Lithium ion battery residual life integrated prediction method considering uncertainty

Abstract

The invention provides a lithium ion battery residual life integrated prediction method considering uncertainty, a data noise quantification method based on model prediction error decomposition is proposed and is applied to residual life prediction based on particle filtering by considering data uncertainty. In consideration the uncertainty of the model, a lithium ion battery residual life prediction method based on model integration is proposed. The method comprises a data acquisition step, an individual model construction step, a model weight acquisition step, an integrated model construction step and a prediction step, the degradation state of the battery capacity is predicted by using an integrated model, and a prediction result of the residual life is obtained according to a failure threshold value of the battery capacity. According to the method, a model integration method based on time varying and degradation weighting is provided, an integration model based on optimal model weighting can be established to reduce the model uncertainty influence caused by model selection, the data uncertainty caused by data noise is reduced, and the prediction accuracy and robustness of the residual life of the lithium ion battery are improved.

Description

technical field [0001] The invention relates to the technical field of remaining life prediction of lithium-ion batteries, in particular to an integrated prediction method for remaining life of lithium-ion batteries considering uncertainty. Background technique [0002] Remaining Useful Life (RUL) prediction is the core problem of failure prediction and health management (Prognostics and Health Management, PHM). Taking lithium-ion batteries as the research object, as the core components for energy supply and storage in electric vehicles, portable electronic devices and aerospace systems, life problems such as degradation and failure of lithium-ion batteries make accurate remaining life prediction of lithium-ion batteries crucial. important. [0003] However, in the remaining life prediction of lithium-ion batteries, there are two types of problems, data and model uncertainties. First, data uncertainty is reflected in data noise caused by equipment, product internal and ext...

AI Technical Summary

Problems solved by technology

[0003] However, in the remaining life prediction of lithium-ion batteries, there are two types of problems of data and model uncertainty

First, data uncertainty is reflected in data noise caused by equipment, product internal and external environments, and measurement errors

Images