Genetic parameter identification of the Doyle-Fuller-Newman model from experimental cycling of a LiFePO4 battery

Joel C. Forman, Scott J. Moura, Jeffrey L. Stein, Hosam Kadry Fathy

Research output: Chapter in Book/Report/Conference proceedingConference contribution

24 Citations (Scopus)

Abstract

This paper examines the identification of the parameters of the Doyle-Fuller-Newman electrochemistry-based Lithium-ion battery model from voltage and current cycling data. The battery used in this study has a lithium iron phosphate cathode chemistry intended for high-power applications such as plug-in hybrid electric vehicles. The variables optimized for model identification include parameterizations of the model's anode equilibrium potential, cathode equilibrium potential, and solution conductivity. A genetic algorithm is used to optimize these model parameters against experimental data. The resulting identified model fits two experimental data sets used for system identification, as well as separate validation data sets corresponding to five different vehicle drive cycles. These drive cycles simulate the current a battery would undergo while used in a plug-in hybrid vehicle battery pack. The accuracy of the parameters is investigated using various validation data sets. This is believed to be the first attempt at fitting nearly all of the parameters and functions in the DFN model simultaneously using only voltage and current data. Computational logistics of using a genetic algorithm to identify 88 parameters of an electrochemistry-based model for 7.5 hours of cycling data are discussed. In addition, a detailed analysis of local parameter identifiability is presented.

Original languageEnglish (US)
Title of host publicationProceedings of the 2011 American Control Conference, ACC 2011
Pages362-369
Number of pages8
StatePublished - Sep 29 2011
Event2011 American Control Conference, ACC 2011 - San Francisco, CA, United States
Duration: Jun 29 2011Jul 1 2011

Publication series

NameProceedings of the American Control Conference
ISSN (Print)0743-1619

Other

Other2011 American Control Conference, ACC 2011
CountryUnited States
CitySan Francisco, CA
Period6/29/117/1/11

Fingerprint

Identification (control systems)
Plug-in hybrid vehicles
Electrochemistry
Cathodes
Genetic algorithms
Electric potential
Parameterization
Logistics
Anodes
Phosphates
Lithium
Iron

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering

Cite this

Forman, J. C., Moura, S. J., Stein, J. L., & Fathy, H. K. (2011). Genetic parameter identification of the Doyle-Fuller-Newman model from experimental cycling of a LiFePO4 battery. In Proceedings of the 2011 American Control Conference, ACC 2011 (pp. 362-369). [5991183] (Proceedings of the American Control Conference).
Forman, Joel C. ; Moura, Scott J. ; Stein, Jeffrey L. ; Fathy, Hosam Kadry. / Genetic parameter identification of the Doyle-Fuller-Newman model from experimental cycling of a LiFePO4 battery. Proceedings of the 2011 American Control Conference, ACC 2011. 2011. pp. 362-369 (Proceedings of the American Control Conference).
@inproceedings{c232ed0a43b840f4962ad972bf97719e,
title = "Genetic parameter identification of the Doyle-Fuller-Newman model from experimental cycling of a LiFePO4 battery",
abstract = "This paper examines the identification of the parameters of the Doyle-Fuller-Newman electrochemistry-based Lithium-ion battery model from voltage and current cycling data. The battery used in this study has a lithium iron phosphate cathode chemistry intended for high-power applications such as plug-in hybrid electric vehicles. The variables optimized for model identification include parameterizations of the model's anode equilibrium potential, cathode equilibrium potential, and solution conductivity. A genetic algorithm is used to optimize these model parameters against experimental data. The resulting identified model fits two experimental data sets used for system identification, as well as separate validation data sets corresponding to five different vehicle drive cycles. These drive cycles simulate the current a battery would undergo while used in a plug-in hybrid vehicle battery pack. The accuracy of the parameters is investigated using various validation data sets. This is believed to be the first attempt at fitting nearly all of the parameters and functions in the DFN model simultaneously using only voltage and current data. Computational logistics of using a genetic algorithm to identify 88 parameters of an electrochemistry-based model for 7.5 hours of cycling data are discussed. In addition, a detailed analysis of local parameter identifiability is presented.",
author = "Forman, {Joel C.} and Moura, {Scott J.} and Stein, {Jeffrey L.} and Fathy, {Hosam Kadry}",
year = "2011",
month = "9",
day = "29",
language = "English (US)",
isbn = "9781457700804",
series = "Proceedings of the American Control Conference",
pages = "362--369",
booktitle = "Proceedings of the 2011 American Control Conference, ACC 2011",

}

Forman, JC, Moura, SJ, Stein, JL & Fathy, HK 2011, Genetic parameter identification of the Doyle-Fuller-Newman model from experimental cycling of a LiFePO4 battery. in Proceedings of the 2011 American Control Conference, ACC 2011., 5991183, Proceedings of the American Control Conference, pp. 362-369, 2011 American Control Conference, ACC 2011, San Francisco, CA, United States, 6/29/11.

Genetic parameter identification of the Doyle-Fuller-Newman model from experimental cycling of a LiFePO4 battery. / Forman, Joel C.; Moura, Scott J.; Stein, Jeffrey L.; Fathy, Hosam Kadry.

Proceedings of the 2011 American Control Conference, ACC 2011. 2011. p. 362-369 5991183 (Proceedings of the American Control Conference).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Genetic parameter identification of the Doyle-Fuller-Newman model from experimental cycling of a LiFePO4 battery

AU - Forman, Joel C.

AU - Moura, Scott J.

AU - Stein, Jeffrey L.

AU - Fathy, Hosam Kadry

PY - 2011/9/29

Y1 - 2011/9/29

N2 - This paper examines the identification of the parameters of the Doyle-Fuller-Newman electrochemistry-based Lithium-ion battery model from voltage and current cycling data. The battery used in this study has a lithium iron phosphate cathode chemistry intended for high-power applications such as plug-in hybrid electric vehicles. The variables optimized for model identification include parameterizations of the model's anode equilibrium potential, cathode equilibrium potential, and solution conductivity. A genetic algorithm is used to optimize these model parameters against experimental data. The resulting identified model fits two experimental data sets used for system identification, as well as separate validation data sets corresponding to five different vehicle drive cycles. These drive cycles simulate the current a battery would undergo while used in a plug-in hybrid vehicle battery pack. The accuracy of the parameters is investigated using various validation data sets. This is believed to be the first attempt at fitting nearly all of the parameters and functions in the DFN model simultaneously using only voltage and current data. Computational logistics of using a genetic algorithm to identify 88 parameters of an electrochemistry-based model for 7.5 hours of cycling data are discussed. In addition, a detailed analysis of local parameter identifiability is presented.

AB - This paper examines the identification of the parameters of the Doyle-Fuller-Newman electrochemistry-based Lithium-ion battery model from voltage and current cycling data. The battery used in this study has a lithium iron phosphate cathode chemistry intended for high-power applications such as plug-in hybrid electric vehicles. The variables optimized for model identification include parameterizations of the model's anode equilibrium potential, cathode equilibrium potential, and solution conductivity. A genetic algorithm is used to optimize these model parameters against experimental data. The resulting identified model fits two experimental data sets used for system identification, as well as separate validation data sets corresponding to five different vehicle drive cycles. These drive cycles simulate the current a battery would undergo while used in a plug-in hybrid vehicle battery pack. The accuracy of the parameters is investigated using various validation data sets. This is believed to be the first attempt at fitting nearly all of the parameters and functions in the DFN model simultaneously using only voltage and current data. Computational logistics of using a genetic algorithm to identify 88 parameters of an electrochemistry-based model for 7.5 hours of cycling data are discussed. In addition, a detailed analysis of local parameter identifiability is presented.

UR - http://www.scopus.com/inward/record.url?scp=80053161005&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80053161005&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:80053161005

SN - 9781457700804

T3 - Proceedings of the American Control Conference

SP - 362

EP - 369

BT - Proceedings of the 2011 American Control Conference, ACC 2011

ER -

Forman JC, Moura SJ, Stein JL, Fathy HK. Genetic parameter identification of the Doyle-Fuller-Newman model from experimental cycling of a LiFePO4 battery. In Proceedings of the 2011 American Control Conference, ACC 2011. 2011. p. 362-369. 5991183. (Proceedings of the American Control Conference).