Blodgett's Allowable Fatigue Stress in Welds 1

I'm not familiar with Blodgett's methods, but in general a partial pen butt weld in fatigue service usually ends up failing prematurely. The only successful way I've seen it done is to treat the partial pen butt weld as a crack, and use a fracture mechanics approach (i.e. assume there is a crack in the material and model the crack growth). This might be conservative, but there is a pretty logical argument that the root of the partial pen weld will be ugly with a sharp point, that it much closer to a "crack" than a finished metal surface that is typical after welding.

I haven't answered your question directly, but thought you'd appreciate this extra bit of advice.

Andrew.

Andrew O'Neill
Specialist Mechanical Engineer
Australia

Maybe calc the max principal tensile stress and calc the max shear stress, both using the combined loads, then check both against the allowable values, and use the most critical one.

BUT, those allowables might not be valid for a partial pen weld.

If you're familiar with the eurocodes, there's an entire section with fatigue details and categories.
Partial pen welds are not included in these tables (as they're not suitable for fatigue design), however you'll find butt welds with transverse cracking there that in practice could be applied for your case (or at least be a good approximation).