Method and apparatus for estimating state of health of battery

Patent metadata

Field	Value
Publication number	US-11422192-B2
Application number	US-202017064013-A
Country	US
Kind code	B2
Filing date	Oct 6, 2020
Priority date	Oct 7, 2019
Publication date	Aug 23, 2022
Grant date	Aug 23, 2022

Abstract

A method of estimating a state of health of a battery, the method being performed by a computing apparatus, the method including: preparing a trained artificial neural network; generating input data by measuring at least one parameter of a battery; acquiring a plurality of output values each corresponding to a plurality of classes by inputting the input data into the trained artificial neural network; and generating a state of health estimation value of the battery using a plurality of preset health state sections each corresponding to the plurality of classes and the plurality of output values each corresponding to the plurality of classes.

First claim

What is claimed is:

1. A method performed by at least one computing apparatus, the method comprising:
preparing a trained artificial neural network;
generating input data by measuring at least one parameter of a battery;
acquiring a plurality of output values corresponding to a plurality of classes by inputting the input data into the trained artificial neural network; and
generating a state of health estimation value of the battery using a plurality of preset health state sections corresponding to the plurality of classes and the plurality of output values corresponding to the plurality of classes,
wherein the state of health estimation value of the battery is calculated by multiplying representative values of the plurality of health state sections corresponding to the plurality of classes and the plurality of output values corresponding to the plurality of classes to compute a plurality of products and summing the plurality of products.

2. The method of claim 1 , wherein the at least one parameter comprises a voltage of the battery and a current of the battery.

3. The method of claim 1 , wherein the at least one parameter comprises a voltage of the battery, a current of the battery, and a temperature of the battery.

4. The method of claim 1 , wherein the generating of the input data comprises:
measuring the at least one parameter of the battery in operation in accordance with a preset sampling period; and
generating the input data based on measurement values of the at least one parameter measured for a preset first time.

5. The method of claim 1 , wherein the generating of the input data comprises:
generating measurement values of the at least one parameter by measuring the at least one parameter of the battery in operation in accordance with a preset sampling period;
generating a plurality of input sub data based on measurement values of the at least one parameter generated for a preset first time after a preset second time has elapsed; and
generating the input data based on the plurality of input sub data.

6. The method of claim 1 , wherein the plurality of output values corresponding to the plurality of classes are probability values that the input data belongs to each of the plurality of classes.

7. A method performed by at least one computing apparatus, the method comprising:
preparing a trained artificial neural network;
generating input data by measuring at least one parameter of a battery;
acquiring a plurality of output values corresponding to a plurality of classes by inputting the input data into the trained artificial neural network; and

generating a state of health estimation value of the battery using a plurality of preset health state sections corresponding to the plurality of classes and the plurality of output values corresponding to the plurality of classes,
wherein the preparing of the trained artificial neural network comprises:
generating an artificial neural network;
preparing measurement data acquired by measuring the at least one parameter of the battery in each of the plurality of preset health state sections;
generating a plurality of training data labeled with a class of the plurality of classes to which they belong, based on the measurement data and the corresponding preset health state sections; and
training the artificial neural network using the plurality of training data to prepare the trained artificial neural network.

8. The method of claim 7 , wherein the preparing of the plurality of training data comprises:
preparing a battery electrochemical model including model parameters;
generating a plurality of model parameter data of the battery electrochemical model each corresponding to the plurality of preset health state sections using the measurement data;
generating a plurality of synthesized voltage data each corresponding to the plurality of preset health state sections by inputting a plurality of current data into the battery electrochemical model to which each of the plurality of model parameter data is applied; and
generating the plurality of training data based on the current data and the plurality of synthesized voltage data.

9. The method of claim 8 , wherein the preparing of the plurality of training data comprises:
further generating a plurality of synthesized temperature data each corresponding to the plurality of preset health state sections by inputting current data into the battery electrochemical model to which each of the plurality of model parameter data is applied; and
generating the plurality of training data based on the current data, the plurality of synthesized voltage data, and the plurality of synthesized temperature data.

10. The method of claim 7 , wherein the artificial neural network is generated based on a multi-layer perceptron (MLP).

11. The method of claim 1 , wherein
the trained artificial neural network comprises an input layer, at least one hidden layer, and an output layer, and
the output layer comprises a softmax function that outputs a probability that the input data input into the input layer belongs to each of the plurality of classes.

12. An apparatus for estimating a state of health of a battery, the apparatus comprising:
memory configured to store input data generated by measuring a trained artificial neural network and at least one parameter of the battery; and
at least one processor configured to acquire a plurality of output values corresponding to a plurality of classes by inputting the input data into the trained artificial neural network and to estimate a state of health estimation value of the battery using a plurality of preset health state sections corresponding to the plurality of classes and the plurality of output values corresponding to the plurality of classes,
wherein
the input data is generated based on a plurality of input sub data, and
the plurality of input sub data are generated based on measurement values of the at least one parameter generated by measuring the at least one parameter of the battery in operation in accordance with a preset sampling period, and
each of the plurality of input sub data is generated based on measurement values of the at least one parameter generated for a preset first time after a preset second time has elapsed.

13. The apparatus of claim 12 , wherein the at least one parameter comprises a voltage of the battery and a current of the battery or comprises a voltage of the battery, a current of the battery, and a temperature of the battery.

14. The apparatus of claim 12 , wherein the input data is generated based on measurement values of the at least one parameter generated by measuring the at least one parameter of the battery in operation in accordance with a preset sampling period for a preset first time.

15. The apparatus of claim 12 , wherein the plurality of output values corresponding to the plurality of classes are probability values that the input data belongs to each of the plurality of classes.

16. The apparatus of claim 12 , wherein the state of health estimation value of the battery is calculated by multiplying representative values of the plurality of health state sections each corresponding to the plurality of classes and the plurality of output values each corresponding to the plurality of classes to compute a plurality of products and summing the plurality of products.

17. The apparatus of claim 12 , wherein
the trained artificial neural network comprises an input layer, at least one hidden layer, and an output layer, and
the output layer comprises a softmax function that outputs a

Assignees

• Samsung Sdi Co Ltd
• Postech Res & Business Dev Found

Inventors

• Lee Jang-Woo
• Kim Jungsoo
• Chun Huiyong
• Han Soohee
• Lee Tae-Kyung

Classifications

• G06N3/0499

  Feedforward networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• H01M10/486

  for measuring temperature · CPC title

• G01R31/3842

  combining voltage and current measurements · CPC title

• G01R31/396

  Acquisition or processing of data for testing or for monitoring individual cells or groups of cells within a battery · CPC title