Intelligent battery cycling for lifetime longevity

Patent metadata

Field	Value
Publication number	US-10958082-B2
Application number	US-201815962148-A
Country	US
Kind code	B2
Filing date	Apr 25, 2018
Priority date	Apr 25, 2018
Publication date	Mar 23, 2021
Grant date	Mar 23, 2021

Abstract

A system for intelligent cycling of a battery for improved lifetime longevity includes a charge interrupt predictor executable to predict a likelihood of a power source disconnect event interrupting a current flow to the battery within one or more segments of a future time interval. The system also includes a charge cycle model executable to model a charge cycle for the battery within the future time interval based on the charge interrupt prediction. The system further includes a charge cycling controller that controls battery circuitry to charge or discharge the battery in accord with the charge cycle model.

First claim

What is claimed is:

1. A system for extending longevity of a battery comprising:
a memory device including memory;
a charge interrupt predictor stored in the memory and executable by a processor to make a charge interrupt prediction, the charge interrupt prediction predicting an interrupt time within a future time interval that a power source disconnect event is most likely to occur, the power source disconnect event interrupting a current flow to the battery;
a charge cycle modeler stored in the memory and executable by the processor to model a charge cycle for the battery within the future time interval, the modeled charge cycle including a charge level reduction to a first target zone prior to a charge level increase to a second target zone such that a charge level of the battery is within the second target zone at the predicted interrupt time, the first target zone and the second target zone being mutually exclusive and the second target zone being defined to include a range of approximately 90 to 100 percent charge; and
a charge cycling controller stored in the memory and executable by the processor to control battery circuitry to charge or discharge the battery in accord with the modeled charge cycle.

2. The system of claim 1 , wherein the charge cycle modeler implements logic to attempt to maximize a total time that a charge level of the battery is in the first target zone while ensuring that the charge level is within the second target zone each time the battery is actually disconnected from an external power source.

3. The system of claim 1 , wherein the charge interrupt predictor is configured to predict a specific time of day of the power source disconnect event.

4. The system of claim 1 , wherein the charge interrupt prediction is based on at least one of device charge history, calendar data, and location data.

5. The system of claim 1 , further comprising:
a charge/discharge rate predictor stored in memory and executable by the processor to dynamically update predicted charge and discharge rates for the battery based on detected changes in system power parameters, wherein the charge cycle modeler updates the modeled charge cycle responsive to each update to the predicted charge and discharge rates.

6. The system of claim 1 , wherein the system further comprises a feedback provider that includes a reinforcement learning algorithm, the feedback provider executable by the processor to:
identify discrepancies between an ideal charge cycle and an actual charge cycle observed; and
provide feedback to the charge cycle model that is based on the identified discrepancies.

7. The system of claim 6 , wherein the feedback provider is configured to provide feedback based on a discrepancy between a target time in the first target zone and an actual time in the first target zone.

8. The system of claim 6 , wherein the feedback provider is configured to provide feedback based on a discrepancy between the second target zone and an actual charge level of the battery at the time of the power source disconnect event.

9. The system of claim 6 , wherein the feedback provider is configured to provide feedback based on a number of fluctuations in charge level of the battery into and out of the second target zone throughout the actual charge cycle observed.

10. A method for extending longevity of a battery, the method comprising:
determining-a charge interrupt prediction, the charge interrupt prediction predicting an interrupt time within a future time interval that a power source disconnect event is most likely to occur, the power source disconnect event interrupting a current flow to the battery;
determining a charge cycle model for the battery within the future time interval, the charge cycle model including a charge level reduction to a first target zone prior to a charge level increase to a second target zone such that the battery is within the second target zone at the predicted interrupt time, the first target zone and the second target zone being mutually exclusive and the second target zone being defined to include a range of approximately 90 to 100 percent charge; and
controlling battery circuitry to charge or discharge the battery in accord with the charge cycle model.

11. The method of claim 10 , wherein the charge interrupt prediction is based on at least one of device charge history, calendar data, and location data.

12. The method of claim 10 , further comprising:
rendering a new charge interrupt prediction responsive to each detected power source disconnect event.

13. The method of claim 10 , further comprising:
dynamically updating predicted charge and discharge rates for the battery based on detected changes to system power parameters; and
updating the charge cycle model responsive to each change in the predicted charge and discharge rates.

14. The method of claim 10 , wherein determining the charge cycle model is based on modeling logic that attempts to maximize a total time that a charge level of the battery is in the first target zone while ensuring that the charge level is within the second target zone each time the battery is actually disconnected from an external power source.

15. The method of claim 14 , wherein the method comprises:
identifying discrepancies between an ideal charge cycle and the actual charge cycle observed; and
providing feedback to the modeling logic that is based on the identified discrepancies.

16. The method of claim 14 , wherein the modeling logic implements logic to maximize total time that charge level of the battery is in the first target zone while mitigating a number of charge cycles incurred during the future time interval.

17. One or more tangible processor-readable storage media encodes instructions for executing a computer process with one or more processors for extending longevity of a battery, the computer process comprising:
determining-a charge interrupt prediction, the charge interrupt prediction predicting an interrupt time within a future time interval that a power source disconnect event is most likely to occur, the power source disconnect event interrupting a current flow to the battery;
determining a charge cycle model for the battery within the future time interval, the charge cycle model including a charge level reduction to a first target zone prior to a charge level increase to a second target zone such that a charge level of the battery is within the second target zone at the predicted interrupt time, the first target zone and the second target zone being mutually exclusive and the second target zone being defined to include a range of approximately 90 to 100 percent charge; and
controlling battery circuitry to charge or discharge the battery in accord with the charge cycle model.

18. The one or more tangible processor-readable storage media of claim 17 , wherein the computer process further comprises:
rendering a new charge interrupt prediction responsive to each detected power source disconnect event.

19. The one or more tangible processor-readable storage media of claim 17 , wherein determining the charge cycle model is based on modeling logic that attempts to maximize a total time that a charge level of the battery is in the first target zone while ensuring that the charge level is within the second target zone each time the battery is actually disconnected from an external power source.

20. The one or more tangible processor-readable storage media of claim 19 , wherein the computer process further comprises:
identifying discrepancies

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Solomon Oz

Classifications

• H02J7/971

  the charge cycle being controlled or terminated in response to non-electric parameters · CPC title

• H02J7/933

  the cycle being controlled or terminated in response to electric parameters · CPC title

• H02J7/92

  with prioritisation of loads or sources · CPC title

• H02J7/44

  between battery management systems and power sources · CPC title

• H02J7/875

  Charging or discharging for charge maintenance, battery initiation or rejuvenation · CPC title