B31.3 Fatigue Curve equivalent
B31.3 Fatigue Curve equivalent
(OP)
In B31.3 we apply a SIF to a pipe bend, calculate the stress range and compare against the allowable stress. the allowable stress is adjusted by the stress range factor f. multiple stress ranges are handled by miners rule as presented in appendix F
As far as i am aware the SIF and F curve arise from the experimental work of Maerkl and the numbers are empirical. the stress calculated does NOT match that which would arise from an FE or theoretical analysis. clearly you cannot mix and match the SIF's with any other code approach.
the other approach is to determine the actual theoretical linear elastic stress and apply a suitable fatigue curve from DNV RP C203 / PD 5500 or similar code
So the the question is
I we perform an FE analsysis and determine the actual stress range, which fatigue curve (DNV RP C203) provides equivalent results?
we might say the following depending on where the stress is located
Curve B1 for parent material and ovalisation
Curve D seam welds due to ovalisation
Curve F1 circumential welds
And the reason for the question is that most of the pipework will be analysed in Caesar but the client wants to validate with abaqus, to get the confidence to proceed, I just wonder if anyone else out there has gone through a simialar approach, especially small differences matter as stress range and life correlate to the power 5.
thanks
donald
As far as i am aware the SIF and F curve arise from the experimental work of Maerkl and the numbers are empirical. the stress calculated does NOT match that which would arise from an FE or theoretical analysis. clearly you cannot mix and match the SIF's with any other code approach.
the other approach is to determine the actual theoretical linear elastic stress and apply a suitable fatigue curve from DNV RP C203 / PD 5500 or similar code
So the the question is
I we perform an FE analsysis and determine the actual stress range, which fatigue curve (DNV RP C203) provides equivalent results?
we might say the following depending on where the stress is located
Curve B1 for parent material and ovalisation
Curve D seam welds due to ovalisation
Curve F1 circumential welds
And the reason for the question is that most of the pipework will be analysed in Caesar but the client wants to validate with abaqus, to get the confidence to proceed, I just wonder if anyone else out there has gone through a simialar approach, especially small differences matter as stress range and life correlate to the power 5.
thanks
donald




