An online prediction method of lithium battery pack health status based on multi-physics simulation and neural network method

Abstract

The invention relates to an online prediction method for the health state of a lithium battery pack based on multi-physical field simulation and a neural network method. The steps include: constructing a multi-physical field simulation model of a lithium battery pack, and carrying out multi-physical field simulation of a lithium battery pack through workload analysis Test, model verification and analysis; based on the simulation test analysis results, combined with the experimental data, construct and train the neural network model for the health state prediction of lithium battery packs, including the neural network model for multi-physics simulation and health state degradation; in lithium battery In the group use stage, local operating data is collected and processed, and the neural network model is used to conduct global physical representation analysis and lithium battery cell degradation analysis, and then predict the health status of lithium battery packs. This method combines the advantages of model-based and data-based methods, and can realize fast online health status prediction of lithium battery packs.

Description

[0001] Technical field [0002] The invention relates to the field of health state prediction, in particular to an online prediction method for the health state of a lithium battery pack based on multi-physics simulation and a neural network method. Background technique [0003] Lithium batteries have the characteristics of good safety performance and long cycle life, and have been widely used in power systems such as aviation, aerospace, and automobiles. Lithium battery pack consists of multiple lithium battery cells in series and parallel. It is a highly nonlinear system including complex physical and chemical changes. The state of health is often used to describe the current performance state of the lithium battery pack relative to the capability of the new battery pack, which is quantitatively described in the form of a percentage. Mastering the health status of the lithium battery pack helps to better manage the battery, avoid abuse such as overcharge and overdischarge, ...

AI Technical Summary

Problems solved by technology

The models involved in lithium battery packs mainly include electrochemical models, equivalent circuit models, thermal models, fluid dynamics models, etc. Due to the complexity of calculating and solving these models, it is difficult to directly apply them to the online real-time health status prediction of lithium battery packs.

Images