|
Jay and Alan,
It might be a usable plan for them to except FEA data,
in a particular format, from what they consider to be properly certified FEA
mechanical engineers. Then ABS could have there own certified FEA mechanical
engineers double check the data, and models. Much like you see the plans
from architects being approved, as well as FEA data given to the FAA.
I've spoken to the lead CAD engineer at Aerocet on
just this matter some time back. He's not a certified engineer, but really
knows his stuff, most of which he's learned on the job over the years.
What he does is produces the 3D parts and/or assemblies, and does some basic FEA
work on them him self, then modifies the design according to what he things it
should be, then checks it out with basic FEA again. This continues until he's
happy with what he sees. Then he sends that data to the certified mechanical
engineers to proof his work. If it needs changes, he makes the changes and then
sends them back to the said engineers. Once every one is happy with the
design, they then submit it to the FAA to be looked over, and approved or sent
back for more work. Once approved they build the parts and/or assemblies and
start the real testing phase.
I have been planning to do some of the above for my
personal subs, that are not classed. I'll do all the CAD work, and as much
FEA work as I can. Then once I've made all the revisions I think need to me
made. I'll send the data on to be proofed by some one with well proven
skills. I'll save myself a lot of money that way, and learn a hell of a
lot in the process. I can export the CAD models data in many universal formats
so they can run there own FEA completely on there own. If may data matches
there's, I'll have a pretty good chance of knowing I configured those FEA test
correctly.
If ABS allowed just one or two particular FEA
software programs, it could simplify things for them. But also exclude other
competing softwares that are always getting better.
Regards,
Brent
From: Jay K. Jeffries
Sent: Sunday, August 30, 2009 2:08 PM
Subject: RE: [PSUBS-MAILIST] FEA Work on Acrylic CT for
KLH-500 Alan, We discussed this article here some time ago. The
problem is not in the FEA programs so much as how the programmer models the
structure...there are many ways to do it with a variety of meshes so it would be
difficult for ABS to certify a given program's output as being correct in all
cases. What ABS has done in this case is significantly reduce the cost of
hull certification by accepting the FEA product that has been verified by strain
gage data. A pretty picture does not verify how accurate the model is
without real world data support for each different model (even different areas
within a model). R/Jay Resepectfully, Jay K.
Jeffries Andros
Is., Bahamas Save
the whales, collect the whole set. -----Original Message----- Hi Jon Below is a link to an article on ABS accepting FEA data
for certifiation on a non standard design. I sent Brent this the other day. http://www.techbriefs.com/component/content/article/1528 There was strain sensor testing done wich verified the FEA
results. Be good to learn a bit more on this. Maybe ABS could
certify various programmes so you'd know you had it right before you submitted
anything to them. Perhaps ABS could analyze everything structural with FEA
& bring their costs right down. Be good to put a question to them as to where they're
going with this. Alan |