The invention belongs to the technical field of battery SOH prediction, and particularly relates to an improved data segmentation
lithium ion battery SOH
estimation method based on a Transform
encoder. Comprising the following steps: S1, inputting
lithium ion battery charge-
discharge cycle data into a model, S2, improving data segmentation, S3, carrying out normalization and alignment
processing to obtain two-dimensional
time sequence change of the data, S4, carrying out
convolution on the processed two-dimensional data by using
convolution blocks with different sizes to extract local features, S5, converting an input sequence into continuous representation by an
encoder, and S5, carrying out data segmentation by using a two-dimensional model. And S6, inputting a full-connection layer to obtain an
estimation result, capturing information in the full-connection layer, modeling features of the full-connection layer, and learning to obtain a dependency relationship in the features. The method provided by the invention has good performance in
root mean square error, mean absolute error and
maximum error, and by comparing different models, the performance of the model provided by the invention is remarkably improved.