Predictive model for estimating battery states

Patent metadata

Field	Value
Publication number	US-2026091684-A1
Application number	US-202519295379-A
Country	US
Kind code	A1
Filing date	Aug 8, 2025
Priority date	Jul 21, 2017
Publication date	Apr 2, 2026
Grant date	—

Abstract

A battery management system (BMS) for a vehicle includes a module for estimating the state of a rechargeable battery, such as its state of charge, in real time. The module includes a learning model for predicting the state of a battery based on the vehicle's usage and related factors unique to the vehicle, in addition to a sensed voltage, current and temperature of a battery.

First claim

1 .- 21 . (canceled)

22 . A method, comprising:
providing a trained machine learning battery model; inputting, into the trained machine learning battery model, obtained data comprising a measured voltage, current, impedance, pressure, temperature of at least one series string of battery cells, or a combination thereof; inputting at least one battery type; using the trained machine learning battery model to predict a battery state; and producing an output based on the predicted battery state.

23 . The method of claim 22 , wherein the trained machine learning battery model was trained using training sets comprising values for each one of a series of input parameters and a corresponding value for the at least one target using at least the obtained data,
wherein the trained machine learning battery model includes additional input parameters, the additional input parameters comprising one or more of (a) the age of the battery, (b) a last battery, or power train servicing date, (c) a vehicle identification number (VIN), or make and model, (d) a driving pattern, (e) a geographic location or predicted weather forecast, (f) an age of a registered driver, (g) number of discharge cycles, or average percentage of battery discharges between charges using an external power source, (h) number of warning conditions issued on the battery.

24 . The method of claim 23 , wherein the trained battery model is trained on one or more sensor data and the additional parameters;
wherein the obtained data is obtained while a series string of battery cells is (a) in an open circuit condition, (b) under an operating load, (b) during a charging event, (c) during a diagnostic event, (d) during a repair event, or (e) during a safety event.

25 . The method of claim 22 , further comprising:
training the battery model using training sets comprising values for each one of the input parameters and a corresponding value for the at least one target using at least the obtained data; receiving new training sets for the battery type over a network; and re-training the battery model on the new training sets.

26 . The method of claim 22 , further comprising
using at least one or more sensors located on a vehicle, obtaining data on at least one physical state of two or more battery cells of the vehicle, wherein the data obtained comprises at least one of a measured voltage, current, impedance, pressure, or temperature of the two or more battery cells; wherein the vehicle has a processor and a non-transitory memory medium, the battery model residing in the non-transitory memory medium; using the vehicle processor to predict the battery state using the trained battery model and a new training set, and producing an output based on the predicted battery state.

27 . The method of claim of claim 26 , wherein the one or more sensors comprising
one or more sensors configured to measure pressure of the series string of battery cells, a module of batteries, and/or individual battery cells, a sensor configured to measure voltage of a series string of battery cells, a sensor configured to measure total charge passed through the series string of battery cells, and/or temperature sensors to measure ambient temperature and battery series temperature in at least one location.

28 . The method of claim 26 , wherein the output is displayed on a user interface (UI) such as a vehicle user interface (VUI) or a mobile device.

29 . The method of claim 28 , wherein the UI displays one or more of the following information in the output,
(a) repair battery when the predicted battery state predicts the SOH is below 50%, (b) warning message when the predicted battery state predicts an unsafe battery condition, (c) warning message when the predicted battery state predicts energy available is insufficient for a predicted event, and (d) warning message when the predicted battery state predicts power available insufficient for the predicted event.

30 . The method of claim 26 , wherein the output is sent to a vehicle control system, and wherein the vehicle control system performs one or more of the following tasks based on the predicted state of the battery:
when SOC is about 80%, allowing the battery to discharge at maximum operating power rate, limiting the power availability when the SOC is below 20% or below 15% or below 10%, notifying a driver when the predicted power availability is below 6 kW or 5 kW or 4 kW or 3 kW or 2 kW or 1 kW, and limiting power to auxiliary systems when the predicted power availability is below 6 kW or 5 kW or 4 kW or 3 kW or 2 kW or 1 kW.

31 . The method of claim 26 , wherein the vehicle has a processor configured for being coupled to a server located over one or more networks and the battery model is accessible through the server, the method further comprising:
uploading to the server the obtained data, and receiving the trained battery model from the server.

32 . The method of claim 26 , further comprising retraining the battery model on a continuous basis as the sensor data provides the prediction module with time varying values for the at least one of a measured voltage, current, impedance, pressure, or temperature.

Assignees

• Quantumscape Battery Inc

Inventors

• Holme Timothy

Classifications

• H01M50/204

  Racks, modules or packs for multiple batteries or multiple cells · CPC title

• H01M10/486

  for measuring temperature · CPC title

• G01R31/3648

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

• B60L58/12

  responding to state of charge [SoC] · CPC title

• B60L50/66

  Arrangements of batteries · CPC title