Data-driven Model for Lithium-ion Battery Capacity Fade and Lifetime Prediction

Patent metadata

Field	Value
Publication number	US-2019113577-A1
Application number	US-201816161852-A
Country	US
Kind code	A1
Filing date	Oct 16, 2018
Priority date	Oct 17, 2017
Publication date	Apr 18, 2019
Grant date	—

Abstract

A method of using data-driven predictive modeling to predict and classify battery cells by lifetime is provided that includes collecting a training dataset by cycling battery cells between a voltage V1 and a voltage V2, continuously measuring battery cell voltage, current, can temperature, and internal resistance during cycling, generating a discharge voltage curve for each cell that is dependent on a discharge capacity for a given cycle, calculating, using data from the discharge voltage curve, a cycle-to-cycle evolution of cell charge to output a cell voltage versus charge curve Q(V), generating transformations of ΔQ(V), generating transformations of data streams that include capacity, temperature and internal resistance, applying a machine learning model to determine a combination of a subset of the transformations to predict cell operation characteristics, and applying the machine learning model to output the predicted battery operation characteristics.

First claim

What is claimed:

1 ) A method of using data-driven predictive modeling to predict battery cells by lifetime, comprising:
a) collecting a training dataset by cycling, using a battery cycling instrument, a plurality of battery cells between a voltage V1 and a voltage V2; b) continuously measuring battery cell physical properties comprising a battery cell voltage, a battery cell current, a battery cell can temperature, and a battery cell internal resistance, or a battery cell internal resistance of each said battery cell during said cycling; c) generating, using an algorithm on a non-transitory computer medium, a voltage curve for each said battery cell, wherein said voltage curve is dependent on a capacity for a given said cycle; d) calculating, using data from said voltage curves, a cycle-to-cycle evolution of a battery cell charge to output a cell voltage versus charge curve Q(V), wherein said output of cell voltage versus charge Q(V) is ΔQ(V); e) generating transformations of said ΔQ(V); and f) applying said machine learning model to output said predicted battery operation characteristics of said cycled plurality of battery cells, or additional battery cells operated at a later date.

2 ) The method according to claim 1 , wherein said battery cell physical properties are selected form the group consisting of a battery cell voltage, a battery cell current, a battery cell can temperature, and a battery cell internal resistance.

3 ) The method according to claim 1 , wherein said continuous measurement further comprising an electrochemical impedance, using spectroscopy, and strain, using a strain gauge.

4 ) The method according to claim 1 , wherein said transformation of said of a ΔQ(V) comprise a value at said V1, and a value at said V2, or between said V1 and said V2.

5 ) The method according to claim 1 , wherein said battery cell operation characteristics are selected from the group consisting of a battery lifetime, a logarithm of said battery lifetime, and a Boolean classification of battery performance, wherein said battery cycle life comprises a lifetime, energy, or power.

6 ) The method according to claim 1 , wherein said output battery cell operation characteristics are selected from the group consisting of a lifetime output, a logarithm of predicted cycle life output, and a predicted classification of battery performance output, wherein said battery lifetime comprises a cycle life, calendar life, energy, or power.

7 ) The method according to claim 1 further comprising generating, using said algorithm, transformations of data streams comprising capacity, temperature and internal resistance, or internal resistance after transformations of said ΔQ(V) are generated.

8 ) The method according to claim 1 further comprising applying a machine learning model, using said algorithm, to determine a combination of a subset of said transformations to predict battery cell operation characteristics after transformations of said ΔQ(V) are generated.

Assignees

• Univ Leland Stanford Junior
• Massachusetts Inst Technology

Inventors

• Severson Kristen Ann
• Braatz Richard Dean
• Chueh William C
• Attia Peter M
• Jin Norman
• Harris Stephen J
• Perkins Nicholas

Classifications

• G01R31/3648

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

• G01R31/3842 Primary

  combining voltage and current measurements · CPC title

• H01M10/0525

  Rocking-chair batteries, i.e. batteries with lithium insertion or intercalation in both electrodes; Lithium-ion batteries · CPC title

• H01M10/44

  Methods for charging or discharging (circuits for charging H02J7/00) · 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