Re: SIF on elbow weldolet

Chris wrote, "Like the value associated with tees
> and laterals. The SIF is usually built into the element formulation along
> with the stiffness correction."

We think this is kinda true. User's of pipe stress programs can also enter SIF's of
their own, which are applied after the solution. They really don't affect the element formulation, (although this may just be a matter of symantics).

"> Given that the fitting includes a fillet reinforced butt weld I really don't
>think FEA is going to tell you
> anything valuable after you have made all the necessary guesses about the
> weld contour, metallurgy and residual stress.

This is often true too, but in the olet case, the olet serves as a localized stiffener, and depending on the d/D ratio, can influence the high stress on the side of the bend due to ovalization. Local effects due the restriction of
radial dilation can also be quantified, if only roughly. Since the B31 codes do
not give SIF's for external loads through these components, nor guidance for the pressure design of these parts, the "approximations" from an FEA tool can at least give answers better than a "guess."

> A SIF value obtained from
> testing will take all of this into account. I've been doing FEA for about
> 30 years, and I think trying to model a welded connection like that
> explicitly is an exercise in futility.

We've worked with several olet manufacturers putting olet models into FE/Pipe (both brick and shell models) and as long as the analyst understands that the olet does not always remove the high stress, but redistributes it in some manner to the often thinner section, the FEA tools have been effectively used to produce good design guidance for the various product lines we've been associated with.

It can't be stated universally, but when the Tolet/T or Tolet/t ratios are large (which is usual), the high stress is in either the attached pipe or the
bend, and an FEA with a locally thick "staunchion" seems to give pretty good predictions of the stresses in the vicintiy, (at least as accurate as the SIF for any other welded component). We have just recently evaluated a furnace tube failure of just such a connection with good results. (Good, in the sense that the high stress was shown where the failure occured, and that the major tensor components were normal to the line of crack propogation.)

1. Paulin

>
> Looking at the construction at the Bonney Forge web site (Elbolet is a
> Bonney trade mark, not a generic name, BTW.)
> <http://www.bonneyforge.com/specs/branch/elbolet.cfm> it looks to me that
> the stress distribution is pretty much the same as a single welded
> coupling welded into a larger pipe. You can probably get guidance from
> the manufacturer or perhaps from a good handbook on piping stress
> analysis.
>
> Do a Google search on Elbolet and get in touch with some distributors.
> This web site has some chatter on the subject
>

<http://www.eng-tips.com/gviewthread.cfm/lev2/16/lev3/58/pid/378/qid/71234>
>
>
> Christopher Wright P.E. |"They couldn't hit an elephant at
> chrisw@skypoint.com | this distance" (last words of Gen.
> ___________________________| John Sedgwick, Spotsylvania 1864)
> http://www.skypoint.com/~chrisw
>
>
>
>
> =========================================
> PipingOffice - Excel Spreadsheets for Piping Calculations
> http://www.pipingoffice.us/
> =========================================
> Main site: http://www.pipingdesign.com
>
> Yahoo! Groups Links
>
>
>
>
>
>
Received on Tue Mar 30 13:59:00 2004