How would you analyze this semi-circular annular flat plate stress?

would considering it as a complete annulus be so "wrong" ?

another day in paradise, or is paradise one day closer ?

rb1957 said:

would considering it as a complete annulus be so "wrong" ?

Click to expand...

rb, I would love to hear other's opinion on using a complete annulus for this type of analysis. The case I'm actually considering is two halves that form a full annulus. See sketch:

I guess that's not what I envisioned. I would have omitted the split lines between the halves ... what does completing the annulus get you if you then split it in two ?

are you hand calcing this or FEA ? I assume hand calc, but the sketch makes me think of FEA ??

another day in paradise, or is paradise one day closer ?

Just treat a thin pie-shaped wedge as a beam.
It depends on what you need out of it, and how expensive the item is. If you need to confirm it's adequate, an analysis like that may do it. If you need the thinnest possible design, then you're going to make a million of them, you go with finite element analysis.

rb, JS:

I'm trying to conduct hand calculations to determine if the thickness is adequate. I can't omit the split line in the final product (this is all part of a larger assembly that must be bolted together). Whether, or not, I can omit the split line for the calculations is a good question that I can't answer right now.

FEA is not suitable because I need to make various sizes (different OD, ID, pressure, thickness, etc...) on a project-by-project basis. Firing up an FEA package for every case is simply not feasible.

ok, I'd analysis as a complete annulus (I think your part was the /12, and you were representing my suggestion of a complete annulus with two 1/2s). then analyze a pie sector slice, say 30deg, to see if the results are similar. if not the easy complete annulus is the least represetnative, and expand the pie to a semi-circle (by integration).

you could FEA a couple of bounding designs.

FEA on this should be really simple ... you're drafting this, right? import the solid model (I'm assuming), tet-mesh, load one surface with pressure, constrain the perimeter, run ...
or 2D model it with plate elements.

if you're clever you can build the FEM using parameters (OD, ID, t) and so quickly generate the model (without a solid drafting model).

how big is this ? how much pressure ?

another day in paradise, or is paradise one day closer ?

May be You can check in literature as per Roark : Wahl - Strength of semicircular plates and rings under uniform external pressure - Trans. ASME vol 54 , no. 23 - 1932

or You can use Roarks as You said plus superimposition of table here attached (rotation at boundary=0 ; radial stress at splin line=0)excerpt from : Richard Bares - tabulky pro vypocet desek a sten - 1979 (sorry, but I don't know the english title : my translation is "calculation of plates and beams")

Since you are concerned about pressure I'm guessing the pressure is pretty high, OR the rings are made of rubber or plastic.
A fixed OD suggests welding, or a larger disk clamped between flanges.

How is the inner bore being sealed to //someThing// to maintain the pressure? Or is this a pressure a brief burst, maybe even a pyrotechnic event?
I'm wondering about even sealing the split line.

I'm wondering why a parametric FEA search isn't suitable. FEA typically is for elastic deformation, so the pressure doesn't matter. If it's into the plastic range, hand calcs won't work well. The range of options is not infinite, so set up a script that steps through them (OD ID Thickness, 5 of each/125 cases to define a surface with decent detail.) Hand calcs will not make up for the stress concentrations in the corners.

i'd rather use 3D analysis for this. either FEA or bundle software such as Solidwork, Creo, etc. much simpler, less headache.