I understand that it is not advisable to connect two equipments with straight run of pipe. But to analyse the situation, I modeled a system with

6 in. pipe, carbon steel material, 185 deg F, 2 m length,

connected straight between pump & buffer tank.
As expected, the nozzle loads in thermal case were very high.
Now if I increase the length of pipe to 10m, the nozzle loads in thermal case of analysis output are the same!

So my questions are:

1) Does the thermal load on nozzle depend on length of straight run of pipe between equipments?
2) Will the B 31.1 Eqn 13 (Thermal) stresses at all nodes be zero for above model ? If yes, why? (that is what I am getting in the output)

Regards,
ASA

1. You must have modelled both ends of your pipe as rigid, giving you a simple unit strain problem. A given length of pipe will expand, but relative to its Young's modulus value it will absorb axial strain producing a certain stress. Regardless of the length, the axial force and axial stress is equal.

If you put in a flexibility for the buffer tank (the real world), results would be quite different.

2. The equation calculates bending stress, you have axial stress only.

3. Try to think for yourself.

Dear MoverZ,

1) Considering the cited problem as 'unit strain' case, does this mean that the analysis softwares would neglect buckling effect in the pipe if the ends are modeled as rigid?

Also, considering the ends to be perfectly rigid, it implies that the amount of axial force that the nozzles are facing in thermal case is equal to the force that is required to compress the pipe by a length equal to the expansion it exhibits. Now, expansion is a function of original length. So, by increasing the length of pipe between the nozzles, the expansion & hence the nozzle axial force has to increase. Then how come the nozzle axial thermal force remain the same?

2) As rightly pointed by you, B31.1 eqn13 is considering only bending stress. So does that mean that in straight pipe runs, the axial forces and hence stresses are not at all considered in the thermal cases? I understand that it is rare for a pipe to fail in pure axial compression, but given the fact that the analysis becomes a unit strain problem, is this consideration (assumption) of only 'bending stress' justified?

Regards,
ASA

Dear ASA,

When we interesting about the stress value generated by Caesar, by starting from the "global element forces" report is a good one. See all three forces (see which is axial force) and all three moments (see which is inplane bending, outplane bending, and torsional moment) in every load case(s) you are interesting to. Consider SIF for each load case(s). Substitute that value(s) to related stresses formula. If we don't miss a thing, the value should match the stress output.

When we interesting with EXP stress, we should start from "global element forces" in case EXP too. For EXP load case(s) you mentioned, did you see non zero value for the bending and torsional moments?

After succeeding, you may ask from where we got three forces and three moments. Well, you have to read book relate to beam element for finite element method. Good luck for the trial .

Dear ASA,

Please see the attachment. The stress (and Force)are not dependant on length of pipe. Hope this clarifies.

Regards
Siv

Dear ASA,

When the piping is connected from one equipment to another by straight run, check if there is any differential thermal expansion between the 2 equipments...
If both of them are at same temperature...then there is only thermal stress from anchor to anchor...

Thermal Expansion : dL = L * alpha * dT

Strain = dL / L = alpha * dt

Therefore,
Thermal Stress = E * strain = E * alpha * dt

Thermal Force = Stress * metal area (Am)

Hence you can see that Thermal stress as axial stress and Thermal force axial force in the caesar output and as the eqaution shows it is independent of length.

Any lateral displacement due to differential movement will result in Bending moment...hence bending stress..
Hope this solves ur query...

Dear Siv,

Did you mean that for ASME B31.3 code stress for EXP load cases = alpha*delta T*E? please clarify.

Asa,

I hope with the help of the other stuff above, you understand the concept of unit strain now.

The best solution is NEVER pipe up two pieces of equipment with a straight shot. Simply installing a gasket would be a problem, let alone resulting thermal stresses.

Read the B31.3 code to see what is included in thermal stresses. Also check ou the effects of changing the Caesar II configuration setting 'Add F/A in Stresses'.

Caesar II does not consider buckling. The analysis is purely elastic. Read the User Guide.

The B31.3 equation for expansion stress range (17) does not include the axial term but see comment in 319.2.3(c). As M/Z says, CAESAR II allows you to include teh axial term. In fact we offer a configuration switch for this paragraph in two different flavors!

Dear Manik,
As you know B31 codes do not take axial term for code stress. The equations were to make it clear to ASA that the axial compressive stress and force are independant of pipe length ( when two ends of a straight pipe are anchored).

Siv

Dear Siv,

As you know B31 codes do not take axial term for code stress ( I disagree w/ this opinion). Especially for SUS code stress for ASME B 31.3.

SUS code stress = Slp + Fax/A +Sb.

There are three options for 'Add F/A in Stresses' in configuration setup [default, yes, no]. Please read related help facility carefully especially for "default" choice.

Dear all,
The above information convinces me. Thanks for your knowledge sharing.

Regards,
ASA

Dear Manik,

1.Since the discussion was around expansion stresses I missed mentioning expansion stresses ,which I should have.

2.You are correct about the application of F/A stress in sustained and occassional code stress check in Caesar.

However this is not a explicit code requirement. caesar adds the F/A term in sustained and occassional loads for B31.3 because the equations are not explicit in code. In B31.1, where the equation is clearly defined F/A is not added by default.Finally it is the engineer's decision to decide if it is critical for the system and apply accordingly.The option to add F/A is always there.

Anyway, thanks to your post I read B31.3 after a long time (using 31.1 for quite some time) to check if the latest edition has some explicit requirements for F/A in sustained loads.


Regards

Dear Siv,

Thank you very much for the clarification and additional explanation.