This paper compares 3 methods for determining residual stresses in inductio
n-hardened shafts, as function of the depth beneath the surface. The method
s are X-ray diffraction, neutron diffraction and a simulation of the induct
ion-hardening process. The simulation is performed with a finite-element pr
ogram called SYSWELD. The residual stresses are needed to be able to calcul
ate the fatigue life for the shafts. Fatigue tests have shown that cracks c
an initiate in the core of the shaft. For both the Xray and the neutron dif
fraction technique, it is necessary to do some machining of the shaft to be
able to measure beneath the surface. Machining of specimens for the neutro
n diffraction measurement means cutting the shaft into a plate and a disc.
Hence, in this case, it is not possible to analytically calculate back to t
he original stresses. Good agreement is found between simulation and X-ray
diffraction measurement, while the conformity with neutron diffraction meas
urements is not so good. Simulations of different hardening depth show a re
latively large influence on the residual stress magnitude.