Degradation modeling and remaining useful life prediction for electronic device under multiple stress influences

Power driver devices have functions such as current amplification and power conversion, making them key components of electronic systems. Their degradation is affected by multiple types of stress, making it difficult to establish an accurate degradation model. To monitor the degradation state of electronic devices and predict their remaining useful life at different moments, this paper obtains the degradation data of the samples by carrying out accelerated degradation experiments of power driver devices under the influence of multiple stresses, and proposes a new multi-stress-coupled accelerated degradation model based on the Wiener process. This model associates the accelerated stress with the drift coefficient of the Wiener stochastic degradation model. Finally, the MLE-SA optimization algorithm is used to obtain the unknown parameter values of the model. The method proposed in this paper incorporates accelerated stress factors into the stochastic degradation model, effectively improving the prediction accuracy and interpretability of the stochastic degradation model. To verify the accuracy of the model, the paper conducted comparative experiments on the accelerated degradation data of power driver devices and publicly available data. The results show that the multi-stress coupled accelerated degradation model based on the Wiener process proposed in this paper can well fit the accelerated degradation data of power driver devices, and the goodness-of-fit for the public dataset above 0.9, indicating that the model proposed in the paper has high accuracy.