Rolling Bearing Stress Based Life—Part II: Experimental Calibration and Validation

The stress based life model described in Paper I was calibrated using a large database of experimental results from a global quality audit test program as well as special development tests conducted for this validation effort. All tests used in the calibration of the new model were case carburized or through hardened tapered roller bearings. The initial model comparison to test results showed very good correlation with a median ratio of test life to calculated life very close to one. Higher accuracy of the stress based model compared to the traditional factor based method was also demonstrated by narrower confidence bands (less data scatter). Validation testing of case carburized tapered roller bearings as well as through hardened spherical roller bearings was also conducted under expanded test conditions beyond those utilized in the quality audit test program (i.e., high and low load, high and low λ, imposed misalignment, and heavy inner ring interference fits). Although the median ratio of relative life for the validation testing showed the stress based method to be conservative and well above one, the stress based method still showed better accuracy than the traditional factor based method as well as narrower confidence bands for this additional body of experimental data. The conservative results can be explained by use of high quality steel and manufacturing processes in a prototype facility and not series production equipment as was the case with the quality audit database.