The requirement you are referring to is quite common, but, what's more important, is a code requirement: see UG-98, where essential parts is to be referred to shell and heads, as in UG-99 the code uses the term each pressure element to refer to to all pressure parts.
So I suggest you not to fight anymore such a requirement.
I think the reason for this is that we don't want localized weak points in a vessel, that might be dangerous especially under fatigue conditions.

prex
motori@xcalcsREMOVE.com
http://www.xcalcs.com
Online tools for structural design

It is a well known fact that the most stressed area in a pressure vessel is the nozzle corner (especially on a large size nozzle). It has been proved by finite element analysis also. This corner is also known as CROTCH CORNER...refer PD5500 (BS 5500) which diagramatically shows this corner. It becomes more important when vessel is subjected to cyclic loading.

 

prex,

Thank your for your response so quick.
I saw the UG-98. But it isnt requiring any condition that nozzle shouldn't limit MAWP.
Let's say with an example.

Each part's MAWP is as follows.

design temp : 250 F
design press : 100 psi
shell a : 300 psi
shell b : 300 psi
head a : 300 psi
head b : 295 psi
nozzle a : 180 psi
nozzle b : 150 psi

Then MAWP of this vessel will be 150 psi at 250 F and is limited by nozzle b.
150 psi is enough to meet DP of 100 psi. But if we have to avoid the condition that
nozzles should'nt limit the MAWP, we need to change nozzles' design so that their MAWP's
exceed at least 295 psi. I think it is overdesign. And in some cases, we have to use XXH schedule
nozzles even though STD or XH shcedule would be enough otherwise.
For 4" nozzle, ID of XXH is 3.152", where ID of XH is 4.026".
Because DP is only 100 psi, probably the connecting piping will be STD of XH schedule.
And then the difference between nozzle ID and ID of mating flange becomes very big. I think it's not
a good situation from the process point of view as well.

Kyong

Your point is clear, but of course structural requirements come before any process requirement.
In my interpretation the code requires that MAWP is not limited by nozzles, where it refers to the calculation of MAWP based onto essential vessel parts. Anyway as you have a client's requirement, I suppose you'll have to resign...
Ways to circumvent this requirement are in adopting lower wall thicknesses, increasing the corrosion allowance (for walls only of course and unless you have the same requirement for uncorroded MAWP!), increasing the thickness allowance in heads for forming.
Good luck.

prex
motori@xcalcsREMOVE.com
http://www.xcalcs.com
Online tools for structural design

Many times, we as Users want to insure that the nozzle won't be the item that renders the vessel inoperable since in a process situation, the nozzle will be more likely to have erosion occur.  Also, depending on the environment in which it's installed, there may be factors that will cause external corrosion.  Another scenario might be to better protect against CUI (Corrosion Under Insulation).  
Nozzles are brutally abused in chemical plants and refineries and as such, we in the mechanical reliability realm want to take steps to insure that a nozzle will not shut down the plant.
As for your example, my first reaction is that the vessel shell and head are way overdesigned.  As PREX stated, make the walls thinner and then the nozzle may not be the limiting factor.
You are definitely correct in your thinking that a nozzle could be the limiting factor for MAWP.  UG-98(b) is very clear that the effect of loadings described in UG-22, along with all rules of the Division, are to be taken into account for the MAWP.  The loadings in UG-22 describe those which can only be transmitted to a vessel through a nozzle.  There are rules in the Division that govern the sizing and attachment of nozzles, and these rules make it clear that nozzles are a pressure component, thus, become an essential part.  

Normally the MAWP of nozzles is determined by the reinforcement of the opening in the shell rather than by the hoop stresses in the nozzle neck. Adding additional reinforcement is not expensive.

Checking mechanical integrity of a vessel after it has been in service is simple enough if the shell or head limits the MAWP. There is no "T-Min" calculation for nozzle reinforcement because both the shell and the nozzle neck will corrode, possibly at different corrosion rates.

Estimating remaining life to set inspection intervals also becomes difficult when nozzle reinforcement is a factor. Here again you have two corrosion rates to consider.

Good points, RichardH and Scrobe,
The othe advantage of not linking the vessel's MAWP to a particular nozzle are pipe loadings.  We generally require manufacturer's to "beef" up nozzles to allow considerable piping reaction loads.  With the nozzle safely considered, it is then an easy matter during the life of the vessel, through thichness monitoring of the shell, to assess its fitness for purpose.

Regards

I still can't fully aggree on the concept you guys raised.

Even though a nozzle's MAWP was higher than shell or heads in the beginning of the vessel's service, you still need to worry about the nozzle to re-assess the whole vessel's MAWP after many years because the corrosion rate can be different between nozzle area and shell as you mentioned.

External load on nozzles should be considered quantitatively. Avoiding being a limiting part is not enough.