Online Battery Monitoring for State-of-Charge and Power Capability Prediction by Larry W. Juang A thesis submitted in partial fulfillment of requirements for the degree of Master of Science (Electrical Engineering) at the University of Wisconsin – Madison 2010 I Online Battery Monitoring for State-of-Charge and Power Capability Prediction by Larry W. Juang Under the supervision of Professor Thomas M. Jahns and Robert D. Lorenz at the University of Wisconsin – Madison Approved by __________________________ Thomas M. Jahns Approved by __________________________ Robert D. Lorenz Date __________________________ II I Abstract This document presents an investigation of a proposed methodology that uses system identification techniques to implement a monitoring system for lead-acid batteries in an electric vehicle. Specifically, the information that the proposed methodology provides can help estimate the energy remained in the battery bank (State- of-Charge (SOC)) and the power capability of the battery bank (State-of-Function (SOF)). A combination of analytical and experimental techniques has been devised to construct an equivalent circuit model that includes a diffusion voltage term for the OPTIMA D34M battery used in the WEMPEC Corbin Sparrow vehicle. In addition, the document analyzes a set of actual road trip data from the WEMPEC Corbin Sparrow to demonstrate the proposed model’s potential use for diagnosing the relative state-of-health (SOH) of individual batteries. Detailed discussions of the limitations and assumptions associated with the model, as well as other possible alternatives, are included in the document. The performance capabilities of the proposed method for estimating battery SOC and SOF are evaluated against experimental data with an emphasis on electric vehicle (EV) operations. I Acknowledgement The author would like to gratefully acknowledge the funding and support provided by the Wisconsin Electric Machines and Power Electronics Consortium (WEMPEC) at the University of Wisconsin at Madison. Many knowledgeable individuals have provided their insights and expertises to help make this project possible. Profs. Jahns and Lorenz of WEMPEC and Prof. Setheras of the Electric and Computer Engineering (ECE) Department were generous with their time to provide necessary guidance, both research direction wise and technology wise. The author’s labortory partners, Phillip Kollmeyer, Adam Anders, and Adam Shea, have always been patient with the author’s shortcomings and eager to share their know-how, especially on the hands-on matters. Micah Erickson, along with many of the author’s colleagues in WEMPEC, has offered much valued opnions on the subject. Finally, the author thanks his parents for their unconditional support for his education and everything else in life. I I Table of Contents Abstract Acknowledgement Table of Contents ......................................................................................................... I List of Figures ..............................................................................................................V List of Tables ...............................................................................................................X Nomenclature .............................................................................................................III Introduction Background ...................................................................................................................1 Thesis Objectives...........................................................................................................1 Thesis Organization .......................................................................................................4 Chapter 1 The State-of-the-Art Review .........................................................................................5 1.1 Battery Overview..................................................................................................5 1.1.1 Battery Chemistries...................................................................................7 1.1.2 Chemical Processes.................................................................................10 1.2 Battery Applications and Monitoring System.....................................................13 1.3 State-of-Charge...................................................................................................15 1.3.1 Coulomb Counting..................................................................................17 1.3.2 Voltage-Based Methods..........................................................................18 1.3.3 Impedance-Based Methods.....................................................................18 1.3.4 Online Estimation Methods ....................................................................22 II 1.3.5 Neural Network and Fuzzy Logic Methods............................................23 1.4 State-of-Health....................................................................................................24 1.4.1 Observations between cycles ..................................................................25 1.4.2 Impedance-Based Method ......................................................................25 1.4.3 Computational Modeling for Aging Prediction......................................26 1.5 State-of-Function ................................................................................................27 1.5.1 Impedance-Based Method ......................................................................28 1.5.2 Nonlinear Modeling................................................................................29 1.6 System Identification Techniques.......................................................................30 1.6.1 Equation Error Algorithm.......................................................................30 1.6.2 Discrete Equivalents to Continuous Transfer Functions ........................32 1.7 Summary.............................................................................................................33 Chapter 2 Lead-Acid Battery Equivalent Circuit Model and State-of-Charge Estimation Methodology .......................................................................................................36 2.1 Statistical RC Model for Lead-Acid Battery ......................................................36 2.1.1 Model Formation ....................................................................................37 2.1.2 Model Theory Discussion.......................................................................46 2.1.3 Parameters Initialization .........................................................................51 2.2 Experimental Comparison between Coulomb Counting Method and Internal EMF Estimation..................................................................................................53 2.2.1 Internal EMF Estimation without the Diffusion Voltage and Coulomb Counting Method ................................................................................................53 III 2.2.2 Internal EMF Estimation with the Diffusion Voltage and Coulomb Counting Method ................................................................................................62 2.3 Summary.............................................................................................................66 Chapter 3 Model-Based Battery Power Capability Prediction ...................................................67 3.1 System Identification in the Developed Model ..................................................67 3.2 Spectral Separation Method for System Identification.......................................76 3.2.1 Simulated R-RC Model and the Built in Assumptions...........................76 3.2.2 Simulation Results ..................................................................................79 3.3 Short Term Power Capability Prediction............................................................86 3.3.1 Evolution of R0 during Discharge ..........................................................87 3.3.2 Short Term Power Capability Prediction Experiment Results ...............90 3.4 The Prospect of Long Term Power Capability Prediction..................................95 3.5 The Case Study of the Relative SOH for Batteries onboard the WEMPEC Corbin Sparrow.................................................................................................103 3.6 Summary...........................................................................................................106 Chapter 4 Conclusions ...............................................................................................................107 4.1 Summary of Contributions................................................................................107 4.2 Avenues for Further Research ..........................................................................109 Bibliography.............................................................................................................111 Appendix A EPA Driving Cycle ....................................................................................................116 IV Appendix B Experimental Set Up .................................................................................................120 Appendix C Matlab Code ..............................................................................................................130 V List of Figures Fig. 1-1 Summary Diagram for Research Approach Relative