Remaining Useful Lifetime Prediction Based on the GRU and Stochastic Degradation Modeling

To further reduce the complexity of the algorithm and improve the performance of remaining useful lifetime (RUL) prediction, an RUL prediction method based on the gated recurrent unit (GRU) and stochastic degradation modeling is proposed. First, a stochastic degradation model based on the Wiener process is established and a weighted kernel function is used to represent its drift increment. Then, a GRU is established to predict the degradation increment of the Wiener process and derive the probability density function (PDF) of the RUL. Finally, based on the lithium battery degradation data, the validity of the method is verified.