Run-out based crossed roller bearing life prediction by utilization of accelerated testing approach and FE numerical models

Surface wear and spalling phenomena are the most important factors affecting roller bearing performance and operational lifetime, in terms of number of cycles the bearing can operate. Although bearing life has been traditionally defined by the onset of spalling, as specified by international standard ISO 281, this definition does not always reflect the real operational conditions, since bearing wear can also represent a failure criterion when excessive run-out occurs. In such cases, bearing life needs to be predicted according to the amount of bearing run-out resulting from wear due to load and number of cycles. In the current study, the wear phenomenon has been modeled using the Archard wear and Lemaitre damage models, both included in a numerical simulation of a roller bearing assembly. Wear and damage models, along with the results of accelerated life experiments, have been used to develop a reliable method enabling bearing life prediction with a minimal number of material characterization tests. Accelerated life testing experiments on real bearings have been performed to confirm the reliability of the methodology, showing good correlation with numerical simulation results and proving that the developed model can be utilized for prediction of the lifetime of bearings.