How to properly analyze a flange at a boiler in stress analysis?

[NickDePinho]

Does anyone have any insight on how to properly model a flange connection on a boiler?

I was given nozzle loads and moments for a boiler that I am connecting to.
I am running a stress analysis of my HW pipe. I typically model anchors as non rigid. allowing for roughly 1/16" of total movement at the connection. This is a rule of thumb that I have learned from an engineer who has taught me.

I have lowered my translational stiffness some in the X,Y, and Z.

But I am unsure of how to treat rotational stiffness of a 10" 150# Flange. Any thoughts on this wold really help.

I found something saying to use 1e5 in*lb/deg another saying allow for .3 deg of rotation.

Curious if there are any codes or insight on how to figure out what is a safe but realistic way to model this flange

 

[KevinNZ]

What programme are you using for the stress analysis? Most have model elements for vessel nozzles. Stiffness is calculated from the vessel diameter and thickness.

1/16” movement rule of thumb could be way off and non-conservative.

Do include the thermal movement of the nozzle away for the vessel anchor point.

You also need to check the loads on the flange for gasket sealing (UG-44). Which not the same thing as allowable loads on the nozzle to vessel connection.

 

[RVAmeche]

I use Caesar II so my terms will be from that program, they all vary slightly.

In my experience using Caesar II, I generally stick to matching the stated thermal growth of the nozzle (ie, at 400F it will grow up 0.25" and laterally 0.125" or something) by inputting it as a displacement rather than anchor. With the nozzle moving how much the vendor says it will, I then work on reducing the loads at that displacement to get below their allowables.

If you know the physical geometry of the nozzle connection, you can do try to do a Nozzle Stiffness calc per WRC 297 or something. I generally try to avoid this unless its absolutely needed tho, since as a consultant I don't want to take on responsibility for saying their nozzle should have been able to move based on xyz calculated stiffness.

Another route/rule of thumb - if the stress at the nozzle is less than 5,000-6,000 psi (Peng's Pipe Stress book mentions this if you need a source), the nozzle is probably okay. Similarly, my ASME B31.3 course instructor said if the pipe stress at the nozzle is < 10% allowable stress, the nozzle is probably okay.

In my Casear II static analysis course we went through a couple examples of nozzle loads and looked at the types of calculation changes you can make and their impacts on stresses/loads -

1. If you can adjust the SIF this can be a quick & easy update
2. WRC 297 generally gives you a bit more flexibility and slightly lower loads, but not a ton in a lot of cases
3. FEA of the nozzle can give you much more accurate data and significantly reduce loads
4. Piping mods via adding expansion loops or accounting for support stiffnesses can significantly reduce loads

 

[NickDePinho]

What programme are you using for the stress analysis? Most have model elements for vessel nozzles. Stiffness is calculated from the vessel diameter and thickness.

1/16” movement rule of thumb could be way off and non-conservative.

Do include the thermal movement of the nozzle away for the vessel anchor point.

You also need to check the loads on the flange for gasket sealing (UG-44). Which not the same thing as allowable loads on the nozzle to vessel connection.

Thanks a lot for the heads up. I used caepipe. I modeled the nozzles as anchors. I used Nuclear branch flexibility reporte by Rodabuagh to find the rotational stiffnesses. I also incorporated the displacement at the nozzle.

 

[NickDePinho]

I use Caesar II so my terms will be from that program, they all vary slightly.

In my experience using Caesar II, I generally stick to matching the stated thermal growth of the nozzle (ie, at 400F it will grow up 0.25" and laterally 0.125" or something) by inputting it as a displacement rather than anchor. With the nozzle moving how much the vendor says it will, I then work on reducing the loads at that displacement to get below their allowables.

If you know the physical geometry of the nozzle connection, you can do try to do a Nozzle Stiffness calc per WRC 297 or something. I generally try to avoid this unless its absolutely needed tho, since as a consultant I don't want to take on responsibility for saying their nozzle should have been able to move based on xyz calculated stiffness.

Another route/rule of thumb - if the stress at the nozzle is less than 5,000-6,000 psi (Peng's Pipe Stress book mentions this if you need a source), the nozzle is probably okay. Similarly, my ASME B31.3 course instructor said if the pipe stress at the nozzle is < 10% allowable stress, the nozzle is probably okay.

In my Casear II static analysis course we went through a couple examples of nozzle loads and looked at the types of calculation changes you can make and their impacts on stresses/loads -

1. If you can adjust the SIF this can be a quick & easy update
2. WRC 297 generally gives you a bit more flexibility and slightly lower loads, but not a ton in a lot of cases
3. FEA of the nozzle can give you much more accurate data and significantly reduce loads
4. Piping mods via adding expansion loops or accounting for support stiffnesses can significantly reduce loads

Thanks for the input. I calculated some nozzle displacements to incorporate in my analysis. I also found the rotational stiffnesses using Rodabaugh and Moore's work on nuclear branch piping flexibility. In caepipe I modeled the nozzle as a non rigid anchor and put in the rotational stiffnessses I calculated. Between the numbers I found as well as the conservative refernece temperature I am using I feel good about my results.

Any input would be helpful

 

[Snickster]

Here are some old references I have which may be helpful.