Fatigue Analysis

[pipesvessels1]

ALL,

I need to a basic fatigue analysis on Vessel nozzles using PD5500 using SN curves. Do I start with the Stress range, Sr, and then get the no. of cycles or do I calculate the no. of cycles and then get the allowable stress range from the curve?

Ultimately I require a suitable thickness of vessel for a given series of stress cycles ie 4 different stress ranges at 4 different no of cycles.

All help appreciated,

Cheers,

JohnOD

 

[ruth44]

Normally one calculates the stress range first based on the loads and piping /vesssel layout. Than get the allowable number of cycles from the S-N curve and compare this with the specified (expected) number of cycles.
If the allowable number is too small you have to change the design.

 

[davefitz]

you can start with either choice, if it is the design stage. If the vessel is already built, then you would calculate the stress range imposed and then determine the cyclic life consumption, n/N.

You may have available some other degrees of freedom. For example , if it is a thick walled vessel, you might be able to define the allowable rate of change of vessel temperature in order to limit the thermal stress for a specific type of cycle.

Also, if there are 4 different cycles and they are imposed in a random manner , then you might be able to calculate a smaller total life consumption by using the rainflow cycle counting algorithm ( use a monte carlo simulation for the sequence of cycles).

There is a good summary in the EU PED standard EN-12952-4:2000 app B