1980c3
Mechanical
- Mar 11, 2022
- 19
Hi all,
I have a scenario in which semi-high cyclical stresses have been found near a steel flange/tube connection via strain gauges. For reference, the OD of the steel tube is >2ft and the wall thickness is >0.25 inches. The steel flange bolts to something else, and I am expecting normal loads in the form of mostly bending and circumferential shear loads in the form of torque to pass through the flange (and ultimately the weld and tube).
Do any of you have any experiencing qualifying this in a FAT class per IIW? Some of the No. 400s are alluring to me but I don't think they really can capture that in the case of a circle weld that there's no inherent eccentricity for the fillet weld to deal with (think a simple flat plate T-joint with one sided fillet welds). Maybe No. 822, but the diagram implies double-sided fillet welds...No. 913 is probably the best fit, but it really only considers weld root failure (through the throat I guess) and no guidance for the toe.
Perhaps I am just better off trying the hot-spot method regardless?
I have a scenario in which semi-high cyclical stresses have been found near a steel flange/tube connection via strain gauges. For reference, the OD of the steel tube is >2ft and the wall thickness is >0.25 inches. The steel flange bolts to something else, and I am expecting normal loads in the form of mostly bending and circumferential shear loads in the form of torque to pass through the flange (and ultimately the weld and tube).
Do any of you have any experiencing qualifying this in a FAT class per IIW? Some of the No. 400s are alluring to me but I don't think they really can capture that in the case of a circle weld that there's no inherent eccentricity for the fillet weld to deal with (think a simple flat plate T-joint with one sided fillet welds). Maybe No. 822, but the diagram implies double-sided fillet welds...No. 913 is probably the best fit, but it really only considers weld root failure (through the throat I guess) and no guidance for the toe.
Perhaps I am just better off trying the hot-spot method regardless?