Nozzle stress intensity factors 1

[tired]

When calculating stresses due to nozzle loads Compress calculates a stress intensity factor in accordance with PVP vol 399. I was under the impression that these need not be calculated unless you are looking for peak stresses for fatigue duties and 1 should be used unless fatigue is an issue. Is this not the case?

In some cases the calculated factor is less than 1 meaning the calculated stresses are reduced. Is it not the case that a minimum of 1 should be used?

 

[TomBarsh]

The stress intensity factor applied in the COMPRESS nozzle WRC-107 report is not that defined in the 107 bulletin for fatigue analysis (COMPRESS does not consider fatigue in nozzle design, this is outside the scope of the software).

The factor is actually the "Pressure stress intensity factor", I. This factor accounts for the increased membrane tensile stress existing in the shell near the nozzle opening. Factor I multiplies the general primary membrane stress to determine the local membrane stress at the nozzle opening. Of course, the general primary membrane stress is that found by elementary methods such as stress = P*R/t for circumferential stress, etc.

It is shown by the theory of elasticity that whenever there is an opening in a plate that is subjected to a membrane stress there will be increased membrane stress near the opening (see any text on strength of materials, etc). The same principle applies for openings in the pressure vessel shell at the nozzle. But the determination of the stress distribution is not as simple as it is for the case of the flat plate with a circular opening (which is not so easy a problem in itself).

The method used by COMPRESS to determine I is that described in paper PVP-399 (pages 77-82) published by ASME. An extract of these pages is available on the Codeware website under Technical Papers. The method was recently adopted in Code Case 2398.

The method described in the PVP paper has been extended to other geometries and continues to be refined with validation by FEA and mechanical testing.

Designers who wish to enter their own values for the SIF can do so by checking the "User defined SIF" option on the WRC-107 dialog, then entering the desired value. As I remember, PD 5500 (the old BS 5500) contains some tabulated listings of SIFs to account for the pressure stress. I doubt these are as accurate as the PVP 399 method but some people prefer these.

The pressure SIF can be less than 1.0 in the case where there is sufficient integral reinforcement at the nozzle such that the calculated membrane stress in the shell at the nozzle opening is less than the general primary membrane tensile stress.

Tom Barsh
Codeware Technical Support

 

[tired]

Thanks Tom,

Is the calculated factor I included in WRC 107 or 297 or the Code. What I really need to know is can I input 1 as a user defined SIF and comply with Code requirements.

Thanks in advance.

 

[TomBarsh]

Welding Research Council bulletin 107 does not calculate stresses in the shell due to internal pressure. Nor does ASME Section VIII Division 1 provide any method of determining this stress.

Division 1 is silent on how to treat this matter. But this doesn't mean the designer can ignore it. The Code requires consideration of all design issues that may affect the vessel but does not provide instruction on how to analyze all situations. Designers must often refer to other bulletins, references, text books, etc.

If you determine that the SIF due to pressure stress at the nozzle opening is "1", then you can enter that into COMPRESS as a user defined value.

Tom Barsh
Codeware Technical Support

 

[tired]

Tom
Thanks for your help.

Further to my original question and on a slightly different topic. When compress calculates the stresses in the nozzle wall itself it doesn't appear to calculate the combined stress intensities as detailed in WRC297. How does compress address the combined stresses due to ML, MC and P in the nozzle wall?

Thanks

 

[TomBarsh]

COMPRESS calculates the "general" stress existing in the nozzle neck due to combination of pressure and external loads. COMPRESS does not use WRC-297 for this purpose although the WRC-297 analysis is available via the FEA-Nozzles finite element analysis option.

But even for WRC-297 the stress intensities must be transformed into resultant stresses for comparison to the allowable stress.

COMPRESS treats the nozzle neck as a hollow circular "beam-column" that is also subject to internal pressure. The combined longitudinal stress is found as resulting from axial load ("P&quot, combined bending moment (resultant of Mx and My), and internal pressure. Thus the stress in the nozzle wall is reported due to internal pressure + external loads as: (see COMPRESS report, below the WRC-107 table)

sn (Pm) = P*Ri/(2*tn) - Pr/(p*(Ro2 - Ri2)) + M*Ro/I

which is the stress due to pressure:

Pressure * R / 2t

combined with stress due to bending:

M / (section modulus)

combined with stress due to axial tension/compression:

(Axial load P) / area


The resulting value is compared to the corresponding allowable stress.

Tom Barsh
Codeware Technical Support