Method and apparatus for learning and estimating battery state information

Patent metadata

Field	Value
Publication number	US-10228425-B2
Application number	US-201514732098-A
Country	US
Kind code	B2
Filing date	Jun 5, 2015
Priority date	Jun 24, 2014
Publication date	Mar 12, 2019
Grant date	Mar 12, 2019

Abstract

A method of estimating battery state information includes collecting battery information; partitioning the collected battery information into interval information in a predetermined range; accumulating the interval information; and estimating the battery state information based on the accumulated interval information and learning information determined in advance. A method of learning battery state information includes collecting battery information; partitioning the collected battery information into interval information in a predetermined range; accumulating the interval information; and determining learning information based on the accumulated interval information and reference information of a battery, the learning information being used to estimate the battery state information.

First claim

What is claimed is:

1. A processor-implemented method, performed in a computing apparatus in a battery management system, of estimating battery state information of a battery, the method comprising:
measuring, using plural sensors for the computing apparatus in the battery management system, respective battery information, associated with the battery, for a plurality of intervals, wherein lengths of the plurality of intervals are variably set dependent on resolutions of the plural sensors;
generating, using one or more processors of the computing apparatus, interval information by partitioning the measured battery information into the interval information in a predetermined range;
accumulating, using the one or more processors, the interval information for each of the plurality of intervals;
generating, using the one or more processors, an input vector corresponding to the accumulated interval information;
transforming, using the one or more processors, the input vector by reducing a dimension of the input vector; and
estimating, using the one or more processors, the battery state information based on the transformed input vector and learning information determined in advance for reducing a time required to estimate the battery state information using the transformed input vector and the learning information and/or improving an accuracy to estimate the battery state information using the transformed input vector and the learning information.

2. The method of claim 1 , further comprising storing the learning information based on reference information of a battery.

3. The method of claim 2 , wherein the storing comprises generating a parameter corresponding to the learning information based on the reference information and sample interval information into which sample battery information is partitioned in the predetermined range.

4. The method of claim 3 , wherein the generating of the parameter comprises learning the parameter by inputting the reference information and the sample interval information to a learner.

5. The method of claim 4 , wherein the estimating comprises estimating the battery state information by inputting the accumulated interval information to the learner in which the parameter is learned.

6. The method of claim 1 , wherein the collecting comprises measuring in real time any one or any combination of a voltage signal, a current signal, and a temperature signal of a battery.

7. The method of claim 1 , wherein the collecting comprises determining, based on a current signal of the battery, whether the battery is being charged or discharged.

8. The method of claim 1 , wherein the partitioning comprises setting a uniform interval in the predetermined range for each of the interval information.

9. The method of claim 1 , wherein the partitioning further comprises setting a first interval in the predetermined range for an interval, among from the plurality of intervals, having a high data generation probability or a second interval in the predetermined range for an interval, among from the plurality of intervals, having a low data generation probability in the battery information, and
wherein the first interval is determined to be narrower than the second interval such that the first interval and the second interval correspond to non-uniform intervals.

10. A processor-implemented method, performed in a computing apparatus in a battery management system, of estimating battery state information of a battery, the method comprising:
accumulating, using one or more processors of the computing apparatus, a usage history of the battery for a plurality of intervals, wherein
the usage history is measured using plural sensors for the computing apparatus, and lengths of the plurality of intervals are variably set dependent on resolutions of the plural sensors;
generating, using the one or more processors, an input data set corresponding to the accumulated usage history;
transforming, using the one or more processors, the input data set by reducing a dimension of the input data set; and
estimating, using the one or more processors, the battery state information based on the transformed input data set and learning information determined in advance for reducing a time required to estimate the battery state information using the transformed input vector and the learning information and/or improving an accuracy to estimate the battery state information using the transformed input vector and the learning information.

11. The method of claim 10 , wherein the accumulating comprises:
partitioning the usage history into interval information in a predetermined range; and
accumulating the interval information and storing the usage history for each interval of the plurality of intervals.

12. The method of claim 10 , wherein the estimating comprises estimating the battery state information by inputting the accumulated usage history to a learner configured to learn the learning information.

13. The method of claim 10 , wherein the learning information comprises a parameter generated based on reference information of the battery and sample interval information into which sample battery information is partitioned in a predetermined range.

14. The method of claim 10 , wherein the accumulating comprises accumulating a per-interval usage history of any one or any combination of a voltage signal, a current signal, and a temperature signal of the battery.

15. The method of claim 10 , wherein the generated input data set corresponding to the accumulated usage history is a generated input vector corresponding to the usage history.

16. The method of claim 10 , wherein the accumulating comprises:
determining, by the one or more processors using the plural sensors and based on a current signal of the battery, whether the battery is being charged or discharged; and
separately accumulating, by the one or more processors using the plural sensors, a usage history corresponding to charging of the battery and a usage history corresponding to discharging of the battery.

17. An apparatus in a battery management system for estimating battery state information of a battery, the apparatus comprising:
a collector configured to measure, using plural sensors for the apparatus, respective battery information, associated with the battery, for a plurality of intervals, wherein lengths of the plurality of intervals are variably set dependent on resolutions of the plural sensors;
an accumulator configured to accumulate interval information, for each of the plurality of intervals, into which the measured battery information is partitioned in a predetermined range;
a generator configured to generate an input vector corresponding to the accumulated interval information;
a transformer configured to transform the input vector by reducing a dimension of the input vector; and
an estimator configured to estimate the battery state information based on the transformed the input vector and learning information determined in advance for reducing a time required to estimate the battery state information using the transformed input vector and the learning information and/or improving an accuracy to estimate the battery state information using the transformed input vector and the learning information.

18. The apparatus of claim 17 , wherein the estimator comprises a learner configured to output the battery state information using a parameter corresponding to the learning information in response to receiving the accumulated interval information.

19. The apparatus of claim 17 ,

Assignees

• Samsung Electronics Co Ltd

Inventors

• Park Sangdo
• You Kaeweon
• Lee Sun-Jae

Classifications

• H02J7/825

  Detection of fully charged condition · CPC title

• H02J7/84

  Control of state of health [SOH] · CPC title

• H02J7/82

  Control of state of charge [SOC] · CPC title

• H01M10/4257

  Smart batteries, e.g. electronic circuits inside the housing of the cells or batteries · CPC title

• B60L2240/549

  Current · CPC title