Nozzle Loads : Allowable for Thermal,Seismic,Weight+Pressure Load 1

[knandwana]

Hi.

The nozzle loads as per the customer spec. includes loads due to (Weight+Pressure),Thermal and Seismic.

Now, Customer wants us to add all these loads and qualify the nozzles.

I am using WRC-297 for stress calculation. Now, simply adding these loads and checking against the Local Membrane Stress Intensity (1.5S) and Primary + Secondary Stress Intensity (3*S) results in overconservative design.

My question is how does the allowables increase in presence of : a) Seismic Load , b) Thermal Load and c) Combination of all three loads ?

Thanks,

Kapil Nandwana
Structural Engg., EWI

 

[TGS4]

See ASME Section VIII, Division 2, Part 5, Table 5.3. After you have done that - some back with any additional questions.

 

[knandwana]

TGS4,

After looking at table 5.3, as suggested by you, I have few questions.

The seven load combination does not consider the nozzle loads due to thermal expansion/contraction. Does it mean per ASME, I need not check for thermal loads at all and only consider the load case 4) D+0.7*E ?

However, the customer specs defines the allowable nozzle load combinations as follows :

Normal Load Case : Weight + Thermal (algebraic summation)
Occasional Load Case : Normal Load Case + Seismic (absolute summation)

How should my allowables change in this case ?

Thanks,
Kapil, Structural Engineer, EWI

 

[TGS4]

What your customer describes as "thermal" piping loads, can better be described as "loads arising from the restraint of piping free thermal expansion" or "loads attributable to restrained free end displacement of attached piping". You will see in Table 5.6 that this term comes up three times in describing "Nozzles".

It is my opinion (and my opinion only) that stresses coming from these "thermal" piping loads should be classified as primary. Therefore, in Table 5.3, these loads would be categorized as "D".

Please note that ALL of the load combinations must be checked, regardless of whether your client wants less or not. Client specs can supplement the requirements of Division 2, but cannot subtract from them.

And in case you are using Division 2 by way of U-2(g) from Division 1, standard industry practice would say the exact same thing - there's no way around it, you have to check ALL of the load combinations.

Your only "out" in the case of your analysis showing higher than permitted stresses, would be to perform an elastic-plastic analysis that includes the piping.

Finally, don't forget the mandatory calculations that must be performed in 5.5.6.

 

[EngAddict]

@TGS4

Why would you include the piping in the elastic-plastic analysis? It could be almost impossible to model accurately, is it worth the over-complication of the model?

Are you also going to consider pressure thrust a variable of the elastic piping system?

 

[TGS4]

How else would you calculate the behavior of the piping if you don't model it? For a load of X, is that from 0.1 or 3 inches of restrained free thermal deflection? That difference would make a huge change in the result of the elastic-plastic analysis of the vessel.

 

[Mandeep1]

Dear TDS4 how do you accurately know the stiffness of the restraints, bends, reducers, T's etc. How can you accurately know the local yielding that might occur at these locations. How can you accurately account for the manufacturing tolerances. etc.

I agree with EngAddict that this will make the analysis too complicated and it may not be accurate given the variation of field conditions mentioned above.

On the other point, if you classify the "thermal" loads as primary you are completely ignoring the self-limiting nature of these loads and not considering any local yielding in the system that might "limit" these loads.

Imho you are being little too conservative if you do so. But, at least you are on the safer side.



best regards,
Mandeep Singh

 

[TGS4]

Mandeep - I would tend to agree with you. Classify the stresses as primary, and you're done. If you want relief, then you need to include EVERYTHING.

As far as how do you accurately do that? Well, the pipe stress guys manage to do that task pretty well...

 

[EngAddict]

We also do pipe stress analysis but it is not included in the nozzle FEA. AutoPIPE will output forces and moments which we would then apply to our FEA model.

 

[TGS4]

EngAddict - that's all find and good. How do you classify the stresses resulting from the loads that are generated from restrained free thermal expansion (the "thermal" loads for your pipe stress program)?

 

[EngAddict]

In the nozzle analysis they are external mechanical loads (primary), not thermal loads in the nozzle. This is typically what causes all external nozzle loads. Think of the piping as a black box type system in relation to the nozzle. The piping stress analysis also takes the nozzle flexibility into consideration.

More to the point that I think you are getting at, is it would most likely cause local primary stress which will transfer some load to other areas and may cause local yielding. It has some similarities to the self limiting affect of secondary stress but I would not classify it as secondary.

 

[TGS4]

For a good coverage of this topic, please refer to the following two papers:

http://www.asmedl.org/getabs/servle...6725000187000001&idtype=cvips&gifs=Yes&ref=no

and

http://www.asmedl.org/getabs/servle...188X000451000001&idtype=cvips&gifs=Yes&ref=no

 

[BPVFEA]

In my opnion,

1) It is irrelevant if piping load coming on nozzle is Mechanincal or thermal. Finally it is the load acting on nozzle.

2) Under any of these loads stresses acting in Nozzle neck (away from junction) shall be classified as primary (Pm < S , Pm+Pb < 1.5S).

3) At the junction the stresses are secondary (Pl < 1.5S & Pl+Pb+Q < 3S)

4) If you are applying different temperatures on shell & Nozzles AND considering Expansion coefficient in the analysys THEN stresses in all locations (neck, junction, shell) shall be considered as secondary (Pl+Pb+Q < 3S)

5) If seismic load is considered then allowable shall be increased by 1.2 times as per old code. OR load shall be reduced by suitable factor keeping the allowable same. (Table 5.3)

6) WRC does not recognise if load is Mechanical or Thermal. It compares (Pl < 1.5S & Pl+Pb+Q < 3S) based on load input. I feel that, in the event like point 4 above, the, WRC can find thermal stresses.

 

[BPVFEA]

Sorry, the last line in my reply shall be "I feel that, in the event like point 4 above, the, WRC can NOT find thermal stresses"

 

[TGS4]

BPVFEA -

Agree on 1).

For 2), define "away from the junction".

For 3), Pl is a primary stress limit. If the stresses are primary, then the Pl limit applies. If they are secondary, then the P+Q limit applies. It's not both. I would define the local membrane stresses as primary (Pl), and the membrane-plus-bending as secondary (which it always is).

For 4), only those stresses that can be attributed to the differential thermal expansion in the shell/nozzle could be excluded from the above checks. And the stresses due to only differential thermal expansion would be F (peak).

For 6), actually, WRC107 doesn't perform ANY comparison. Your program might, and I bet that it's doing it wrong - most are. WRC107 calculates a membrane and a membrane-plus-bending stress at the 4 cardinal locations around the attachment (remember, WRC107 doesn't actually include nozzles with holes - we've used it for that, but if you look at the riginal document, it doesn't).

 

[BPVFEA]

Thank you TGS for the corrections. Some points I agree with you and some points I have different opinion. I am rewriting my explanation.

In my opinion,

1) It is irrelevant if piping load coming on nozzle is Mechanical or thermal. Finally it is the load acting on nozzle.

2) Under any of these loads stresses acting in Nozzle neck away from junction [ 2.5 SQRT(rt) ] shall be classified as primary (Pm < S , Pm+Pb < 1.5S).
(I am referring to fig 7 of guidelines given by

http://pm-engr.com/aboutpmi/pmipapers/pvp1999-vol388_2.pdf

. Thanks to Porter & Martens.)

3) At the junction the stresses are
Local Membrane (Primary + Secondary*) Pl < 1.5S
Local Membrane + Bending (Primary + Secondary) Pl+Pb+Q < 3S

* From page10 of WRC429 “The Code recognises that primary local membrane stresses have some characteristics of Secondary Stresses. Conservatism requires that such a stress be classified as primary membrane stress even though it has some characteristics of a secondary stress”. (Also refer 5.12.10 of Div 2)

4) If you are applying different temperatures on shell & Nozzles AND considering Expansion coefficient in the analysis THEN stresses in all locations (neck, junction, shell) shall be considered as secondary (Pl+Pb+Q < 3S)
To get the stress component induced due to diff. thermal expansion
Step 1 - Run analysis considering piping loads & w/o considering ALFA. But apply the Temperature distribution (to catch the actual Young’s modulus). Let’s say you get linearised stress (P+Q)1
Step 2 - Re-run analysis considering ALFA. Let’s say you get linearised stress (P+Q)2
Step 3 – The difference (P+Q)2-(P+Q)1 would give the Stress component of (P+Q) that is induced due to diff. thermal expansion.

If location under consideration is local structural discontinuity (notch, fillet etc.) THEN the peak stress F would in come into picture. F can be obtained by subtracting linearised stress (P+Q) from the Max stress appearing at the corner.
Again here peak stress would have two components.
F1- Due to loads but w/o Diff Expansion
F2- Only due to Diff Expansion

5) If seismic load is considered then allowable shall be increased by 1.2 times as per old code. OR load shall be reduced by suitable factor keeping the allowable same. (Table 5.3)

6) WRC107 & 297 does not recognise if load is Mechanical or Thermal. WRC does not find the stresses induced due to differential thermal expansion.

 

[TGS4]

BPVFEA, I think that we are essentially in agreement. Good to see.

Now that we can see the difficulties with stress classification and linearization, can we all just get on-board the elastic-plastic FEA ship. Factor the loads, run the analysis, and you're done.

 

[BPVFEA]

That's great!
Elastic Plastic FEA has been paid little attenstion. I will create a new post for Elastic Plastic FEA. I will post my understanding as well as doubts. Let's learn & improve throgh discussion.
Regards
BPVFEA