Yves Champoux has some interesting measurements of vertical acceleration from bumps in his 2007 Paper "Bicycle Structural Dynamic"
He measured vertical shock forces of ~800 and ~1100 N for front and rear wheels traversing a bump. I can't find any reference to the size of the bump but it appeared to be a piece of 12mm dowel. These loads would be of the order of twice the static load in the tests and since the accelerometers show the static load as a zero baseline you'd need to add them for total load eg 1200 N and 1600 N respectively.
This would probably make a reasonable base for estimation of repetitive load for the bearing. The maximal load could be calculated by considering the usual 15% rule for tyre compression: turning that on its head, the largest load before rim damage will be around 7 times the static load.
The maximal bending moment could be calculated by considering the shock load as vertical and mutliplying be the moment arm which will be the horizontal disatance between the hub centre and the headset centre and thus equal to axle to crown distance x cos (headtube angle) + rake x sin (headtube angle).
Last edited by Mark Kelly; 06-15-11 at 05:57 AM.