Hi,
I have a piping system with a nozzle on a bend. I have to calculate SIF for this type of intersection because there is not any formula in B31. I calculate first SIF for the bend, next SIF for the intersection, and I put the higher value in the node. Is corect this calcultion?

Thanks
Dragos

Hello,

NO, this calculation will not provide the correct flexibility factor nor the corrects SIF. If you really must analyze this abomination, you must use finite element methods.

In the component that you describe, much of the natural flexibility of the elbow is lost by changing its geometry. Consequently, the elbow SIF is no longer correct. But arbitrarily using a branch connection SIF is equally incorrect.

Remember SIF’s only apply to beam theory models and the stress analysis that you need to do cannot be approached by beam theory (local membrane bending stresses will be the determining stresses, not the gross section beam bending stresses).

I would not willingly allow such a component in a piping system that I would be responsible for. However, having said that, I think the best approach to analysis here is to model the component (within a comprehensive CAESAR II model) as a series of rigid links with nodes at the tangent points (weld lines) of the elbow and at its mid-bend point (presumably the intersection with the centerline of the branch). There should probably be some mention of local and system stiffness matrices here but let me skip that part. Also place nodes some distance (maybe 5 to 8 pipe diameters) away from the tangent points (use regular "pipe" elements) so that forces and moments will be calculated at those points. Then, make your FEA model. Add enough straight pipe distance to the FEA model (the 5 to 8 diameters) so that the "end effect" will not transfer (will be "damped-out")to the curved section of the model. Apply the forces and moments calculated by CAESAR II at the remote ends of the FEA model (remember the summation of forces and moments must be zero).

Remember that CAESAR II is a beam theory analytical tool. CAESAR II applies the rules provided by the piping Codes. The piping Codes do not attempt to provide rules for every possible (albeit ill-advised) configuration. If bending stresses that are calculated by beam theory are not the limiting stresses (or the configuration does not lend itself to analysis by beam theory), the engineer is responsible to find another approach to the analysis.

Have you decided how you will qualify this component for pressure design? Perhaps you should configure your FEA model so that you can also use it to determine internal pressure stresses. Remember CAESAR II has no way of knowing that this "special" geometry exists - Code equations for straight pipe minimum wall thickness are not applicable here.

Or, replace the elbow with a standard one that is not modified and find another place in the piping system for the branch connection.

Good luck with your projects.

Regards, John.

In the case of a pipe component not the constructions requirement the respective pipe code fill, then my opinion is the following :
1. Calculate the local bending stiffnesses(In/Outplane)with Finite Element
2. Calculate the force and moments with a pipe calculation programm under consideration of a local bending stiffnesses. Be mindful of the pipe flexibilitätsfactors, so you have the correct local bending stiffnesses
3. Calculate with the force and moments and pressure the stresses with FE
4. Estimation the stress categorie Pm,Pl,Pb,Q,F . The stress limits from the pipe equation are not acceptable. Use the stress limits for the stress catogerie from ASME Code.

Thank you Mr. Ohliger,

Mr Ohliger makes several excellent points. The appropriate flexibility factors can be determined FIRST by using the FEA model, and then those flexibilities can be used in the comprehensive CAESAR II model.

Mr. Ohliger also points out a very important consideration - when using the FEA approach, the designer must use the correct allowable stresses for comparison to the stresses calculated in the FEA model. The B31 Pressure Piping Codes give the designer the latitude to employ more rigorous methodologies for design by analysis - but within any set of Code rules it is important to use the correct allowable stresses. In this case (for this component), it would be appropriate to employ analysis methodologies described in the ASME B&PV Code, Section VIII, Division 2 - i.e., FEA. It is very important, of course, that the designer also apply the "Div. 2" allowable stresses to the FEA analysis.

For completeness, the designer should look at the "Div. 2" rules for fabrication and nondestructive examination because the allowable stresses are predicated upon the increase in rigor in those areas too.

Regards, John.

[ September 05, 2001: Message edited by: John Breen ]

Thanks for all explications!
I modeled the junction with FE-Pipe software (PRG) and I obtained necessary rezults: flexibility and SIF. I modeled in FE-Pipe also a part of my piping system and I put all limits conditions (I read them from the CAESAR output). I obtained stresses according ASME Sec. VIII, but for B31 in expansion case too. I introduced the SIF and the flexibility of the bend in the CAESAR model and I compared the rezults. They was very close. Because the nozzle has a free end (is a thermo-rezistance) there was not high values for ASME stresses

Thanks for all your advices!

Best regards,
Dragos Aron