Time-domain fitting of battery electrochemical impedance models
Electrochemical impedance spectroscopy (EIS) is an effective technique for diagnosing the behaviour of electrochemical devices such as batteries and fuel cells, usually by fitting data to an equivalent circuit model (ECM). The common approach in the laboratory is to measure the impedance spectrum of...
| Main Authors: | , , |
|---|---|
| 格式: | Journal article |
| 出版: |
Elsevier
2015
|
| _version_ | 1826276072801435648 |
|---|---|
| author | Alavi, S Birkl, C Howey, D |
| author_facet | Alavi, S Birkl, C Howey, D |
| author_sort | Alavi, S |
| collection | OXFORD |
| description | Electrochemical impedance spectroscopy (EIS) is an effective technique for diagnosing the behaviour of electrochemical devices such as batteries and fuel cells, usually by fitting data to an equivalent circuit model (ECM). The common approach in the laboratory is to measure the impedance spectrum of a cell in the frequency domain using a single sine sweep signal, then fit the ECM parameters in the frequency domain. This paper focuses instead on estimation of the ECM parameters directly from time-domain data. This may be advantageous for parameter estimation in practical applications such as automotive systems including battery-powered vehicles, where the data may be heavily corrupted by noise. The proposed methodology is based on the simplified refined instrumental variable for continuous-time fractional systems method ('srivcf'), provided by the Crone toolbox [1,2], combined with gradient-based optimisation to estimate the order of the fractional term in the ECM. The approach was tested first on synthetic data and then on real data measured from a 26650 lithium-ion iron phosphate cell with low-cost equipment. The resulting Nyquist plots from the time-domain fitted models match the impedance spectrum closely (much more accurately than when a Randles model is assumed), and the fitted parameters as separately determined through a laboratory potentiostat with frequency domain fitting match to within 13%. |
| first_indexed | 2024-03-06T23:08:29Z |
| format | Journal article |
| id | oxford-uuid:64ab13db-a77e-47ae-b0ed-4d87a364a0c4 |
| institution | University of Oxford |
| last_indexed | 2024-03-06T23:08:29Z |
| publishDate | 2015 |
| publisher | Elsevier |
| record_format | dspace |
| spelling | oxford-uuid:64ab13db-a77e-47ae-b0ed-4d87a364a0c42022-03-26T18:20:19ZTime-domain fitting of battery electrochemical impedance modelsJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:64ab13db-a77e-47ae-b0ed-4d87a364a0c4Symplectic Elements at OxfordElsevier2015Alavi, SBirkl, CHowey, DElectrochemical impedance spectroscopy (EIS) is an effective technique for diagnosing the behaviour of electrochemical devices such as batteries and fuel cells, usually by fitting data to an equivalent circuit model (ECM). The common approach in the laboratory is to measure the impedance spectrum of a cell in the frequency domain using a single sine sweep signal, then fit the ECM parameters in the frequency domain. This paper focuses instead on estimation of the ECM parameters directly from time-domain data. This may be advantageous for parameter estimation in practical applications such as automotive systems including battery-powered vehicles, where the data may be heavily corrupted by noise. The proposed methodology is based on the simplified refined instrumental variable for continuous-time fractional systems method ('srivcf'), provided by the Crone toolbox [1,2], combined with gradient-based optimisation to estimate the order of the fractional term in the ECM. The approach was tested first on synthetic data and then on real data measured from a 26650 lithium-ion iron phosphate cell with low-cost equipment. The resulting Nyquist plots from the time-domain fitted models match the impedance spectrum closely (much more accurately than when a Randles model is assumed), and the fitted parameters as separately determined through a laboratory potentiostat with frequency domain fitting match to within 13%. |
| spellingShingle | Alavi, S Birkl, C Howey, D Time-domain fitting of battery electrochemical impedance models |
| title | Time-domain fitting of battery electrochemical impedance models |
| title_full | Time-domain fitting of battery electrochemical impedance models |
| title_fullStr | Time-domain fitting of battery electrochemical impedance models |
| title_full_unstemmed | Time-domain fitting of battery electrochemical impedance models |
| title_short | Time-domain fitting of battery electrochemical impedance models |
| title_sort | time domain fitting of battery electrochemical impedance models |
| work_keys_str_mv | AT alavis timedomainfittingofbatteryelectrochemicalimpedancemodels AT birklc timedomainfittingofbatteryelectrochemicalimpedancemodels AT howeyd timedomainfittingofbatteryelectrochemicalimpedancemodels |