	US 12,241,938 B2
	Method for estimating the state of an energy store
Christoph Kroener, Rosstal (DE); Felix Kleinheinz, Blaufelden (DE); Marco Stroebel, Stuttgart (DE); and Peter Birke, Wallhausen (DE)
Assigned to Robert Bosch GmbH, Stuttgart (DE)
Appl. No. 17/781,080
Filed by Robert Bosch GmbH, Stuttgart (DE)
PCT Filed Nov. 24, 2020, PCT No. PCT/EP2020/083112 § 371(c)(1), (2) Date May 31, 2022, PCT Pub. No. WO2021/105071, PCT Pub. Date Jun. 3, 2021.
Claims priority of application No. 10 2019 218 591.6 (DE), filed on Nov. 29, 2019.
Prior Publication US 2023/0003805 A1, Jan. 5, 2023
Int. Cl. G01R 31/382 (2019.01); G01R 31/367 (2019.01); G01R 31/374 (2019.01); G01R 31/389 (2019.01)

CPC G01R 31/382 (2019.01) [G01R 31/367 (2019.01); G01R 31/374 (2019.01); G01R 31/389 (2019.01)]

19 Claims

1. A method for estimating the state of an energy store comprising at least one electrochemical battery cell by means of a battery management system (BMS), which comprises an impedance spectroscopy chip, said method comprising at least the following method steps:

a) determining the frequency-dependent impedance of the at least one electrochemical battery cell from a data set recorded in real time,

b) training an artificial neural network with temperature-dependent training spectra from the data set as input and receiving a preset for a temperature value belonging to each training spectrum as output,

c) determining weighting functions based on the input and the output,

d) testing the artificial neural network by receiving a battery cell-to-battery cell variance from the data set as test spectra and estimating the temperature values belonging to the test spectra in accordance with the weighting functions determined in method step c), and

e) estimating at least one internal state of the at least one electrochemical battery cell of the energy storage with the trained artificial neural network, and

f) disconnecting the at least one electrochemical cell when the estimated temperature value exceeds a thermal load threshold.