Multiple skillets 4

[wyogoob]

Say I need a 1 1/2" thick skillet between two flanges to end a 2160 psi piping hydrostatic test section. Is it OK to make the skillet out of multiple plates; one 1" thick plate plus one 1/2" thick plate?

 

[jte]

No, since the thickness of the blinds is limited by bending stress, adding a 0.5" blind plus a 1.0" blind does not equal a 1.5" blind.

Can you really afford to hydro a line for which you cannot afford a 1.5" blind? Seems like an MBA brilliant idea to save $50 worth of materials by spending $500 worth of engineering time.

jt

 

[wyogoob]

I don't agree with the practice and am trying to remove multiple skillets before hydro. Project engineering group has ok'd the multiple skillets. I cannot support my opinion with code criteria. I've just never seen multiple skillets in 35 years of construction.

What code or engineering standard can I reference?

 

[GregLamberson]

I'm like you, never seen it or even had it tried before (25 years).

You might want to turn it around and ask them to justify their decision with calculations and code references. Might mention the buzzwords - liability, risk, litigation.

Greg Lamberson
Consultant - Upstream Energy
Website:

http://www.oil-gas-consulting.com

 

[dcasto]

If you can't use multiple blinds, then how can we make clad vessels. I have a vessel with 3 - 3/8" steel shells stacked together, the last is 304 SS. The design for pressure would have called for a 1" plate rolled.

 

[TGS4]

Layered vessels retain the pressure through membrane action, and successive yielding of the inner layers (you're actually referring to layered vessels - not clad vessels).

In the case of a blind flange, as jte said, the pressure is retained by bending action, so the successive yielding of the different flat plates would lead to a cumulative effect, as you might in layered vessel construction.

wyogoob, as far as justification, you can start with Code of construction. After that, I would take a look at what ASME Sec. VIII, Division 1 says about layered vessels, and what you need to do for flat heads.

 

[jte]

wyogoob-

Have the project folks buy two 2x4 x 8' studs. Add a 4x4 x 8' to the list (same flavor of wood). Take the project guys outside and set up the two 2x4's stacked 4" side to 4" side (the weak direction) with bricks or similar at each end. Set up the 4x4 with the same supports. Stand in the middle of the 2x4's. Have the project folks measure deflection. Now stand in the middle of the 4x4's. Have the project folks use their measuring tapes again. End of discussion.

If the project folks do not understand bending stress then they really should not be implementing designs which have a bending failure mode. You are absolutely right to question their approach and dig your heels in to oppose them. Presumably you calculated your 1.5" value using B31.3 304.5 or pulled the thickness from B16.48 (note new 2005 ed.) which uses the B31.3 equation. This is not some esoteric code equation with no backup. It is straight from Timoshenko who used Poisson's (heard of him?) solutions for the basic differential equations which govern the (linear) behaviors of circular plates subject to pressure.

Now, you also stated that the application was for hydrotesting and not for process service… If you are willing to go nonlinear and accept some deformation you might find this paper interesting:

http://store.asme.org/product.asp?c...apers&category_name=&product_id=PVP2005-71164

jt

 

[wyogoob]

jte,

Great example.

I am familiar with S.P. Timoshenko's old work; THEORY OF PLATES AND SHELLS. I am also familiar with ASME paper pvp2005-71164, SLIP BLINDS AT PRESSURES CAUSING PERMANENT DEFORMATION, by J. Taaepera and FLOUR's T, Seipp. Their work curtails how to judge minimum thickness for skillets from destructive tests and FEA.

Today we used two 1/2" thick plates where B31.3 304.5 called for a 1" thick.

 

[jte]

wyogoob-

Presumably the test with two 1/2" plates turned out ok, with no bending of either plate. Out of curiosity, what thickness plate did the paper recommend for your pipe size and test pressure for a 1/8" residual deformation?

jt