A method for predicting a remaining useful life of a lithium battery based on a wavelet denoising and a relevance vector machine, relating to a method for estimating health condition and predicting remaining useful life of lithium battery, includes steps of: (1) obtaining health condition data of each of charge-discharge cycles of the lithium battery by measurement; (2) processing capacity data measured of the lithium battery with wavelet double denoising; (3) calculating a capacity threshold where the lithium battery fails; (4) referring to capacity data and charge-discharge cycle data of the lithium battery, applying a differential evolution algorithm for optimizing a width factor of the relevance vector machine; and (5) predicting the remaining useful life of the lithium battery with the relevance vector machine optimized by the differential evolution algorithm. The method is simple and effective, which can accurately predict remaining useful life of lithium battery.
展开▼
