Lithium battery nuclear temperature evaluation method and system based on transfer learning

Abstract

The invention provides a lithium battery nuclear temperature evaluation method and system based on transfer learning. The method comprises the following steps: acquiring target domain lithium battery parameter data; obtaining a lithium battery nuclear temperature evaluation result according to the obtained lithium battery parameter data and a target domain neural network model, wherein the target domain neural network is obtained by training according to historical data of a source domain lithium battery to obtain a source domain neural network model, and retraining a full connection layer of the source domain neural network model according to target domain data. According to the invention, transfer learning is adopted, and the established source domain model is transferred to the target domain lithium battery, and the nuclear temperature of other lithium batteries in the target domain can be accurately estimated with only a small amount of information of the target domain being used as training data, so that the model training progress is accelerated, and the training efficiency is improved.

Description

technical field [0001] The invention relates to the technical field of lithium-ion batteries, in particular to a method and system for evaluating the nuclear temperature of lithium batteries based on transfer learning. Background technique [0002] The statements in this section merely provide background art related to the present invention and do not necessarily constitute prior art. [0003] Lithium-ion batteries are widely used in the power supply of electric vehicles and energy storage systems due to their excellent properties such as high power, high energy density, and long cycle life. However, the heat generated during charging and discharging of lithium-ion batteries will directly affect the cycle life, reliability and safety of the battery. The battery management system (Battery Management System, BMS) is the key to ensure the safe and efficient operation of lithium batteries. Therefore, it is very important to monitor the battery temperature through the BMS to pr...

AI Technical Summary

Problems solved by technology

[0006] However, the inventors have found that the gradient disappearance and explosion problems are prone to occur during the training process of the recurrent neural network (RNN), which is only suitable for processing short time series; in addition, the data used for training the model is often not the same distribution or between distributions. There are obvious differences, such as between different battery data sets, due to different internal electrochemical compositions and complex external environmental factors, etc., even the temperature curves of the same type of battery are different, so the battery temperature data does not belong to the same distribution. Therefore, the model built with one battery data set is difficult to generalize to other battery data sets

Secondly, the large amount of data required for model training is difficult to obtain or takes a long time, so the obtained data is very limited and precious. Traditional machine learning methods require more data for long-term training, which limits the applicability of the algorithm.

Images