Method and system for estimating battery pack balance state of new energy vehicle

Patent metadata

Field	Value
Publication number	US-11513164-B2
Application number	US-202017082051-A
Country	US
Kind code	B2
Filing date	Oct 28, 2020
Priority date	Oct 28, 2020
Publication date	Nov 29, 2022
Grant date	Nov 29, 2022

Abstract

A method and system for estimating a battery pack balance state of a new energy vehicle are provided. The method includes: acquiring an actual cell voltage difference measured for a battery pack of a new energy vehicle, and acquiring one or more operation condition parameters for an operation condition under which the actual cell voltage difference is measured, wherein the actual cell voltage difference is the difference between the maximum and minimum battery cell voltages within the battery pack; determining an expected cell voltage difference according to a model based on the one or more operation condition parameters; determining a compensated cell voltage difference according to the actual cell voltage difference and the expected cell voltage difference; and estimating a battery pack balance state of the new energy vehicle according to the compensated cell voltage difference.

First claim

What is claimed is:

1. A method for estimating a battery pack balance state of a new energy vehicle, comprising:
acquiring an actual cell voltage difference measured for a battery pack of a new energy vehicle, and acquiring one or more operation condition parameters for an operation condition under which the actual cell voltage difference is measured, wherein the actual cell voltage difference is the difference between the maximum and minimum battery cell voltages within the battery pack;
determining an expected cell voltage difference according to the one or more operation condition parameters;
determining a compensated cell voltage difference according to the actual cell voltage difference and the expected cell voltage difference; and
estimating a battery pack balance state of the new energy vehicle according to the compensated cell voltage difference.

2. The method as claimed in claim 1 , wherein the one or more operation condition parameters for the operation condition comprise at least one of:
a vehicle current, a state-of-charge, a battery temperature, a vehicle speed, or an odometer value.

3. The method as claimed in claim 1 , wherein determining the expected cell voltage difference according to the one or more operation condition parameters comprises:
determining, by using a machine learning model, the expected cell voltage difference according to the one or more operation condition parameters.

4. The method as claimed in claim 3 , wherein the machine learning model is a linear model, and determining, by using the machine learning model, the expected cell voltage difference according to the one or more operation condition parameters comprises:
determining the expected cell voltage difference δ ref based on the following formula:
δ ref =w 0 +w 1 *P 1 + . . . +w M *P M ,

where w 0 , w 1 , . . . , w M are weights of the linear model, P 1 , . . . , P M are the one or more operation condition parameters, and M is the number of the one or more operation condition parameters.

5. The method as claimed in claim 4 , wherein the weights w 0 , w 1 , . . . , w M of the linear model are obtained through training based on first training data, wherein the first training data comprises one or more operation condition parameters for multiple different operation conditions, and actual cell voltage differences measured for battery packs of multiple new energy vehicles under the multiple different operation conditions.

6. The method as claimed in claim 5 , wherein before the weights w 0 , w 1 , . . . , w M of the linear model are obtained through training based on the first training data, actual cell voltage differences measured for battery packs of new energy vehicles satisfying a preset condition are removed from the first training data, wherein the preset condition comprises: new energy vehicles with abnormal fleet performance, new energy vehicles with fleet performance at top or bottom percentiles among the multiple new energy vehicles.

7. The method as claimed in claim 3 , wherein the machine learning model is a nonlinear model, and determining, by using the machine learning model, the expected cell voltage difference according to the one or more operation condition parameters comprises:
determining, based on the nonlinear model by using the one or more operation condition parameters as an input, the expected cell voltage difference δ ref , wherein the nonlinear model comprises a neural network-based model or a random forest-based model.

8. The method as claimed in claim 7 , wherein weights in the nonlinear model are obtained through training based on first training data, wherein the first training data comprises one or more operation condition parameters for multiple different operation conditions, and actual cell voltage differences measured for battery packs of multiple new energy vehicles under the multiple different operation conditions.

9. The method as claimed in claim 8 , wherein before the weights of the nonlinear model are obtained through training based on the first training data, actual cell voltage differences measured for battery packs of new energy vehicles satisfying a preset condition are removed from the first training data, wherein the preset condition comprises: new energy vehicles with abnormal fleet performance, new energy vehicles with fleet performance at top or bottom percentiles among the multiple new energy vehicles.

10. The method as claimed in claim 1 , wherein determining the compensated cell voltage difference according to the actual cell voltage difference and the expected cell voltage difference comprises:
determining the compensated cell voltage difference Δ based on the following formula:
Δ=δ act −δ ref ,

where δ act is the actual cell voltage difference, and δ ref is the expected cell voltage difference.

11. The method as claimed in claim 1 , wherein the actual cell voltage difference measured for the battery pack of the new energy vehicle comprises measured actual cell voltage difference for each time record; and
determining the compensated cell voltage difference according to the actual cell voltage difference and the expected cell voltage difference comprises:
determining the compensated cell voltage difference for each time record, and aggregating the compensated cell voltage difference for each time record within a preset time unit to obtain mean compensated cell voltage difference;
or,
determining the compensated cell voltage difference for each time record, and aggregating the compensated cell voltage difference for each time record by trip-on-battery period basis to obtain mean compensated cell voltage difference;
or,
determining the compensated cell voltage difference for each time record, and aggregating the compensated cell voltage difference for each time record by charging period basis to obtain mean compensated cell voltage difference.

12. The method as claimed in claim 11 , wherein estimating the battery pack balance state of the new energy vehicle according to the compensated cell voltage difference comprises:
in a case where the mean compensated cell voltage difference is above a first threshold or mean compensated cell voltage differences in multiple preset time units or multiple trip-on-battery periods or multiple charging periods show a degradation trend, determining that the battery pack of the new energy vehicle is in an unbalanced state.

13. The method as claimed in claim 1 , wherein
before estimating the battery pack balance state of the new energy vehicle according to the compensated cell voltage difference, the method further comprises: determining normalized weighted compensated cell voltage difference within a preset period; and
estimating the battery pack balance state of the new energy vehicle according to the compensated cell voltage difference comprises: estimating the battery pack balance state of the new energy vehicle according to the normalized weighted compensated cell voltage difference within the preset period.

14. The method as claimed in claim 13 , wherein determining the normalized weighted compensated cell voltage difference within a preset period comprises:
determining normalized weighted compensated cell voltage difference based on the following formula:

Δ
norm

=

∑

n
=
1

Assignees

• Guangzhou Automobile Group Co

Inventors

• Zhang Yizhen
• Shang Jin
• Zhang Xiong
• Xu Yonggang

Classifications

• H02J7/82 Primary

  Control of state of charge [SOC] · CPC title

• H02J7/50

  acting upon multiple batteries simultaneously or sequentially · CPC title

• B60L58/22

  Balancing the charge of battery modules · CPC title

• B60L2240/545

  Temperature · CPC title

• B60L2260/46

  by self learning · CPC title