Bonded repairs can replace mechanically fastened repairs for aircraft struc
tures. Compared to mechanical fastening, adhesive bonding provides a more u
niform and efficient load transfer into the patch, and can reduce the risk
of high stress concentrations caused by additional fastener holes necessary
for riveted repairs. Previous fatigue tests on bonded Glare (glass-reinfor
ced aluminium laminate) repairs were performed at room temperature and unde
r constant amplitude fatigue loading. However, the realistic operating temp
erature of -40 degrees C may degrade the material and will cause unfavourab
le thermal stresses. Bonded repair specimens were tested at -40 degrees C a
nd other specimens were tested at room temperature after subjecting them to
temperature cycles. Also, tests were performed with a realistic C-5A Galax
y fuselage fatigue spectrum at room temperature. The behaviour of Glare rep
air patches was compared with boron/epoxy ones with equal extensional stiff
ness. The thermal cycles before fatigue cycling did not degrade the repair.
A constant temperature of -40 degrees C during the mechanical fatigue load
had a favourable effect on the fatigue crack growth rate. Glare repair pat
ches showed lower crack growth rates than boron/epoxy repairs. Finite eleme
nt analyses revealed that the higher crack growth rates for boron/epoxy rep
airs are caused by the higher thermal stresses induced by the curing of the
adhesive. The fatigue crack growth rate under spectrum loading could be ac
curately predicted with stress intensity factors calculated by finite eleme
nt modelling and cycle-by-cycle integration that neglected interaction effe
cts of the different stress amplitudes, which is possible because stress in
tensities at the crack tip under the repair patch remain small. For an accu
rate prediction it was necessary to use an effective stress intensity facto
r that is a function of the stress ratio at the crack tip R-crack tip inclu
ding the thermal stress under the bonded patch.