Pressure fatigue in B31.3 piping 1

[mech2926]

Is evaluation for pressure fatigue necessary for B31.3 piping in "normal" service? I've been looking through the Code and pressure fatigue seems to mainly be a concern for high-pressure piping (defined as P/SE > .385) and/or piping with "thick" walls.

I could just go ahead and calculate fatigue life based on hoop stress versus endurance limit, but I'm not sure what to do about elbows, tees, etc. For thermal fatigue (longitudinal stress), I would just use the SIFs in the Code, but I'm not sure whether these apply to pressure (hoop stress).

For the particular system in question, there are < 10,000 cycles of pressure.

 

[aroundhere]

What kind of pressures are you looking at? Is it 10K cycles of ramping up/down in temp? THermal fatigue I would think depends on temp.
Is the 10K cycles going to be over weeks, months, years?

 

[mech2926]

Pressures are less than or equal to 2400 psig. There would be temperature cycles occuring simultaneously with the pressure cycles. I'm assuming the cycling would be over a period of years.

 

[a234f56]

Fatigue can lead to fracture under repeated or fluctuating stresses having a maximum value less than the tensile strength of the material, known as the fatigue or endurance limit. Sources for fatigue include thermal/stress cycling, rotation or vibration, like that produced by reciprocating compressors and positive displacement pumps. The guidance presented in the ASME B31.3 for checking cyclic stress levels is based on low cycle/high stress fatigue, e.g. thermal stresses associated with infrequent start-up/shutdown cycles. In a vibrating system, the concern is high cycle/low stress cycling but ASME B31.3 does not explicitly address high-cycle fatigue.

High cycle fatigue is of particular importance in the presence of flaws, e.g. fabrication cold (hydrogen) cracks. Under these circumstances, the majority of the components’ life will be spent propagating the crack. In a nominally defect free welded joint, fatigue life will incorporate a substantial crack initiation period, as well as a crack propagation period. Understandably therefore, the fatigue design guidance presented in ASME B31.3 is based on nominally defect free welds. The sources of fatigue loading that have to be considered in the design of pipes are more numerous than those for pressure vessels.

In addition to internal pressure fluctuations, pipes may also be subjected to external forces from direct loads, bending moments, and torques resulting from. (These low cycle fatigue loads are accounted for in pipe flexibility design analyses). In relatively flexible small diameter pipes, a number of failures have been caused by high-cycle, resonant vibrations due, for example, to external vortexes, internal turbulent flow regimes, sustained relief valve discharge, etc. If the frequency of any of the modes of these vibrations coincide with the natural frequency of the pipe, substantial resonant vibrations can be produced. Nevertheless, in the absence of complex time-history cumulative damage analyses of the small diameter piping systems, small diameter piping support design is more often than not based on field experience.

Corrosion fatigue: In addition to design features that cause stress concentration, deep scratches, sharp corners, weld profiles etc, all can act as initiation points for fatigue. Further, in the presence of a corrosive environment, a pit can initiate fatigue. Often, multiple initiation points result dependent on the pit frequency. Short of changing the environment or stresses, shot peening is used to place the surface of a potential initiation site in compression. Stress relief, use of corrosion inhibitors and protective coatings, all have had some success in resisting corrosion fatigue.

Thermal fatigue occurs in equipment that experiences frequent changes in temperature. For instance, each start-up and shutdown induces thermal stresses, which, if significant in number, can lead to thermal fatigue. In particular, coke drums and reactors (heavy section welds) in cyclic temperature service are prone to thermal fatigue. Austenitic stainless steel is often used to clad the internal surfaces of thick walled vessels to protect the alloy steel substrate from, say, H2S/H2 environments.


Austenitic stainless steel exhibit significantly higher thermal expansion (more than 30%) than low alloy steels and start-up and shutdown can cause plastic deformation of the plastic layer and adjacent base material. Repeated thermal cycles can induce high strain, low cycle fatigue of the cladding. Roll bonded cladding is significantly more resistant to fatigue than weld overlay cladding. In the latter case, the requirement for some ferrite in the weld deposit induces sigma phase formation during post weld heat treatment that reduces fatigue resistance. To minimise the risk of thermal fatigue it is recommended that the heating and cooling rates in hydrotreater plants are slower than 40°C/hr.

 

[Korichnevijgigant]

Have you taken a look at API 579 annex B?

There is a quick fatigue assessment you can do to see if you need to consider it or not.

 

[rconner]

While I can't answer as to all the nuances of the latest code, it has been known for more than 30 years that some weaker piping materials can exhibit fatigue cracking issues in applications with generally far less pressure than is normally considered "high pressue" on such forums. An example are issues with pvc sanitary sewer force pipelines, where I believe disproportionate problems were reported as early as the First International Conference oo Plastic Pipe in New Orleans in 1981(see

http://www.drpipe.ir/pages/images/stories/pdf/pvc/21.pdf

). As others have hinted, I also would not assume that fatigue is only pressure induced, although that is of course the first thing one thinks about particularly in applications like some sewer forcemains, where there is by design frequent starting and stopping of pumps. I believe a quick search of these (with "Advanced Search" feature)and other forums with a few key words will likely reveal that some such problems continue to the present day.

B31.3 however might exclude extremely low pressure pipelines (e.g. 0<=P<=15psi)

 

[Korichnevijgigant]

Yes you should not only consider pressure, but also thermal effects and mechanical loading et cetera

 

[mech2926]

Thank you all for your posts.

The system I'm working on uses stainless steel (304, 316). Thermal expansion and pressure are the only loads that are expected to occur cyclicly. For these materials, since yield is about half of ultimate stress, and allowable stress is much lower than yield, theoretically I'm under the endurance limit (for hoop stress anyway) for pressure cycling, right?
Pipe fittings are a different matter (I think) because there could be stress concentrations, and that's what is bothering me. If I bought a copy of API 579 and found out that I need to consider pressure cycling fatigue: how would I do that? It seems like I would need to determine the actual stress in an elbow/tee/whatever, but B31.3 only has formulas for longitudinal "effective stress" (I think that's the term), not real longitudinal stress or hoop stress.

 

[if7005]

Here is what this code ( ASME B31.3 )said, if pressure rating excess #2500 psig :
K304.8 Fatigue Analysis
K304.8.1 General. A fatigue analysis shall be performed
on each piping system, including all components
and joints therein, and considering the stresses.....etc
this analysis shall be in accordance with the BPV Code, Section VIII,Division 2 or Division 3.9 The cyclic conditions shall include pressure variations as well as thermal variations or displacement stresses.


But within # 2500, fatigue stress only considered for displacement stress(thermal or imposed movement) as stress range factor (f) in it's allowable stress (?)