Apparatus and method with battery state estimation

Patent metadata

Field	Value
Publication number	US-11630157-B2
Application number	US-201916281438-A
Country	US
Kind code	B2
Filing date	Feb 21, 2019
Priority date	Sep 20, 2018
Publication date	Apr 18, 2023
Grant date	Apr 18, 2023

Abstract

A processor-implemented method of estimating a state of a battery includes acquiring current information and voltage information of a battery; determining time interval values based on the acquired current information such that current integration values corresponding to the time variation values satisfy a condition; determining voltage values corresponding to the determined time interval values in the acquired voltage information; and determining state information of the battery based on the determined voltage values.

First claim

What is claimed is:

1. A processor-implemented method of estimating a state of a battery, the method comprising:
acquiring current information and voltage information of a battery;
determining time interval values based on the acquired current information such that current integration values corresponding to the time interval values satisfy a condition, wherein the current integration values are equal to each other, and at least one of the time interval values is non-equal to others of the time interval values;
determining voltage values corresponding to the determined time interval values in the acquired voltage information;
recognizing a voltage variation pattern during a time interval, including two or more of the time interval values, based on the two or more time interval values and voltage values corresponding to the two or more time interval values;
determining state information of the battery using the recognized voltage variation pattern based on the determined voltage values; and
outputting the determined state information to a display.

2. The method of claim 1 , wherein the determining of the time interval values comprises:
acquiring current integration information as an integral of the acquired current information over times;
dividing the current integration information to obtain the current integration values; and
extracting a corresponding time interval value for each dividing point of the current integration values.

3. The method of claim 1 , wherein the determining of the state information comprises:
dividing the acquired voltage information based on the determined time interval values.

4. The method of claim 3 ,
wherein, among dividing points of the divided voltage information, a time interval between a pair of adjacent dividing points of the voltage information is greater than another time interval between another pair of adjacent dividing points, and
wherein each of the time interval and the other time interval comprises one or more of the time interval values.

5. The method of claim 4 , wherein a current of the battery decreases in the interval between the portion of dividing points.

6. The method of claim 1 , wherein the determining of the state information comprises:
extracting a voltage value corresponding to one or more time interval values associated with a scan order among the determined time interval values from the acquired voltage information;
generating a feature vector based on the one or more time interval values and the extracted voltage value; and
inputting the generated feature value to a state estimation model.

7. The method of claim 6 , wherein the state estimation model comprises a recurrent neural network (RNN) including a long short-term memory (LSTM).

8. The method of claim 6 , wherein a length of the feature vector is the same in a case in which the battery is fast-charged and in a case in which the battery is slow-charged.

9. A non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform the method of claim 1 .

10. An apparatus for estimating a state of a battery, the apparatus comprising:
one or more processors configured to:
acquire current information and voltage information of a battery,
determine time interval values based on the acquired current information such that current integration values corresponding to the time interval values satisfy a condition, wherein the current integration values are equal to each other, and at least one of the time interval values is non-equal to others of the time interval values,
determine voltage values corresponding to the determined time interval values in the acquired voltage information,
recognize a voltage variation pattern during a time interval, including two or more of the time interval values, based on the two or more time interval values and voltage values corresponding to the two or more time interval values,
determine state information of the battery using the recognized voltage variation pattern based on the determined voltage values, and
output the determined state information to a display.

11. The apparatus of claim 10 , wherein the one or more processors are configured to:
acquire current integration information as an integral of the acquired current information over times,
divide the current integration information to obtain the current integration values, and
extract a corresponding time interval value for each dividing point of the current integration values.

12. The apparatus of claim 10 , wherein the one or more processors are configured to divide the acquired voltage information based on the determined time interval values.

13. The apparatus of claim 12 , wherein:
among dividing points of the divided voltage information, a time interval between a pair of adjacent dividing points of the voltage information is greater than another time interval between another pair of adjacent dividing points, and
each of the time interval and the other time interval comprises one or more of the time interval values.

14. The apparatus of claim 13 , wherein a current of the battery decreases in the interval between the portion of dividing points.

15. The apparatus of claim 10 , wherein the one or more processors are configured to:
extract a voltage value corresponding to one or more time interval values associated with a scan order among the determined time interval values from the acquired voltage information,
generate a feature vector based on the one or more time interval values and the extracted voltage value, and
input the generated feature value to a state estimation model.

16. The apparatus of claim 15 , wherein the state estimation model is based on a recurrent neural network (RNN) including a long short-term memory (LSTM).

17. The apparatus of claim 15 , wherein a length of the feature vector is the same in a case in which the battery is fast-charged and in a case in which the battery is slow-charged.

18. A processor-implemented method of estimating a state of a battery, the method comprising:
acquiring current information and voltage information of a battery;
determining, based on the current information, current integration values that are equal in value;
determining time interval values based on the determined current integration values, wherein at least one of the time interval values is non-equal to others of the time interval values;
recognizing a voltage variation pattern during a time interval, including two or more of the time interval values, based on the two or more time interval values and voltage values corresponding to the two or more time interval values;
determining state information of the battery using the recognized voltage variation pattern based on voltage values corresponding to the determined time interval values in the acquired voltage information; and
outputting the determined state information to a display.

19. The method of claim 18 , wherein the determining of the state information of the battery comprises:
generating a feature vector based on one or more time interval values, of the time interval values, and one or more first corresponding voltage values, of the voltage values;
inputting the feature vector into a long short-term memory (LSTM) to generate an LSTM output vector;
generating a subsequent feature vector based on one or more subsequent time interval values, of the time interval values, and one or more subsequent corresponding voltage values, of the v

Assignees

• Samsung Electronics Co Ltd

Inventors

• You Kaeweon
• Park Sangdo

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• Y02E60/10

  Energy storage using batteries · CPC title

• G01R31/3648

  comprising digital calculation means, e.g. for performing an algorithm · CPC title

• G01R31/3828 Primary

  using current integration · CPC title