Battery state estimation apparatus and method

• US 10,310,022 B2
• Filed: 04/17/2017
• Issued: 06/04/2019
• Est. Priority Date: 10/05/2016
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Park, SangDo, You, Kaeweon
• Primary Examiner(s): Raymond, Edward
• Application Number:
  US15/488,588
• Publication Number:
  US 20180095140A1
• Time in Patent Office: 778 Days
• Field of Search: 702 63, 702 64, 702182-185
• CPC Class Codes: G01R 31/367 Software therefor, e.g. for battery testing using modelling or look-up tables
• CPC Class Codes: G01R 31/378 specially adapted for the type of battery or accumulator
• CPC Class Codes: G01R 31/392 Determining battery ageing or deterioration, e.g. state of health
• CPC Class Codes: Y02E 60/10 Energy storage using batteries

Abstract

Provided is a battery state estimation apparatus and method that determine a validity of a battery model, which is dependent on a parameter, based on state information of a battery that is estimated from battery information of the battery, transmit an update request for the battery model to an external battery model provider in response to a result of the determining indicating that the battery model is invalid, receive another parameter in response to the update request, and update the battery model based on the other parameter.

Claims

1. A method to estimate a state of a battery, the method comprising:

determining, by a processor, a validity of a battery model configured by a parameter, based on state information of the battery;

controlling, by the processor, a communicator to transmit an update request for the battery model to an external battery model provider, in response to a result of the determining indicating that the battery model is invalid;

receiving, by the processor, another parameter, in response to the update request; and

updating, by the processor, the battery model to be configured by the other parameter.

2. The method of claim 1, further comprising:

estimating, by the processor, the state information of the battery using the updated battery model and battery information.

3. The method of claim 1, wherein the determining comprises:

estimating the state information of the battery from battery information using the battery model;

determining whether the estimated state information of the battery is outside of an estimation range of the battery model;

determining that the battery model is valid in response to the estimated state information of the battery being within the estimation range; and

determining that the battery model is invalid in response to the estimated state information being outside of the estimation range.

4. The method of claim 3, wherein the determining of whether the estimated state information of the battery is outside of the estimation range comprises determining that the estimated state information of the battery is outside of the estimation range in response to the estimated state information of the battery being less than a predetermined minimum level of state information set to be accurately estimated by the battery model.

5. The method of claim 1, wherein the transmitting comprises transmitting, to a battery state estimation server that generates battery state models based on different parameters, a signal to request at least one parameter having a range overlapping at least a portion of an estimation range of the parameter that the battery model is dependent on.

6. The method of claim 1, wherein the determining comprises determining that the parameter that the battery model is dependent on is invalid in response to the estimated state information of the battery being outside of an estimation range of the battery model, and

the transmitting comprises transmitting, to a battery state estimation server that generates battery state models based on different parameters, a signal to request at least one parameter trained to estimate state information less than or equal to a predetermined minimum level of the estimation range of the parameter that the battery model is dependent on in response to the estimated state information of the battery being less than or equal to the predetermined minimum level.

7. The method of claim 6, wherein the receiving further comprises receiving, from the battery state estimation server, one or more parameters trained to a different predetermined minimum level, of a different estimation range predetermined for accurate estimation, based on all or a select portion of reference data collected by the battery state estimation server up to a point in time at which the update request is received.

8. The method of claim 1, further comprising:

searching, by the processor, for a parameter corresponding to a sensed change in the battery, from among parameters stored in a memory;

updating, by the processor, when the searched for parameter is found, the battery model based on the found parameter;

performing, by the processor, when the searched for parameter is not found, the transmitting, the receiving of the other parameter, and the updating of the battery model based on the other parameter; and

estimating, by the processor, state information of the battery with the sensed change based on the updated battery model.

9. The method of claim 1,

wherein the determining further includes searching for a parameter corresponding to a changed battery, among parameters stored in a memory, in response to a change of the battery being sensed, and

wherein the transmitting further includes transmitting a signal to request the parameter corresponding to the changed battery in response to the parameter corresponding to the changed battery being not found.

10. The method of claim 1, further comprising:

continuing, by the processor, an estimating of the state information of the battery for continued changes in battery information based on the parameter that the battery model is dependent on, until the result of the determining indicates that that the estimated state information is invalid.

11. The method of claim 1,

wherein the battery model is a battery model neural network and the parameter is a representation of specially trained connection weights within the battery model neural network trained for estimating a battery state for a first state information estimation range, and

wherein the updating of the battery model includes applying connection weights represented by the other parameter to a neural network structure to generate another specially trained battery model neural network trained for estimating a battery state for a different second state information estimation range, with the generated other specially trained battery model neural network being the updated battery model.

12. The method of claim 1, wherein the battery model is a battery model neural network and the parameter that the battery model is dependent on is a trained connection weighting matrix, and the other parameter is a different trained connection weighting matrix.

13. The method of claim 1, wherein the parameter and the other parameter are respectively trained parameters for different battery state estimation ranges.

14. A non-transitory computer-readable medium storing instructions, that when executed by the processor, cause the processor to perform the method of claim 1.

24. The method of claim 1, further comprising storing, by a memory, the state information of the battery.

25. The method of claim 24, wherein the communicator is configured to communicate with the external battery model provider.

26. The method of claim 25, wherein the battery model is used to estimate the state information of the battery.

27. The method of claim 1, wherein the parameter represents a collection of connection weights between nodes of a neural network of the battery model.

28. The method of claim 1, wherein the battery model is a machine learning (ML) model including any one or any combination of any two or more of a neural network, a hidden Markov model, a Beyesian network, a support vector machine (SVM), a decision tree (DT), and a nearest neighbors algorithm (k-NN).

15. An apparatus to estimate a state of a battery, the apparatus comprising:

a communicator to communicate with an external battery model provider;

a memory to store a battery state estimation model that is dependent on a parameter; and

a processor configured to determine a validity of the battery state estimation model based on state information of the battery that is estimated from battery information of the battery, control the communicator to transmit an update request for the battery state estimation model to the external battery model provider in response to a result of the determining indicating that the battery state estimation model is invalid, and update the battery state estimation model based on another parameter received in response to the update request.

16. The apparatus of claim 15, further comprising the battery.

17. The apparatus of claim 15, wherein the processor is configured to estimate the state information of the battery from the battery information using the battery state estimation model, determine whether the estimated state information of the battery is outside of an estimation range of the battery state estimation model, determine that the battery state estimation model is valid in response to the estimated state information of the battery being within the estimation range, and determine that the battery state estimation model is invalid in response to the estimated state information being outside of the estimation range.

18. The apparatus of claim 17, wherein the processor is configured to determine that the estimated state information of the battery is outside of the estimation range in response to the estimated state information of the battery being less than a predetermined minimum level of state information set to be accurately estimated by the battery state estimation model.

19. The apparatus of claim 15, wherein the processor is configured to control the communicator to transmit, to a battery state estimation server that generates battery state estimation models based on different parameters, a signal to request at least one parameter having a range overlapping at least a portion of an estimation range of the parameter that the battery state estimation model is dependent on.

20. The apparatus of claim 15, wherein the processor is configured to determine that the parameter that the battery state estimation model is dependent on is invalid in response to the estimated state information of the battery being outside of an estimation range of the battery state estimation model, and control the communicator to transmit, to a battery state estimation server that generates battery state estimation models based on different parameters, a signal to request at least one parameter trained to estimate state information less than or equal to a predetermined minimum level of the estimation range of the parameter that the battery state estimation model is dependent on in response to the estimated state information of the battery being less than or equal to the predetermined minimum level.

21. The apparatus of claim 15, wherein the parameter and the other parameter are respectively trained parameters for different battery state estimation ranges.

22. An apparatus to estimate a state of a battery, the apparatus comprising:

a processor configured to

determine a validity of a battery state estimation model neural network, which is dependent on first trained connection weighting information, based on state information of a battery that is estimated from battery information of the battery,

control a transmitting of an update request for the battery state estimation model neural network to an external battery model provider in response to the determining indicating that the battery state estimation model neural network is invalid, and

update the battery state estimation model neural network based on second different trained connection weighting information received in response to the update request.

23. The apparatus of claim 22, wherein the first trained connection weighting information is trained connection weighting information for a battery state estimation range different from a battery state estimation range for which the second trained connection weighting information is trained.