[PSUBS-MAILIST] GPR pressure hull

Thu Aug 18 21:13:02 EDT 2016

Hi Sean,Thank you for the great explanation.  Also I had a feeling you could be familiar with the Wrangler sub, since you worked at ISE.  Since posting the question to you, I determined that the cost of a GRP hull is quite cheap material wise, in fact it is cheaper than steel, (maybe)  I concluded that it must be engineering and fabrication that make it expensive.  On another note, I came across a test study of concrete spheres both 16 in dia and 66 inch dia.  The wall thickness worked out to 4.16 inches on the 66 inch spheres.   I think 18 spheres were tested to 4,200 feet. Don't hold me to the figures, it was a lot of information to digest.  The interesting thing is that the spheres were 10,000 psi concrete, hardly high performance.  The other interesting thing is that an opening up to 40 percent of the area is not detrimental, in fact no problem.  The study explains that the opening angle must intersect with the centre of the sphere as the important feature.  Also spheres were built as hemispheres glued together with epoxy.  Even that proved to work well in failure events.  The epoxy held up with no problem, and I was surprised that the hemispheres only had to fit together with a better that 1\8 inch gap.  Amazing!  If you want the study i can send it to you or a link.Thank you again, that was a great explanation, and I understood it lol.Hank 

    On Thursday, August 18, 2016 6:39 PM, Sean T. Stevenson via Personal_Submersibles <personal_submersibles at psubs.org> wrote:

 Hi Hank. I am well acquainted with the Wrangler - I actually did a study on that sub in the late 90s for a potential refit for the Canadian Amphibious Search Team, but it didn't end up making sense at the time.Fiber reinforced composites are tricky when it comes to analyzing their performance in comparison to quasi-isotropic materials like steel alloys.  Composites are strongly anisotropic in single layers, so you need to lay up multiple layers in different directions in order to approximate isotropic behaviour, but then you introduce additional failure modes: fiber shear failure, fiber buckling failure, matrix failure, interlayer delamination, and of course complete brittle failure in shear.In general, you can estimate the strength of a fiber reinforced composite material as a weighted average of material properties in proportion to their relative volume fractions, and this behaviour holds true up to the first failure or yield of the first component to do so. The problem is reliably predicting when these failures will occur. There are computer programs out there that will simulate composite performance, where you enter parameters like number of layers, principal fiber lay direction in each layer, matrix and fiber materials, volume fractions, average continuous fiber length, etc., and it will spit out the failure modes and associated stresses - I have done such analyses on simple parts - flat plates, beams and so forth, but a spherical layup I find intimidating. It is so difficult to obtain theoretically perfect construction in composites that destructive testing is generally recommended to validate predicted failures, and this is reflected in the ABS rules for suitable hull materials. Essentially, they say that alternative materials will be considered on a case-by-case basis, but that results from tests to failure of similar hulls may be necessary to validate performance.If you, for example, have a particular glass fiber of specified mechanical properties, and wish to combine it with a polyester resin of particular mechanical properties at a 60%/40% volume ratio, there are still several ways of doing that that will result in a wide range of possible failure strengths of the finished composite.  This is not to say that GRP or other fiber reinforced composites are not good materials for pressure hulls - indeed, they can be extremely efficient, but each composite part design is an exercise in custom engineering which does not lend itself well to "rules of thumb", or in any case, not to the algorithmic design methodology used by my software.If you are interested in composite construction, you may wish to consult someone with more expertise in that area, or alternatively, if you think you can be consistent in your layup process, build a dozen identical spheres, test ten of them to failure, and then you have a statistically valid sample to give the other two a depth rating.This is, I'm afraid, as much detail as I am comfortable providing.  You can get a ballpark estimate by using that weighted average of material properties in proportion to volume fraction, and then using those numbers to analyze as if it were an isotropic and tough material, but don't expect it to behave as such when you're anywhere near the performance limits.  I certainly will not warrant my software for the purpose.Sean

On August 18, 2016 8:42:29 AM MDT, hank pronk via Personal_Submersibles <personal_submersibles at psubs.org> wrote:
Sean,I have been thinking about GPR sphere's and wonder at what point they make sense, if ever.  When you consider the huge cost for foam buoyancy and factor that into the total cost.  If a sub can be built with little or no need for additional buoyancy, then  the cost of GPR does not make sense.  My sub Elementary 3000 is at that limit.  The cost for foam buoyancy is under 10K, but if I want to go deeper it gets crazy pretty fast.  So it makes me wonder if a GPR sphere makes more sense   than exotic steel or thicker 516-70.    To figure this out I have to start with a sample depth rating for a GPR sphere.  Could you run a couple samples through your program for me? &nbs! p;Iwould like to compare 50 inch ID  at 2 inches thick and 4 inches thick.  I had a chance to look at a GPR sub (Wrangler built by ISE) in Nelson of all places.  I was amazed really, the hull is about 2 inches thick with no rings and the two aluminum dome seats sealed  right on the GPR  at each end of the hull.The dome on Gamma is the spare for the Wrangler sub.  The people selling it now claim it could be certified to 2,000 feet, that is quite impressive for GPR if it is true.Hank

Personal_Submersibles mailing list
Personal_Submersibles at psubs.org
http://www.psubs.org/mailman/listinfo.cgi/personal_submersibles

_______________________________________________
Personal_Submersibles mailing list
Personal_Submersibles at psubs.org
http://www.psubs.org/mailman/listinfo.cgi/personal_submersibles

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.whoweb.com/pipermail/personal_submersibles/attachments/20160819/ca5b9138/attachment.html>