ELM-UKF-based lithium battery remaining service life prediction method

The invention discloses an ELM-UKF-based lithium battery remaining service life prediction method. The method comprises the specific steps that (1) lithium battery constant pressure drop discharge time is selected as a lithium battery service life characteristic parameter; (2) the lithium battery constant pressure drop discharge time data are used to construct a lithium battery state update equation based on an extreme learning machine (Extreme Learning Machine, ELM); (3) the lithium battery state update equation and a constant pressure drop discharge time observation equation are used as a lithium battery performance degradation model; (4) based on the established lithium battery performance degradation model, a multi-stage unscented Kalman filter (Multi Unscented Kalman Filter, MUKF) method is used to carry out constant pressure drop discharge time prediction; (5) the model of the relationship between the constant pressure drop discharge time and the lithium battery practical capacity is built, wherein the model is based on the extreme learning machine; (6) the constant pressure drop discharge time predicted in the step (4) is used as the input of the model determined in the step (5) to acquire the future lithium battery practical capacity value; and according to a specified lithium battery failure threshold, the remaining lithium battery recycling cycle is finally estimated. According to the method provided by the invention, the health state of a lithium battery is monitored on-line; the future lithium battery service life characteristic parameter is predicted; and the remaining service life of the lithium battery is effectively evaluated.