Seems to me that axial tension won't affect the base plate, but only the anchor rods. Draw the FBD of the resultant force if you have tension and moment combined in your connection. The tensile side of your moment will be additive with the tension in the column and will pull on your anchor rods with a greater force. The base plate shouldn't be affected.

I agree with NetTension that the anchor rods will see the combined flexural tension plus direct tension. I disagree that it won't affect the base plate.

Presumably, the tension from the connected element to the anchor rods is resolved through flexure of the base plate. More tension in the rod equates to more flexure in the base plate. If the anchor bolts sit on chairs or straddle the connected element then them effect on the base plate might be minimal. It all depends on geometry.

You just need to look at the load path and figure out where the loads go.

That was an oversight on my end JLNG of course an increase in tension force in the rods needs to be reacted somewhere. Most likely in the base plate. Try this Link. See also page 25 of DG1, which describes how to handle tensile axial loads "For tensile loads, a simple approach is to assume the an-chor rod loads generate bending moments in the base plate consistent with cantilever action about the web or flanges of the column section (one-way bending)" (AISC DG1). This is only for the case where your anchor rods are outside of the column flanges. For the case where anchor rods are inside of the flange two-way bending must be considered.

Direct to your question:

1) For large moment that resulted in only a pair of anchors in tension, I think the case from DG1 is still applicable.

2) For small moment that resulted in all bolts are in tension, I suggest to add some weight to counter the lift force. As this case is not well addressed in literatures, as long as I can remember. (Note that net tension case does not apply, because part of the base plate will still in contact with the floor, prying action will occur)

I’m not sure I agree that all bolts in tension aren’t covered in literature, because DG1 directly addresses the case as I quoted. Do you disagree retired13?

Edit: this is a question not a challenge

I just enjoy the debate

NetTension,

No, I wouldn't suggest that is well covered, as the word, "assumed bending line" (Fig 3.1.1) is vague to me. The method DG1 employed was a modification to the ultimate strength design of reinforced concrete, with a pair of anchor bolt in tension, and a compressive stress block under the column compression flange. It works fine as the case I stated in my comment #1 above, but when all bolts are in tension, that is liking both top and bottom reinforcing steel are all in tension, how does that compressive stress block look then? That's the reason I wouldn't suggest it without further understanding. I think evaluation of prying action maybe one way to go, and FEM is another, if exact solution is desired.

If case 2 in my comment is the predominate case in your design, as suggested, add weight to diminish the lift force, or modify you framing, so the structure is not standing on its toes.

For simplicity, I use a rectangular stress block for compression.

Dik