bolt design against fatigue

huangyhg

bolt design against fatigue
in bs4604 it refers to calculating the torque applied to a bolt based upon 0.6 of the proof load, or uts as i read it. for fatigue applications, however, it says that only 0.5 of the proof load should be used. this appears contradictory to the advice usually given to torque the bolt to almost yield stress to reduce the load variation in the bolt and failure from fatigue. is there any reason why the design standard should recommend reducing the bolt torque for fatigue load applications?
corus
from my experience, higher than recommended pretension results in early fatigue failure of bolts, thus logically a lower pretension should prolong the life of the same bolt.
hth
vod
i agree with your thoughts, a higher preload reduces the load cycle reacted by the bolt.  however, i don't think i'd load as much as you're suggesting (all the way up to fty) ... i'm nervous about yielding bolts on installation (together with the in-service inspection issues ... maintaining the preload).
i might select a torque so that the maximum preload (the upper range of the scatter for the method used to measure torque, plis, torque wrench, ...) was fty (maybe 95%fty).  a consideration should also be how much tension is expected in service ? (lower tension loads, i'd reduce the preload slightly).
rb1957, the amount of preload depends on the grade of bolt and the ductility as i understand it. a grade 4.6 bolt can be preloaded to yield as the margin between uts and yield is large. for a higher grade bolt like an 8.8 the preload is about 75% of yield as the margin between yield and uts is less.
from the theory if the maximum preload can be achieved then the stress range on the bolt should be less and as such the chance of fatigue failure reduced. i can only presume that a higher preload increases the mean stress in the bolt and thus hastens failure even though the tress range has been reduced, as voyagethingy implies. i wondered if there was any reference to this to satisfy other's arguments.
corus
i think this link explains the purpose of proper preloading quite well: