According to WRC-368 if the pad width is greater that 1.65 SQRT (RT) then it is assumed that the stresses at the pads edge will be reduced to an acceptable value at that location (i.e, all local nozzle stress effects will be essentially gone). Also some FEA calculation reports have also indicated that at about 9T distance from the edge of the nozzle the local discontinuity nozzles stresses due to piping loads have all diminished so that only general membrane stresses exists.

Question: if this is the case then why should a WRC-107/297 stress check be performed at the R-pad edge at all if the width of the pad is 1.65 SQRT (RT)or greater? I have come across a case where an FEA report calc predicts a stress of about 20,000 psi, which is all just general membrane, at the pads edge but WRC-107/297 predict close to 140,000 psi total membrane + bending for the same configuration and loading (although at the nozzle to shell/pad connection the stresses calculated are in close agreement). Seems like this is erroneous results from the WRC-107/297 calc since even WRC-368 ignores stresses at the pads edge if the pad width is greater than 1.65 SQRT(RT). In my understanding of this, if the pad is of a width 1.65 SQRT(RT) then no stress check is required at the pad edge or at least should not be done with WRC-107 or 297, in paticular since there will be super high erroneous stresses calculated by these programs there.

Could you explain why Caesar II WRC calcs consider stresses at pads edge in light of the above potential for super high erroneous results? And do you agree that if the pad is at least 1.65 SQRT(RT) wide then no stress check is required at the pad edge?

I can see your point of view. If the re-pad is so wide that it covers the area of high stress around the nozzle (opening) then stress beyond the re-pad would be typically membrane stress in the shell/head.

If I remember from WRC-368, T here is re-pad thickness plus vessel thickness and R is mean vessel radius. So, it has to be a very wide pad to see this case. We probably do not need to compute the stress at the edge of the re-pad in this case.

WRC-368 ignores the re-pad if it is less wide than 1.65* SQRT (RT).

Another point, if re-pad is so wide, it may not behave as one unit with shell. Are you adding re-pad thickness to the shell thickness?

Can you share your dimensions?

Thanks for your reply Mandeep. I was out on vacation and I just returned so I could not get back with you last week. the example configuration I was looking at is as follows:

Vessel: 96" dia. 1/2" wall
Nozzle: 24" dia. 1/2" wall
R-Pad: 42" dia. 1/2" thick
Material SA516-70
Temp: 500F

As you can see the R-pad is 9" wide.

The above configuration came from a FEA analysis report I found on the internet at the following site which you might want to look at: www.dynamicanalysis.com/pvp96a.htm

The WRC 368 equation 1.65 SQRT (RT)= 11.43" with R (vessel)= 48" and T=1" using the thickness of the vessel + R-pad. Note that in the sample FEA calculation even though the pad was only 9" wide the local membrane plus bending stresses were still attenuated at the pads edge indicating that the WRC 368 equation above gives conservate pad widths.

I ran the same configuration on the Caesar WRC 107 and 297 and got very high results compared to the FEA report as I indicated in my first post. It appears that the way the Caesar WRC programs models the configuration is what leads to the erroneous results. The Caesar WRC models the R-pad, when calculating the stresses at the edge, as if it were just a rigid pipe attachment with an OD equal to the R-pad OD (in this case 42"). Since the highest local membrane and bending stresses are always located at the nozzle connection to the vessel, then the R-pad considered as just a large nozzle results in these very high bending stresses at the pad OD.

However in the FEA it is modeled just as it is in reality, i.e, a 24" nozzle connecting to a flexible and relatively thin vessel + pad which in this case the local bending stresses are attenuated very quickly so at the edge of the pad they are no longer present (in other words the pad does not act like a rigid attachment as assumed by the Caesar WRC calc).

Therefore to me it appears to be not correct to model the pad as just a larger and stiff nozzle and will lead to beefing up the vessel or trying to lower piping loads when not really necessary.