We are analysing piping as per B31.3 for one of our projects and have been requested to add SIFs at all branch connections, irrespective of branch diameter. As a result we are ending up defining SIFs at all drain ,vent ,temperature,pressure connections.
Caesar calculates and applies same value of SIF on branch and header irrespective of branch size as most of SIF calculations are based on Header diameter and not on branch diameter.

As a case ,we have an 18” X 1.5” weldolet connection . Caesar uses an SIF of 3.31 for both header as well as branch. Due to this SIF value ,Caesar is reporting failure in the 18” header in some load cases.

Practically we are aware that a 1.5 “ connection will not matter at all for an 18” Header.
Is there any clause in B31.3 that speaks of any limitation of SIF values for very low d/D ratios.?
Any ideas or suggestions?

hi dear .you can use fe pipe software to calculate sif

I am also facing the same problem as that of manoj.

Is this right to model SIF for Temp. gauge,pressure gauge and drain as theses connction are free

can anybody throw light on thsi subject.

Why don't one of you give me a real world example, such as header diameter, wall thickness, material...branch diameter, wall thickness, material, etc. and I will run it in an FEA program and give you my results for calculated SIFs. Then you can decide if it is even worth worrying about.

In my case I have 450 NB header of SS 304 (sch 10S) and there is a drain line 50NB (sch 5S)(with end free with blind flange).

Most of us believe that if line is free then there is no need to model SIF, but one of our clent insisted to model SIF event attemp Indicator and Pressure indicator, due to this we required to add more flexibility to line to pass line in expansion stresses

Here are your results. Your header SIF outplane is less than 1 so use 1. (Read the FEA line for the FEA SIF results of course)

In my opinion, it is irrelevant to include small bore lines including drains, vents, TI, PI connections in the comprehensive analysis of piping system.

Thanks threeouts

As per B31.3 Appendix-D, Note-(1)- If you have more applicable direct data (like derived thru FEA), can be used in calculation rather using B31.3 SIF where branch diameter effect is not properly addressed.

Nozzle Pro-SIF can be used in CAESAR-II calculation.

I have one question to the extensive users of Nozzle Pro who uses Nozzle Pro SIF in CAESAR-II calculation. Actually per my understanding one should use seperate SIF for Branch and Header derived from Nozzle Pro. Then accordingly define in CAESAR-II at all three elements (two header element and one branch). Please correct me if i am wrong.

Pradeep, I believe you are correct. This is what I have been doing. Yeah, NozzlePro gives you the option in the options menu to calculate only the header SIF. It says that when you use the default branch sif that it calulates for both branch and header, that this could be way too conservative for the header, especially with small d/D ratios. Not sure why it just doesn't calulate both as default now that I think about it.

Anyways, in Caesar I manually apply SIFS at the base of the branch element, and define two SIFs for the header, one for node going into intersection and one for the node going out of intersection. So yeah you have three SIFS all defined at one point essentially, right at the branch to header junction.

Thanks threeouts!
Also I understand that when one using code (e.g. B31.3)that based on branch effective section modulus rather exact section modulus, at one of the element out of three, joint type (welding / reinforced/ unreinforced etc.) must be defined along with user defined SIF in three elements.

Request to Dave and Richard to confirm above understanding whlile using CAESAR-II.

I never define branch type if I have already defined SIFs manually at that intersection. For 31.3, flexibility factor, k is 1 for all branch types for D/T<=100.

However, for larger bore pipe, maybe 24" and larger, and certainly for pipe with D/T>100, I will model the branch in FEA and get more accurate flexibilities.

My opinion is to define connection type at one of the three element while entering user defined SIF at all three elements. Specifically it is essential when code is using effective section modulus for branch stress stress calculation. For example B31.3 code.
I once again request to Dave and Richard to clarify our doubts on user defined SIF in CAESAR-II.

DAVE & RICHARD please clarify our doubt in CAESAR II modeling procedure.
Thanks.

I think you have it right. If you call out a tee in B31.3, CAESAR II will use Z(effective) for the branch and you can override the SIFs by explicitly defining them for each leg into the tee.
Here's your test: define SIFs for each leg and run the job twice - once with a tee defined and once without. Does the branch stress change? Header stress should remain the same in any case. (Be sure you have a tee where Z(effective) will be different from Z(branch)).

Thanks Dave for your response.

You are right. Yes stress at branch node is different in both run - with a tee defined and second without tee defined. Header stress remain same in both the case.

May I request you to clarify one more doubt that why effective section modulus is used by code (B31.3) over exact section modulus for branch stress calculation? What is real theory / expanasion behind this?

Thanks!

It's one way to account for reduced strength, locally, associated with reduced branch size.

Thanks Dave for clarification.