RUL prediction of lithium batteries based on DLUKF algorithm

摘要

Lithium batteries have been widely used in various fields, but there are still some safety problems. Once an explosion occurs, people's lives and property will be damaged seriously. In order to strengthen the safety of lithium batteries, it is necessary to master the maintenance of lithium batteries technology. Accurate prediction of the remaining useful life (RUL) of lithium batteries is beneficial to the maintenance of the battery, so that its safety can be improved. The traditional unscented Kalman filtering (UKF) method has been used in RUL prediction of lithium batteries, but there are still some problems such as low prediction accuracy because of the system’s high non-linearity. In this paper, an improved unscented Kalman filtering method for predicting the RUL of lithium batteries is used. The formula for calculating the weights by particle filter (PF) is used to change the weights in the UKF, and then in order to get the state and covariance at the next moment, the changed weights are additionally used in the measurement mechanism. The new method is verified more accurately by the open source battery capacity decay data from Center for Advanced Life Cycle Engineering.