Calculating number of cycles in a piping system 2

[krspack]

I have several piping systems that are approximately 30 years old. I need to calculate the approximate number of cycles they have been subjected to during their life. I have been reviewing B31.3 para 302.3.5, but I am having trouble understanding the steps involved in calculating the number of cycles. Can someone clarify this method and/or have a simpler method for calculating cycles that is in conformance with the code.

Thanks!

 

[BigInch]

Are you having trouble with getting the value for N? Why?


**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[ApC2Kp]

krspack,

The calculation of intended number of cycles for a new plant in design is based on expected startup / shutdown sequences per year multiplied by number of years of required service life. For an existing plant you would be counting (not calculating) from the service records the plant's startup / shutdown service events. For a process plant with multiple batches each day, then it could be the number of batches per day multiplied by days of operation in each year multiplied by years of operation. A petroleum refinery might have a continuous process (batch time of 12 to 18 months) that could represent 1 cycle for the startup to shudown for the piping's thermal cycle.

There are calculations in the B31.1 / B31.3 Codes for partial cycles to include in summation of total cycles. The basic 20 year life of B31.3 process plant piping is based on 1 cycle per day for total of 7,000 cycles full stress range. Service at reduced stress ranges could be used to establish extended life beyond 7,000 cycles.

 

[krspack]

Thanks ApC2Kp,

That seems pretty straigtforward for a non-batch (continuous operation) like ours. I can certainly go back and estimate the number of startup and shutdowns.

I guess where I was getting caught in the weeds was looking at determining number of cycles through the calculation of bending and torsional stress based on equation 1d found in section: 302.3.5 Limits of Calculated Stresses Due to Sustained Loads and Displacement Strains. In this equation, it refers you to calculating # of cycles using a stress range determined in para. 319.4.4 (using bending and torsional stresses).

So to clarify, since we a continuous operation (with little temperature fluctuations), if I just look at the startups and shutdowns, and count each event as one cycle, this should be sufficient.?.

 

[Artisi]

Is this a homework problem?

 

[ApC2Kp]

krspack,

The thermal expansion cycles (or lack of cycling) is only one possible failure mode that might determine remaining life of a piping system. Pressure cycling could result in fatigue - maybe not common for small diameter piping, but the DeHaviland Comet airliner of 1950's suffered from fatigue failures due to cabin pressurization cycles. If there have been portions of piping with vibrations, then there could have been several repairs in those vibrating portions over the years. The wall thickness of all the 30 year old pipe needs to be surveyed for remaining thickness to establish its pressure capability and any remaining corrosion life. The pipe could have corrosion from outside due to environment as well as the inside due to the fluid characteristics. It could be prudent to perform a service pressure test of piping systems if they have been idle for an extended time. One of the more common causes of failure for old piping systems would be the structural steel supports being corroded - threaded rod hangers rusted away and dropping the pipe for example. One resource you may want to get would be ASME B31-G. There are other industry standards for evaluating existing pipelines or vessel / piping systems.

 

[racookpe1978]

Good caution about how the "remaining lives" in an older piping system may be affected by internal (or external) corrosion and wear.

But fatigue failure can get you several ways - some of which will actually be reduced by internal erosion and corrosion:

Consider "simple" fatigue failure: a high frequency vibration but relatively low movement like what caused failure in the Comets. As a pipe or vessel gets eroded/corroded, it flexes easier at places of high stress, so small cracks are slightly less likely. Large or catastrophic cracks from internal pressure become more likely. Stress corrosion becomes more dangerous if corrosion can start down in the little cracks.

But larger, slower movement like what he described from each plant startup and shutdown will usually induce lower stresses, and there be far lower number of plant startup cycles than if he had daily (or hourly) "reloading" heatup and cooldown cycles. Note the keyword and tricky phrase "slower cycles" though.

Slower, less frequent heating of large thick parts and flanges and vessel heads are the key to figuring his lifecycle: but it will be hard to tell exactly how many of his startup's and shutdown's the past 30 years were "nice and slow" compared to how many shutdown's were dangerously "quick and short". A power failure for example will cause immediate cooling of all the lines, no process flow, no cooling water slowdown period or pump flow to empty the pipes and vessels, rapid stress rises at vessel flanges, and (probably) a rapid "bump" on a cold system when power gets turned back on again.

10 or 16 quick shutdowns may have put him further down the fatigue path than 64 slow startups.

 

[davefitz]

An atuomated method of calculating the number of cycles , now used in european PED design, is the "rainflow cycle counting algoriythm". But to use this method , there must be a model of the piping system part that estimates the max stress or strain as a function of (a) rate of change of fluid pressure vs time ( if thermal stress is important) plus pressure vs time .

Piping loads ( static and vibration) would presumably be counted separately as the max stress for these loads is not likely to coincide with the peak thermal + pressure stress.

As the fatigue life consumption increases geometrically with stress range ( a doubling of stress range reduces the number of fatigue life cycles by a factor of 20) , high stress events ( ie pressure releif valve events, hydrostatic tests, cold startups) will neeed to be counted expecially accurately. For the same reason, the fatigue life will vary greatly according to the presumed stress concenration factor- an assumed pre-exisiting crack at a weld ( eg weld root pass, in a location that cannot be inspected and ground smooth) would have a much lower fatigue life than an assumed smooth surface.

And finally, the fatigue model used- if one assumes the likely location of first failure is the weld crack, then one wouuld use a fracture mechanics approach- this differs from the 1950's era fatigue model that remains in most sections of the ASME code.

 

[davefitz]

error in prior post:

should be "rate of change of fluid temperature vs time" , not pressure. must be slipping.

one common approximation is that the thermal stress of a pressure vessel is proportional to the (rate of change of temperature)* ( wall thickness squared)/ (thermal diffusivity) * SCF * constant. Roughly valid for thermal transients that are of a duration longer than 2*s^2/a, where s= wall thickness and a= thermal diffusivity.