Method and system for key predictors and machine learning for configuring cell performance
A method for key predictors and machine learning for configuring battery cell performance may include providing a cell that may include a cathode, a separator, and a silicon-dominant anode; measuring a plurality of parameters of the cell; and using a machine learning model to determine cell performance based on the plurality of measured parameters. The plurality of parameters may include initial coulombic efficiency and/or second cycle coulombic efficiency. Cells may be classified based on the determined cell performance and similarly performing cells may be binned together. A battery pack may be provided with a plurality of cells. The plurality of cells may be assessed during cycling using the machine learning model. One or more of the plurality of cells may be replaced when the assessing determines a different performance of the one or more of the plurality of cells. The battery pack may be in an electric vehicle.
1. A method of predicting battery performance, the method comprising:
providing a cell comprising a cathode, a separator, and a silicon-dominant anode;
measuring, via a measurement apparatus configured for obtaining particular types of measurements relating to cells and operations of cells, a plurality of parameters of the cell;
using a machine learning model to determine cell performance based on the plurality of measured parameters; and
measuring a third subset of the plurality of parameters during cycling of the cell;
wherein the plurality of parameters comprise a comparison of two different voltage curves of the cell within a first ten cycles, wherein the voltage curves comprise voltage vs. capacity data points that are acquired as the cell is cycled.
2. The method of claim 1 , wherein the plurality of parameters comprise initial coulombic efficiency.
3. The method of claim 1 , wherein the plurality of measured parameters comprises second cycle coulombic efficiency.
4. The method according to claim 1 , comprising classifying cells based on the determined cell performance and binning similarly performing cells.
5. The method according to claim 1 , comprising providing a battery pack with a plurality of cells.
6. The method according to claim 5 , comprising assessing the plurality of cells during cycling using the machine learning model.
7. The method according to claim 6 , comprising replacing one or more of the plurality of cells from the battery pack when the assessing determines a different performance of the one or more of the plurality of cells.
8. The method according to claim 7 , wherein the battery pack is in an electric vehicle.
9. The method of claim 1 , wherein the plurality of parameters comprise cell impedance values.
10. The method of claim 1 , wherein the plurality of parameters comprise open-circuit voltage.
11. The method of claim 1 , wherein the plurality of parameters comprise cell thickness.
12. The method of claim 1 , wherein the plurality of parameters comprise impedance after degassing.
13. The method of claim 1 , comprising measuring a first subset of the plurality of parameters before a formation process.
14. The method of claim 1 , wherein the plurality of parameters comprise a charge capacity of the cell within a first ten cycles.
15. The method of claim 1 , wherein the plurality of parameters comprise a discharge capacity of the cell within a first ten cycles.
16. The method according to claim 1 , wherein the determining of cell performance comprises determining cycle life defined as a number of cycles to reach 60-80% of initial capacity.