II would post your model and your reference for the results you expect.
Short answer: I don't know.
The most recent thinking is that, if all of your cycles are below the endurance limit, they do no damage. But if one cycle is above the endurance limit, then the endurance limit temporarily goes to zero and all following cycles do damage. This would account for the fact that Miner's rule doesn't seem to work, even though it seems reasonable.
This effect is relatively easy to account for in strain based fatigue with time history. But maybe not so easy with stress based fatigue and no time history.
I don't know how SW does fatigue studies, but one reasonable way to account for this effect and remain conservative is to assume materials have no endurance limit, and that all cycles do damage as if the S/N curve continues to zero.
It is a constant amplitude analysis, so all cycles should be below the endurance limit.
You did give me an idea though, just make the last data point in the SN curve some ridiculously large number of cycles, say 100 biillion cycles, and the program will round the resulting damage to zero. This tricks it into giving me the result I wanted.