Vehicle battery analysis system

Patent metadata

Field	Value
Publication number	US-11555859-B2
Application number	US-202017017016-A
Country	US
Kind code	B2
Filing date	Sep 10, 2020
Priority date	Sep 10, 2020
Publication date	Jan 17, 2023
Grant date	Jan 17, 2023

Abstract

In one embodiment, a vehicle battery diagnostics system forecasts a future state for a battery by monitoring, over a period of time, one or more of voltage, current or temperature signals from at least one battery of the vehicle, storing information from the voltage, current or temperature signals as time-series data, obtaining a forecasting model from a server, the forecasting model indicating at least one shapelet feature that corresponds to a forecast categorization, identifying, in the time-series data, a shapelet that matches the at least one shapelet feature to a degree exceeding a predetermined similarity threshold, and providing a notification indicating the forecast categorization.

First claim

What is claimed is:

1. A vehicle battery diagnostics system for a vehicle, comprising:
a communication device configured to wirelessly transmit data to a server and receive data from the server;
one or more processors; and
a memory communicably coupled to the one or more processors and storing:
a monitoring module including instructions that when executed by the one or more processors cause the one or more processors to monitor, over a period of time, one or more of voltage, current or temperature signals from at least one battery of the vehicle and store information from the voltage, current or temperature signals as time-series data;
a forecast module including instructions that when executed by the one or more processors cause the one or more processors to obtain a forecast model from the server, the forecast model indicating at least one shapelet feature that corresponds to a forecast categorization, the forecast categorization indicating a suitability of the battery for a second life application, the second life application being as an energy source for a device different from the vehicle, wherein the forecast model is trained to classify the battery using a drive-cycle dataset associated with failure of a similar battery under a plurality of conditions, the plurality of conditions determined by a closed-loop iterative process that adjusts condition profiles to diversify extracted shapelets, the shapelets being associated with battery cells that exhibit one or more predetermined life-cycle capacity retention traits; and
a classifier module including instructions that when executed by the one or more processors cause the one or more processors to identify, in the time-series data, a shapelet that matches the at least one shapelet feature to a degree exceeding a predetermined similarity threshold, and to provide a notification indicating the forecast categorization.

2. The vehicle battery diagnostics system of claim 1 , wherein the forecast categorization further indicates one or more of: an anomalous behavior of the battery or a projection of an amount of performance capacity the battery would retain at a future time T.

3. The vehicle battery diagnostics system of claim 1 , wherein the at least one shapelet feature is derived from a voltage or current signal, the forecast model configured to predict a temperature of the battery from the at least one shapelet feature, and the classifier module further includes instructions to:
compare a present temperature of the battery detected when the shapelet that matches the at least one shapelet feature is identified; and
determine that the battery exhibits anomalous behavior when the present temperature differs from the temperature that was predicted by a threshold amount.

4. The vehicle battery diagnostics system of claim 1 , wherein the forecast model is trained to classify the battery based on training data obtained from emulating driving conditions for a plurality of batteries identical or similar to the at least one battery of the vehicle.

5. The vehicle battery diagnostics system of claim 4 , wherein the forecast model is trained by:
creating the drive-cycle dataset, the drive-cycle dataset being of one or more of voltage, current or temperature values by using cycler equipment to cycle the plurality of batteries to failure under the plurality of conditions;
extracting, from the drive-cycle dataset, the shapelets;
enhancing information derived from the plurality of shapelets to create a plurality of features; and
training the forecast model based on the plurality of features to forecast, based on input time-series data, whether the battery will retain a threshold percentage of capacity at a future time T.

6. The vehicle battery diagnostics system of claim 5 , wherein the plurality of conditions are determined by a drive-cycle profile generator that generates the condition profiles in the closed-loop iterative process that receives the extracted shapelets as input and adjusts the condition profiles to diversify the extracted shapelets.

7. The vehicle battery diagnostics system of claim 5 , wherein the plurality of features are created by one or more of:
identifying a minimal distance of shapelets to a given segment of a timeseries;
identifying correlative features of shapelets across different signal types; or
identifying shapelet co-occurrence in a time segment of the time-series data.

8. A method of diagnosing a battery for a vehicle, comprising:
monitoring, over a period of time, one or more of voltage, current or temperature signals from the battery of the vehicle;
storing information from the voltage, current or temperature signals as time-series data;
obtaining a forecast model from a server, the forecast model indicating at least one shapelet feature that corresponds to a forecast categorization, the forecast categorization indicating a suitability of the battery for a second life application, the second life application being as an energy source for a device different from the vehicle, wherein the forecast model is trained to classify the battery using a drive-cycle dataset associated with failure of a similar battery under a plurality of conditions, the plurality of conditions determined by a closed-loop iterative process that adjusts condition profiles to diversify extracted shapelets, the shapelets being associated with battery cells that exhibit one or more predetermined life-cycle capacity retention traits;
identifying, in the time-series data, a shapelet that matches the at least one shapelet feature to a degree exceeding a predetermined similarity threshold; and
providing a notification indicating the forecast categorization.

9. The method of claim 8 , wherein the forecast categorization further indicates one or more of: an anomalous behavior of the battery or a projection of an amount of performance capacity the battery would retain at a future time T.

10. The method of claim 8 , wherein the at least one shapelet feature is derived from a voltage or current signal and the forecast model predicts a temperature of the battery from the at least one shapelet feature, the method further comprising:
comparing a present temperature of the battery detected when the shapelet that matches the at least one shapelet feature is identified; and
determining that the battery exhibits anomalous behavior when the present temperature differs from the temperature that was predicted by a threshold amount.

11. The method of claim 8 , wherein the forecast model is trained to classify the battery based on training data obtained from emulating driving conditions for a plurality of batteries identical or similar to the battery of the vehicle.

12. The method of claim 11 , wherein the forecast model is trained by:
creating the drive-cycle dataset, the drive-cycle dataset being of one or more of voltage, current or temperature values by using cycler equipment to cycle the plurality of batteries to failure under the plurality of conditions;
extracting, from the drive-cycle dataset, the shapelets;
enhancing information derived from the plurality of shapelets to create a plurality of features; and
training the forecast model based on the plurality of features to forecast, based on input time-series data, whether the battery will retain a threshold percentage of capacity at a future time T.

13. The method of claim 12 , wherein the plurality of conditions are determined by a drive-cycle profile generator that generates the condition profiles in the closed-loop iterative process that receives the extracted shapelets as input and adjusts the condition profiles to diversify the extracted shapelets.

Assignees

• Toyota Res Inst Inc

Inventors

• Aykol Muratahan
• Balaji Gopal Chirranjeevi
• Herring Patrick K
• Anapolsky Abraham S

Classifications

• G06F2119/06

  Power analysis or power optimisation · CPC title

• H01M10/48

  Accumulators combined with arrangements for measuring, testing or indicating the condition of cells, e.g. the level or density of the electrolyte (constructional details of current conducting connections for detecting conditions inside cells or batteries, e.g. details of voltage sensing terminals, H01M50/569) · CPC title

• Y02E60/10

  Energy storage using batteries · CPC title

• B60L2240/545

  Temperature · CPC title

• G01R31/374

  with means for correcting the measurement for temperature or ageing · CPC title