In the last posting, a mathematical model for predicting fatigue performance from hardness was introduced. The following equation, developed by Roeselle and Fatemi, describes the prediction of strain-life curves from Brinell hardness values(1):
It becomes of interest to determine if good a correlation between predicted and measured strain life curves can be obtained for high hardness low alloy steels. These have martensite microstructures as opposed to the ferrite-pearlite microstructures found in normalized carbon steels.
Figure 1 shows a comparison of the strain-life fatigue results obtained from several sets of test data obtained for SAE 4140 quenched and tempered alloy steel. The hardness values of the steels ranged from 353 to 390 BHN. Superimposed on the data is a strain-life curve predicted from the above equation using an average hardness value of 375 BHN.
As can be seen, the predicted strain-live curve correlates quite well with the experimental data. The predicted exhibits somewhat more conservative performance than the experimental data, especially at lives approaching run-out. This is largely due to the fact that the above equation cannot predict run out per-se.
As was noted in the earlier posting on normalized steels, the prediction methodology cannot account for the effects of variations in microstructure, non-metallic inclusions, prior austenite grain size, surface condition or residual stress. Thus, care must be exercised in using the prediction model for design purposes.
(1) M. Roessle and A. Fatemi, International Journal of Fatigue, Vol. 22, 2000