PD5500 Annex G Applicability

[Spoonful]

G'day All,


Another question here:

PD5500 Annex G Section: G2.2.1 " for values of Cx/r>0.25, the data should be used with caution(see 2.3 of ref[22])"

and also the last paragraph of this section.

I do not have a copy of ref[22] "a review of the method of calculating stresses due to local loads and local attachment of pressure vessels, PD6439" and my case Cx/r is about 0.7, which is way higher than 0.25.

As, Cx=0.85r0, if Cx/r needs to less than 0.25 then r0/r needs to be less than 0.29. Then your nozzle size can not be greater than 30% of the size of your shell. Which really narrows the applicability of this Annex G.


Could someone comment on this? Or comment on what exceptions were given in section 2.3 of ref 22?

Thanks

Spoonful

 

[Karloss12]

Appendix G and WRC 107 are designed to model the stresses in the shell local to a nozzle.

When you have a large nozzle, loads on it will create all sorts of strains and stresses all over the place far away from the nozzle. There becomes a need to analyse the entire cylinder.

This is where other more suitible analysis techniques need to be considered.

 

[Spoonful]

Karloss12,
when you say "other more suitable analysis techniques", do you mean FEA or other method? could you give some reference?

Thanks

Spoonful

 

[Spoonful]

Just another concern about the size ratio, WRC 297 have a d/D ratio limit of 0.5.
and also section G2.8.3.1 of appendix G also have a d/D ratio of 0.5 limit,

Just confused why section G2.2.1 give a Cx/r ratio of <0.25, which gives a equivalent d/D ratio of <0.29, which is far less that what is allowed in others.

Could someone comment on this?

Thanks

Spoonful

 

[Karloss12]

Spoonful,

As described above,

PD 5500 Appendix G and WRC 107 predict stresses around an osolated nozzle that is reletively small compared to the cylinder or head it is attached to.

If the nozzle is to large then the stresses and strain spread far and wide. Engineer/Scientist have not been able to predict or mathmatically model this accurately with a WRC type research paper.

There for your logical next option is to search for research papers that atleast give you guidance on larger nozzles and carry out an FEA analysis. A cheaper more reliable option may be to get the best you can out of WRC/Appendix G and make the head an extra 20% thicker and then load test it.

I understand that one method returns more conservative stresses and in exchange is applicable to a wider range of geometries (i.e larger d/D ratio).

But they are all different methods developed by different people on different continents. This is why they cover different geometries and return different stresses.