to know the exact failure mode, if only to better understand proper structural design.
Cheers!
Yep, Chuck, in fact, I also agree with your guess that the best shape would be a sphere. It seems clear to me that, since for a given enclosed volume, a sphere has the least surface area, it would be the strongest shape for any given skin thickness. It would be a neat manufacturing trick to arrange for pretty much continuous layers at angles to one another, but it could be done. a hatch and a window could be let it with a bonded joint that sees the external pressure pressing it into place, working with the seal. Just thinking out loud. Thanks for reading and commenting. Cheers!
Peter, for the record, this is from Robert Tagg, Senior Principal Naval Architect, Hebert Engineering, in the SNAME group:
.
“ Phil, thanks for adding some common sense to the discussion. I'm not a CF guy, but the blanket statements about CF not being suitable for compression loads seemed like nonsense to me. CF-metal transitions and portlight design seem much more likely”