Stress analyse and ASME Codes 1

[frd03]

Hello everyone !

I have studied the codes Asme B31.1 e B31.3 and I still did not find some answers for these questions which I would like to receive your comments and opinions:

a) The self-springing effect from A.R.C. Markl paper, which introduced the displacement allowable stress SA, it had your simulation and tests with specimes under 950 oF so inside the Creep range and consequently with stress relaxation on time.
Is dificult to understand and prove this effect with intermediate temperatures like 570 oF (bellow the creep range)and with no cold spring.
I think that the code should divide the allowabe stress SA on two groups, up to 700oF and above, with different procedures.

b) The code B31.3 (F301.5) presents a warning about the dynamics effect on pipings but it did not make a orientation about some procedure and I think that pratically all pipings are under vibration ( acoustic or mechanical ) from fluids or rotary machines. To displacement stress is presented the factor f but in general thermal cycles are very lower frequency if compared with primary stresses.
How ASME got the allowable stresses Sc and Sh (to hoop and longitudinal stresses )? would have been from some fadigue criterium ?

c)The solutions applied to solve the "structural analyse" of a piping still are supported with "beam theory" (and also from Kellogg Inc.) coupled of Asme rules but I think that we have very good technical resources through shell teory and finite elements method which we would allow to get a optimal design reducing material and costs applied on the piping but in another side the schedule series of pipes do not offer a large range for this (maybe by self interest of suppliers ).

Thanks for your help

Regards

Francisco Dominguez
frdominguez@ig.com.br

 

[JohnBreen]

Hello Francisco Dominguez,

You are asking good questions. They have been asked before. They will be asked again.

1) Self springing is an inelastic behavior that can (and often does) take place below the creep temperatures of materials. On the first (and perhaps the next few) thermal cycles, some plastic deformation may take place (generally at elbows) and upon cool-down the piping will be permanently deformed (as a result there will be a residual stress in the piping). The B31 Codes recognize the beneficial effects of self springing but caution that the piping should “shake down” to purely elastic behavior in a few thermal cycles. If the plastic deformation continues for many thermal cycles, the piping system may fail due to ratcheting. This plastic deformation should not be confused with creep.

The thermal expansion (displacement) stress ranges are compared to the Code maximum allowable stress range (SA). The equation that gives us the maximum allowable stress range include the terms Sh and Sc. The values for Sh and Sc are to be found in Appendix A of the Code(s). Please note that the values of Sh and Sc are based upon a percent of the material’s yield strength or upon a percent of the material’s ultimate strength (whichever is the lesser) but only up to the creep temperature of the material. Beyond the creep temperature the values of Sh and Sc are based upon the percent of creep elongation over a certain number of hours. So, since creep can determine the values for Sh and Sc, creep (when it is an issue) will also affect the calculated Code maximum allowable stress range (SA).

2) As you point out, continuous vibrations are typically a high frequency, low amplitude phenomenon while cyclic thermal expansions (displacements) are a relatively low frequency, relatively high amplitude phenomenon. It should be noted that although the cyclic thermal expansion range may result in the largest single displacement, other loadings may contribute additional lesser displacements. These other displacements include such pipe displacing loadings as wave action and vibration. The number of equivalent full temperature cycles that determines the value for the factor “f” should represent ALL these cyclic loadings and the number of equivalent full temperature cycles should be computed using the power equation show in B31.1, 102.3.2(c ) equation (2). The values for Sh and Sc were described above and although they do not address fatigue, it is addressed in the equation for SA. Stresses due to internal pressure are circumferential (hoop) stress (P x Do / 2 t) and longitudinal stress (P x Do / 4 t) and these are primary sustained stresses. The B31 Codes address circumferential (hoop) stress in pressure design (e.g., B31.1, 104.1.2) by providing equations for calculating minimum allowable wall thickness.

3) The equations provided by the B31 Codes for calculating bending stresses are indeed based upon beam theory. Other more rigorous methodologies are not precluded by the B31 Codes. The applications of shell theory or finite element theory are not required but they are not precluded by the Code. The B31 Codes revolve around a SIMPLIFIED engineering approach. It is intended that an engineer capable more complete and rigorous analyses to special or unusual problems shall have latitude in the development of such designs and the evaluation of complex or combined stresses. Simplified analyses necessarily include conservatisms and these may include larger margins of safety. I can assure you however, this is not “due to the self interests of suppliers”. Similarly, Standards for pipe that set standard schedules (thicknesses) for piping are in the interest of controlling the number of components that must be kept in inventories – again no “self interests” are involved. In some cases pipe mills are asked to manufacture special thicknesses of pipe for some specific application. There is a significant additional cost involved in the manufacture of such “non-standard” pipe.

Thanks for your questions. There are likely other young engineers that have also thought of these same questions.

Regards, John.

 

[frd03]

Hello John Breen,



Thanks for your traditional attention!



If possible, I would like to continue a quite more this discussion and I really would apreciate your comments.



1) I did a plastic simulation with Abaqus FEA (including cyclic hardening) using a small pipe ( 11/2" sch40) including a 90 degree RL curve ( the lenghts and temperature were choosed to create a yelding on specific points ), in steel. I applied 15 sucessives loading cycles from 21 up to 300 oC.

I noticed, near the anchors/ends and also in the middle of internal radius of the curve, the similar stress values since the first and the last cycle ( around 27 kgf/mm2 >= Sy=25 kgf/mm2, so plasticity, ok ) when I take out the loading I can see the residual stresses but the final value under temperature loading I can not noticed a stress reduction after some cycles !



2) Concerning vibrations on pipings I am worried because, for example a hot oil line or a steam line can pulse to 10 or 30 Hz (fluid)due to the propellers and the alternate pressure following this frequency so the S x N curve can present a low life for the material pipe ( ex.100.000 hours)while the temperature of line keep it very regular ( f=1 !? ). How should be the safety procedure for this design ? ( and the Asme orientation ? ).



3) There are many cases where I could apply smaller thickness of carbon steel pipes if they were avaliables, keeping the safety design, but generally I do not have this option.

You know the amount of piping in middle petrochemical plant, and for ex. 20 % of reduction on weight materials could represent some hundreds of toneladas.

In the steel structural building market we have a large range of W-profiles (wide beams)so we can develop a design very adjusted for our needs. Why we are not a similar situation with pipes and components ?



Thanks Again !!



Francisco Dominguez