Lifetime Prediction of Electrical Machines Under Nonstationary Loads by Coupled Thermal Model and Insulation Aging Assessment

Electrical machines (EMs) in electromechanical actuators need to be highly reliable, especially when driving safety-critical systems in more electric aircraft. However, the non-stationary loads faced during mission profiles will cause the EM to overheat, raising the risk of insulation failure. Motivated by the issue, this paper proposes a lifetime prediction framework based on multi-field coupling models. The main objective is to obtain the temperature curve during a non-stationary mission, and then predict the lifetime of the EM in the context of insulation thermal aging. Specifically, a lumped parameter thermal network model is calibrated based on electromagnetic field, fluid field simulations and temperature tests, which achieves the accurate temperature prediction of EMs under non-stationary loads. Accelerated thermal aging tests are carried out on typical insulation specimens for EMs. The obtained data are processed by a thermal degradation model based on the Wiener process and Arrhenius law, then a lifetime prediction model is developed at several lifetime percentiles for variable temperature operations. Finally, the proposed model is employed for predicting the lifetime of the EM under typical non-stationary load conditions.