MAWP vs. Design Pressure 5

[jproj]

Recently while trying to finalize a nameplate issue, I had an interesting discussion with one of our fab shops. They kept showing the vessel design pressure (50 psig) as the MAWP on their code nameplate drawings. The code calculations showed an actual vessel MAWP to be much higher (~140 psig). After several discussions, our fabricator said their A.I. refused to stamp anything other than the design pressure on the nameplate "MAWP" section. Vessel was designed to ASME Section VIII Div. 1 (current ed.).

While this isn't the first time I've heard of this practice, I'm curious about the reasoning behind it. As I understand it, the MAWP is a product of actual materials / thickness used for fabrication and really has very little (if anything at all) to do with the design pressure (unless of course components are designed to have a MAWP equal to the design pressure).

Any insight?

Best Regards,
Jproj

 

[deanc]

See SecVIII Div1 Appendix 3

 

[jproj]

Please elaborate. Per the last sentence of the MAWP definition (sorry if this is old, but I only have the 1995 edition to reference):

"The design pressure may be used in all cases in which calculations are not made to determine the maximum allowable working pressure"

In this case (as explained above), calculations were made and the MAWP was determined. Are you referring to something else? The rest of the MAWP definition is in accordance with my previous understanding of vessel MAWP.

Regards,
Jproj

 

[TomBarsh]

Due to the practicalities of construction and fabrication, the vessel MAWP will likely be somewhat higher than the design pressure. eg: if the required shell thickness for design pressure plus static head is 0.6123", it wouldn't be unusual to specify 5/8" (0.625") plate unless you are placing a mill order to 1/100ths of an inch. This slight thickness over the required thickness then provides a higher MAWP, if analyzed. Of course, all components in the vessel would have to have their MAWP determined, including nozzles, cone-cylinder junctions, etc, which are not as elementary to determine MAWP for as are cylinders and heads.

To have a vessel MAWP that is almost 3 times the design pressure indicates a very conservative design and might mean that money could've been saved in the construction. But there's nothing really wrong with hell-for-stout.

The real issue here for stamping the nameplate is what is the test pressure based on? Consider a hydrostatic test. Per UG-99(b) the test pressure is based on the chamber MAWP or the design pressure if the MAWP is not calculated.

If the test is performed as based on the chamber MAWP (eg: 1.3*MAWP*stress ratio) then the nameplate may legitimately be stamped with the chamber MAWP. It would be inconsistent to require a test based on the MAWP but allow the name plate to reflect only the design pressure.

 

[djack77494]

Vessel fabricators, in my opinion, are prone to balk when asked for a vessel's MAWP. To some degree, that makes sense because the MAWP is sort of a "guarantee point". The fabricator warrants that the vessel meets code requirements for the stated MAWP pressure. Typically, they don't want to guarantee any more than they need to (would you?). The buyer is typically only concerned with a vessel capable of withstanding the pressure he/she has specified - the design pressure. In times past, there were extra efforts in calculating the MAWP, and the fabricators wanted to be paid if you required those extra steps. With modern vessel design software, the extra work is minimal, and the fabricators should be able to easily extract the MAWP. If they feed you back the design pressure and call it the MAWP, that means they are NOT supplying you with the real MAWP; there is always some difference between the two.
Doug

 

[athomas236]

In my opinion the MAWP is the design pressure unless calculations are made to determine the MAWP. If calculations are made then clearly whatever the calculations show is the MAWP.

The difference between a MAWP of 140psig and a design pressure of 50psig may seem to be excessive but sometimes the practicalities of having materials of a thickness that can be handled in the works takes precedence.

That's what I remember from my days as a pressure vessel designer, although it was 40 years ago.

athomas236

 

[gr2vessels]

The MAWP is the lowest pressure which one or several pressure vessel components can withstand under the basic design conditions.Typically must be higher than the design pressure, because of reasons listed by TomBarsh above.
The 140psig compared to 50psig seems a bit odd, however, is most likely to be the rating of a flange, rather than a shell section rated for the 50 psig design pressure (or any other weaker component).
However, a carefully calculated MAWP for all the vessel components is unlikely to be much higher than the design pressure and I'd be worried with a high one (incomplete calculations, over-design, incompetence???) picked-up from blue sky. Sometimes, the fabricator is too lazy to do the calcs, particularly the manual ones and as noted above, will nominate, incorrectly, the design pressure as MAWP. Please note that the up-rating of the vessel for any reason, without correctly calculated MAWP at the design stage will be very difficult, if not impossible.

Cheers,

gr2vessels

 

[chaulklate]

As it works in my area (I will assume it to be the same everywhere) the pressure stamped on the nameplate is the presure used for "design" during calculations. If the calculation package indicates 50 psig as "design", then it is 50 psig that goes on the nameplate. If someone would like the unit rated to 140 psig (or other calc'd MAWP), all calculations must be done again using 140 (or calc'd MAWP) as "design". Using a design of 50 psig, and having the calc report a MAWP of 140 psig does not necessarily mean you can stamp the vessel to 140. This is why most fabricators don't like registering to MAWP, it means they have to do their work twice, and it also erases any safety factors they have built in beyond code requirements.

 

[jproj]

Thank you all for the responses.

Regarding concerns about the MAWP being much larger than the design pressure, it is a combination of several other design factors (full vacuum, allowable nozzle loads, MAWP limited by head / shell, etc.).

It seems like the consensus is in accordance with the Appendix 3 definition of "MAWP"... if calculations are made, show MAWP, if not, show the design pressure. This makes sense to me, but I'm still confused at their AI's reasoning (since it seamingly contradicts the above referenced definition).

Regards,
jproj

 

[jproj]

chaulklate,

This is exactly what their AI said. Can you explain why the caluclations must be run twice? It doesn't make sense to me. Using the same equations to calculate different terms should give you the same result (if they don't I think we have bigger problems...).

The way I see it, entering the actual vessel components into the calculations and calculating the MAWP is the same as entering the MAWP as the design pressure and calculating the minimum thickness required.

If I'm not looking at this correctly, please correct me.

Regards,
Jproj

 

[jte]

The hazard I see with not running the calc's twice is that some of them are interdependent. If, for example, you run an opening reinforcement calc and take credit for the excess area in the shell based on the DP of the vessel, without checking for adequate reinforcement based on the MAWP, you could find yourself in a trap. Here's how I usually see it done when the design process is done well:

Necessary design pressure and design temperature are determined by a process engineer.

This DP/DT is used as a basis for the design by the mechanical engineer for the major components: The heads, shell, and perhaps other large cost items such as body flanges.

The MAWP of the major components is determined. (One could actually optimize one or a combination of the DP, DT, or corrosion allowance. I have a bias towards optimizing corrosion allowance based on my experience.)

The chosen MAWP (and corresponding DT and CA) is used as the new design basis. The major components are re-run with the MAWP as the design basis and the minor components such as nozzles are added to the calculation at this point.

The design is evaluated one more time for MAP (max allowable pressure, new and cold) which is then used as the basis of the hydrotest pressure.

jt

 

[jproj]

Thanks for the details, jt. I know I am making certain assumptions that do not necessarily apply to all cases... I didn't think about it until I read your reply, but it probably matters that our fabricator is using COMPRESS for their code calculations.

It is my understanding that (in general, with COMPRESS at least) the calculations are run to determine the minimum required thickness (for internal / external pressure). Based on the results, head / shell thickness are chosen (next largest commercially available thickness). Next, the actuall shell / head thickesses are used to check the rest of the vessel (supports, nozzles, etc.). The final results (design pressure, MAWP, MAP, etc) are detailed in the "Pressure Summary". Obviously, one can run the calculations using minimum calculated thicknesses, but the pressure summary in this case would just show a MAWP equal to the DP, right?

Regards,
Jproj

 

[jte]

jproj-

Yes, I believe that in the scenario that you describe - a once thru evaluation - that the MAWP would be very close to the DP since the nozzle reinforcement calc's will most likely take full credit for the excess thickness in the head/shell. Thus a minor component such as a repad will wind up governing the MAWP since its design is optimized and not the shell or head which it is reinforcing. This kind of approach frustrates me as an engineer who then has to deal with the inevitable rerate and simply since I've been brought up as an engineer to optimize resources.

A once thru design optimizes the fabricator's resources by saving them perhaps 15 minutes (I'm being generous...) of engineering time.

jt

 

[TomBarsh]

Using COMPRESS software (and likely other software as well), it is actually a pretty simple, one-step process. Or perhaps several of your steps are combined...think of tying your shoes, exlaining this to someone in detail may give them the impression that it is more complicated than it really is.

As the designer enters the information into COMPRESS, the software determines the minimum thickness required by Code rules. The designer can enter this thickness or some larger value as the "nominal thickness" of the component. In COMPRESS, if a thickness needs to be increased to meet changing design conditions the software increases it automatically.

Depending on the calculation/reporting options selected, COMPRESS reports the minimum thickness required for design pressure (plus static head, if any), and optionally reports the component's MAWP and MAP. MAWP is defined in Section VIII Division 1, Appendix 3. MAP (maximum allowable pressure) is not defined in this Code but is usually taken (as by COMPRESS) to be the maximum pressure that the component may withstand in the new (uncorroded) condition at ambient temperature, with no operating liquid.

These principles apply to the major "shell" components of the vessel; eg: formed heads, cylinders and transitions. Finding the MAWP for such components is usually very simple.

For example, for an ellipsoidal head there is only a single formula relating the thickness of the head to internal pressure. Solve for required thickness as a function of pressure, or maximum pressure as a function of pressure using only one formula.

A cylindrical shell has two different Code requirements for thickness: one based on circumferential stress and one based on longitudinal stress. In most cases the circumferential stress will govern. But longitudinal stress may govern for cases of tall towers under high wind. Finding the MAWP of the cylinder requires that both formulas be investigated, the lower value will govern.

Things get more complicated with a cone. Again there are two different formulas required to consider required thickness based on circumferential and longitudinal stress, and these must be checked at both the small and large diameters of the cone (of course, you can make a priori conclusions that one end or the other will govern). So now we have 4 formulas required to check the MAWP of the cone. But wait, we're not done yet! For cones we also have to check the cone-cylinder juncture requirements of Appendix 1-5. Strength of either of the junctions may also govern the cone MAWP. It's at this point that obtaining the MAWP is not a trivial task involving only 1 formula or several independent formulas.

Determining MAWP of nozzles gets even more complicated. There is no single formula that you can express to get the nozzle MAWP. For a given nozzle construction the MAWP may be limited by available reinforcing area, minimum nozzle neck thickness, minimum weld sizes, weld path strength requirements, rating of attached flange, stresses in shell or nozzle neck if there are external loads (WRC-107) on the nozzle, stresses per Appendix 1-7(b) if a large opening, MDMT rating, and likely many, many more criteria. These many unrelated criteria must all be met at the MAWP. Because they are generally unrelated it is impossible to define an algebraic system (such as an 8 x 8 matrix, etc) of equations to solve for the nozzle MAWP. Instead, each Code requirement must be checked one by one for each possible "candidate" MAWP. Using computer software makes the job of finding the nozzle MAWP tractable; by hand it would be very laborious and not practical. This explains why in the olden days vessel owners’ would indicate in their specifications things like "full nozzle area replacement", or "nozzles shall not limit MAWP", etc, so that a nozzle or some inconsequential part on a vessel would limit the vessel MAWP.


Tom Barsh
Codeware Technical Support

 

[jproj]

Tom,

Thanks for the insight, your last paragraph on the nozzle MAWP was especially enlightening (particularly the part about prohibiting nozzles limiting the MAWP).

I understand that these are very detailed and lengthy calculations. If, however, the MAWP was determined, shouldn't all component parts "pass" if the MAWP of the limiting component was entered into the program as the design pressure (keeping all materials & thicknesses the same)?

Regards,
Jproj

 

[TomBarsh]

jproj,

Yes, if the limiting MAWP (the "chamber MAWP") is re-entered into COMPRESS there should be no change to the other components, all components should "pass" at this pressure (this is the same as what I would expect if I were doing the calculations by hand). By definition the MAWP of the pressure vessel will be the lowest of the MAWPs calculated for each component of the vessel.

I accidentally omitted in the list above of possible limiting factors for nozzle MAWP that the thickness of the shell to which the nozzle is attached may also limit the nozzle MAWP.

Note that, as jte alluded to, when using COMPRESS it really costs your fabricator no more time to include MAWP calculations for the vessel than not. The reports will be longer and may lead to more questions from a knowledgable vessel owner or their representative. Additional engineering time may be required to permit the designer/fabricator to optimize the construction as per jte's valuable comments.


Tom Barsh
Codeware Technical Support

 

[jproj]

That's pretty much what I expected (basic logic the way I look at it... 2+2=Y is the same as Y-2=2).

What still confuses me is why an AI would insist on new calc's using the MAWP as the DP when the current calc's (with the actual DP) already show the MAWP. Is this just an incompetent AI or just one that is unwilling to stamp anything with less than a 20x overall safety factor?

Regards,
Jproj

 

[TomBarsh]

That's a good question. UG-116 "Required Marking" for the nameplate even shows that the MAWP (not "design pressure") is to be on the nameplate (see UG-116(a)(1)(3)). A footnote to this paragraph states "The maximum allowable working pressure may be assumed to be the same as the design pressure when calculations are not made to determine the maximum allowable working pressure." Sounds like the AI doesn't have much of a leg to stand on. But he's the "judge, jury, and executioner" in this case. The work required to produce the COMPRESS calculations for design pressure = chamber MAWP would be minimal; the hardest part is printing them, getting them distributed, approved, etc.

 

[chaulklate]

jproj,

I can't disagree with you, it doesn't necessarily make sense, and I don't see how it differs from 'solving for thickness' vs 'solving for pressure', just stating my experience which seems to be the same experience as you. The long and short of it is, the AI's want to see everything calculated on a set of 'control' parameters. Which for me basically means determining MAWP on my weakest componenet (or as set out by jte) and then rerunning all calc's based on the 'new' pressure as indicated in MAWP report. And as Tom indicated, the AI gets the final say. But at least they seem to be consistent with their wants.

 

[djack77494]

I fail to understand your comments about "rerunning the calcs". As powerful as software like COMPRESS is, are you saying that significant extra efforts are required to go from the vessel design mode to (essentially) a vessel rating mode. All the data is already in the program, so I'd think very little effort would be needed. However, I admit my background is not in vessel mechanical design, and I've never personally run COMPRESS. I more looking to be educated on this subject. Thanks.