Model-based Battery Management Systems Research Articles

Online parameter identification for equivalent circuit model of lithium-ion battery

Parameter identification is the most fundamental task for the model-based battery management system. However, there are some difficulties in completing this task since most of the existing methods require at least one known parameter or a time-consuming offline procedure to extract parameters from measured data. Based on the well-known thevenin equivalent circuit for battery, this paper determines the unique purpose is introducing the bounded varying forgetting factor recursive least square approach which identifies online all the parameters of the battery model at the same time. The precision of the proposed method is verified by simulation with the error converged to zero and the maximum error less than 1% of the nominal value.

Impedance response simulation strategies for lithium-ion battery models

The battery is an electrochemical system which may be considered a black box with no practical way of observing processes occurring within in a nondestructive manner at an affordable cost. Fortunately, most physical and chemical processes in electrochemical systems can be distinguished by their distinct characteristic time constants. Electrochemical impedance spectroscopy (EIS) is a powerful technique to distinguish internal processes within batteries based on their frequency response. EIS has been successful at identifying relevant electrochemical mechanisms and battery parameters and therefore can be integrated with model-based battery management systems (BMS) which are critical for improving the battery life and performance. In this article, we provide our perspective on different simulation strategies for modeling the impedance response of lithium-ion batteries, implementation of EIS models in BMS, and some challenges associated with achieving a computationally efficient approach. • EIS is a powerful technique to characterize and identify internal processes within batteries based on their frequency response. • Battery models for EIS can largely be categorized into Empirical and Physics-based models. • Reduced order physics-based models (SPM,P2D) are proven to be faster, accurate with higher fidelity. • Battery management systems (BMS) integrated with EIS can prove to a good diagnostic tool for batteries in operation. • Robust and efficient state/parameter estimation techniques are critical for BMS implementation.

Lithium-Ion Battery Parameter Identification for Hybrid and Electric Vehicles Using Drive Cycle Data

This paper proposes an approach for the accurate and efficient parameter identification of lithium-ion battery packs using only drive cycle data obtained from hybrid or electric vehicles. The approach was experimentally validated using data collected from a BMW i8 hybrid vehicle. The dual polarization model was used, and a new open circuit voltage equation was proposed based on a simplification of the combined model, with the aim of reducing the number of parameters to be identified. The parameter identification was performed using NEDC data collected on a rolling road dynamometer; the results showed that the proposed model improved the accuracy of terminal voltage estimation, reducing the peak voltage error from 2.16% using the Nernst model to 1.28%. Furthermore, the robustness of these models in maintaining accuracy when new drive cycles were used was evaluated by comparing WLTC simulations with experimental measurements. The proposed model showed improved robustness, with a reduction in RMS error of more than 50% compared to the Nernst model. These findings are significant because they will improve the accuracy of model-based battery management systems used in electric vehicles, allowing for improved performance prediction without the requirement of recharacterization for different drive cycles or individual cell characterization.

Parameter Estimation of Lithium-Ion Battery Models Using a Novel Tanks-in-Series Approach

Lithium-ion batteries play a vital role in electric vehicles and energy storage systems. In order to monitor, predict and control the performance of lithium-ion batteries, advanced Battery Management Systems (BMS) are key. Intensive efforts are being pursued to develop electrochemical model-based battery management systems (BMS).1 The accuracy and predictability of a model depends on the precision of its parameters. Precise parameter estimation is critical for efficient BMS performance. Moreover, model parameters tend to change with battery aging. Efficient BMS performance requires the accurate tracking and update of relevant parameters.Estimating parameters for electrochemical models is challenging due to the complexity of the governing equations, and the possibility of degeneracy, with multiple sets of parameter values that give the same accuracy for fitting charge/discharge data. Parameter estimation of various lithium-ion battery systems has been attempted with different model types, including equivalent circuit model2, single particle model3 and pseudo 2D (P2D) model.4,5 Literature approaches to estimate parameters using GA6, heuristic algorithms7 etc., are difficult to implement in real-time due to computational complexity. In the past, we proposed and implemented reformulated p2D models for real-time parameter estimation.8,9 While the reformulated models are computationally inexpensive, the large set of parameters and resulting likelihood of degeneracy introduce substantial complexity.In this presentation, we propose to estimate parameters of the recently published Tanks-in-Series model10. This model is generated by systematic volume-averaging of the p2D model that enables efficient and real-time simulation. The number of parameters to be estimated in the tank model can be reduced by grouping the transport, design and kinetic parameters. The estimation is performed on experimental charge/discharge data for cylindrical cells. The accuracy of the estimated parameters is validated by using a dynamic charging profile (FUDS cycle test) and measuring the error in voltage-time curves. Acknowledgements The authors acknowledge funding from the Department of Energy, Office of Electricity, Energy Storage Program. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525 and the Texas Materials Institute at The University of Texas at Austin. References V. Ramadesigan, P. W. C. Northrop, S. De, S. Santhanagopalan, R. D. Braatz, and V. R. Subramanian, J. Electrochem. Soc., 159, R31–R45 (2012).Y. Hu, S. Yurkovich, Y. Guezennec, and B. J. Yurkovich, Control Eng. Pract., 17, 1190–1201 (2009).A. P. Schmidt, M. Bitzer, Á. W. Imre, and L. Guzzella, IFAC Proc. Vol., 43, 198–203 (2010).J. C. Forman, S. J. Moura, J. L. Stein, and H. K. Fathy, in 5th Annual Dynamic Systems and Control Conference,, p. 1–8 (2012).S. Santhanagopalan, Q. Guo, and R. E. White, J. Electrochem. Soc., 154, 198–206 (2007).A. Jokar, B. Rajabloo, M. Désilets, and M. Lacroix, J. Electrochem. Soc., 163, A2876–A2886 (2016).J. Li, L. Zou, F. Tian, X. Dong, Z. Zou, and H. Yang, J. Electrochem. Soc., 163, A1646–A1652 (2016).V. Boovaragavan, S. Harinipriya, and V. R. Subramanian, J. Power Sources, 183, 361–365 (2008).V. Ramadesigan, K. Chen, N. A. Burns, V. Boovaragavan, R. D. Braatz, and V. R. Subramanian, J. Electrochem. Soc., 158, A1048–A1054 (2011).A. Subramaniam, S. Kolluri, C. D. Parke, M. Pathak, S. Santhanagopalan, and V. R. Subramanian, J. Electrochem. Soc., 167, 013534-013534–18 (2020).

Parameter sensitivity analysis of electrochemical model-based battery management systems for lithium-ion batteries

Accurate identification of physical parameters of a lithium-ion electrochemical model is of critical importance for next-generation battery management systems. The complexity of the electrochemical model increases the difficulty of the identification process, and hence the analysis of parameter identifiability is the cornerstone for accurate parameter identification. The overarching goal of this paper is to analyze the parameter sensitivity of an electrochemical model under both the charging process and real-world driving cycles. The boundaries for the sensitivity analysis of 26 physical parameters are determined with a systematic benchmarking of published parameters for lithium Nickel-Manganese-Cobalt-Oxide/graphite cells. In particular, the sensitivity of the parameters is analyzed not only for terminal voltage but also for essential states in an electrochemical model-based battery management system, e.g., cathode bulk state of charge, cathode surface state of charge and anode potential. The sensitivity matrices of the parameters under different C-rates and depth of discharge regions clearly show their different influences on capacity-related parameters and other parameters. Furthermore, the rankings of the normalized parameter sensitivity index provide us the identifiability of the parameters, as well as the influence of parameter inaccuracy on the main functionalities in an electrochemical model-based battery management system.

Real-Time Parameter Estimation of Lithium-Ion Battery Models Using a Novel Tanks-in-Series Approach

Lithium-ion batteries play a vital role in electric vehicles and energy storage systems. In order to monitor, predict and control the performance of lithium-ion batteries, advanced Battery Management Systems (BMS) are key. Intensive efforts are being pursued to develop electrochemical model-based battery management systems (BMS).1 The accuracy and predictability of a model depends on the precision of its parameters. Precise parameter estimation is critical for efficient BMS performance. Moreover, model parameters tend to change with battery aging. Efficient BMS performance requires the accurate tracking and update of relevant parameters. Importantly, the estimation and update tasks need to be performed online in real-time, making computational efficiency a critical consideration.Estimating parameters for electrochemical models is challenging due to the complexity of the governing equations, and the possibility of degeneracy, with multiple sets of parameter values that give the same accuracy for fitting charge/discharge data. Parameter estimation of various lithium-ion battery systems has been attempted with different model types, including equivalent circuit model2, single particle model3 and pseudo 2D (P2D) model.4,5 Literature approaches to estimate parameters using GA6, heuristic algorithms7 etc., are difficult to implement in real-time due to computational complexity. In the past, we proposed and implemented reformulated p2D models for real-time parameter estimation.8,9 While the reformulated models are computationally inexpensive, the large set of parameters and resulting likelihood of degeneracy introduce substantial complexity.In this presentation, we propose to estimate parameters of the recently published Tanks-in-Series model.10 This model is generated by systematic volume-averaging of the p2D model that enables efficient and real-time simulation. The number of parameters to be estimated in the tank model can be reduced by grouping the transport, design and kinetic parameters. The estimation is performed on experimental charge/discharge data for cylindrical cells. The accuracy of the estimated parameters is validated by using a dynamic charging profile (FUDS cycle test) and measuring the error in voltage-time curves. Acknowledgements This research was supported by the Assistant Secretary for Energy Efficiency and Renewable Energy, Office of Vehicle Technologies of the US Department of Energy (DoE) through the Advanced Battery Materials Research Program (Battery500 Consortium) and the Texas Materials Institute at The University of Texas at Austin.

(Invited) Model-Based Battery Management System of Lithium-Ion Batteries

The advances in the automotive sector for plug-in hybrid electric vehicles (PHEV) and battery electric vehicles (BEV) along with penetration of renewables for power generation, cell phones and electric-flights have increased the use of batteries, as energy storage devices to a great deal. Some limitations of existing lithium-ion battery technology include underutilization, stress-induced material damage, capacity fade, and the potential for thermal runaway.1-3 Model-based control approaches (Model predictive control) can be used to design a smarter Battery Management System (BMS) that guarantees safety, efficiency, lifetime and performance of lithium-ion batteries at the stack and module level.4-7 This talk will present approaches for determining model-based optimal charging profiles for batteries, and experimental validation of the same. Optimal profiles that increase the cycle life of the cells were implemented on 16 Ah cells. An improvement of more than 100% in cycle life is observed experimentally, for our test conditions on this cell design.8 To test across multiple chemistries and form factors, optimal profiles were also implemented on a 1.5 Ah cylindrical cell, leading to an improved cycle life for these cells. The interplay between the fundamental depth in modeling, choice of numerical algorithms, and application-driven problem formulation will be presented. Acknowledgements The work presented herein was funded in part by the Advanced Research Projects Agency – Energy (ARPA-E), U.S. Department of Energy, under Award Number DE-AR0000275 along with the Clean Energy Institute (CEI) at the University of Washington (UW) and the Washington Research Foundation. References P. Arora, R.E. White, and M. Doyle, Journal of the Electrochemical Society, 145, 3647 (1998).V. Ramadesigan, P.W.C. Northrop, S. De, S. Santhanagopalan, R.D. Braatz, and V.R. Subramanian, Journal of the Electrochemical Society, 159, R31 (2012).S. Santhanagopalan, Q. Guo, P. Ramadass and R.E. White, Journal of Power Sources, 156, 620 (2006).B. Suthar, P.W.C. Northrop, R.D. Braatz,and V.R. Subramanian Journal of the Electrochemical Society, 161, F3144 (2014).B. Suthar, V. Ramadesigan, P.W.C. Northrop, B. Gopaluni, S. Santhanagopalan, R.D. Braatz, and V.R. Subramanian, in American Control Conference (ACC), p. 5350, Seattle, WA, USA (2013).M. Torchio, L. Magni, R.D. Braatz, and D.M. Raimondo, in 11th IFAC Symposium on Dynamics and Control of Process Systems, including Biosystems, p. 827, Trondheim, Norway (2016).M. Torchio, N.A. Wolff, D.M. Raimondo, L. Magni, U. Krewer, R.B. Gopaluni, J.A. Paulson, and R.D. Braatz, in American Control Conference (ACC), p. 4536, Chicago, IL, USA (2015).M. Pathak, D. Sonawane, M. Powell, S. Santhanagopalan, R.D. Braatz, and V.R. Subramanian, ECS Transactions, 75, 51 (2017).

(Invited) Open Software for Electrochemical Battery Modeling, Estimation, and Control

Batteries are ubiquitous. However, today’s batteries are expensive, range-limited, power-restricted, die too quickly, charge too slowly, and susceptible to safety issues. For this reason, model-based battery management systems (BMS) are of extreme interest. In this talk, we discuss eCAL’s open code for electrochemical-thermal battery models. Specifically, we also discuss applications of this code for (i) parameter identification, and (ii) optimal charging control. Finally, we close with exciting new perspectives for next-generation battery systems.

Is There Room for Theory in Data Science? Encoding Physics into Machine Learning Algorithms

Batteries are complex electrochemical devices whose short-term and long-term operational lives are highly dependent on their construction, materials, and use cases. Models have been developed and used to improve the performance of batteries by providing different material design, model based control and model based battery management systems1. There is a recent trend in data-science-based approaches for modeling different devices and processes2,3. It has been demonstrated that using machine learning techniques can retain 99.9% of the accuracy of a complex model while improving solve time by four orders of magnitude, which can allow for the use of high accuracy models in a real-time control scenario3. When data science techniques, and machine learning in particular, are implemented in battery systems, they are typically used as black boxes to predict battery behavior4. In this talk, we will highlight why this approach is ill suited for use in physics-based battery models, which are highly nonlinear, and how an understanding of the system can improve the result of the machine learning effort for use in optimization and control, while still allowing for the prediction of important internal states. Additionally, we will highlight ways to employ the flexibility of machine learning algorithms to allow for more accurate state estimation and voltage prediction. Real-time model-based control has demonstrated significant improvements in time remaining, state of charge, and state of health, even using simple equivalent circuit models5. By reducing the order and solve time of higher complexity models, we believe that better predictability is possible. Acknowledgments The authors thank the United States Department of Energy (DOE) for the financial support for this work though the Advanced Research Projects Agency – Energy (ARPA-E) award #DEAR0000275, as well as the Clean Energy Institute (CEI) through the DIRECT program at the University of Washington. References Y. Dai, and V. Srinivasan,. J. Electrochem. Soc., 163, A406 (2016).D. A. C. Beck, J. M. Carothers, V. R. Subramanian, and J. Pfaendtner, “Data science: Accelerating innovation and discovery in chemical engineering,” AIChE J., vol. 62, no. 5, pp. 1402–1416, May 2016.Blake R. Hough, author. Jim Pfaendtner, degree supervisor. 2016 Thesis (Ph. D.)--University of Washington,2016.J. Liu, A. Saxena, K. Goebel, B. Saha, and W. Wang, “An Adaptive Recurrent Neural Network for Remaining Useful Life Prediction of Lithium-ion Batteries,” NATIONAL AERONAUTICS AND SPACE ADMINISTRATION MOFFETT FIELD CA AMES RESEARCH CENTER, NATIONAL AERONAUTICS AND SPACE ADMINISTRATION MOFFETT FIELD CA AMES RESEARCH CENTER, Oct. 2010.Williard, Nick, Wei He, and Michael Pecht. "Model Based Battery Management System for Condition Based Maintenance." Technical Program for MFPT 2012, The Prognostics and Health Management Solutions Conference - PHM: Driving Efficient Operations and Maintenance, 2012

Model Based Battery Management Systems (BMS) – from Theory to Practice

Lithium-ion batteries are now extensively being used as the primary storage source. Capacity and power fade, and slow recharging times are key issues that restrict its use in many applications. Battery management systems (BMS) are critical to address these issues, along with ensuring its safety.1 An efficient BMS should be able to accurately predict the internal states of the battery such as the sate-of-charge (SOC), state-of-health (SOH) using the voltage and current measurements from the battery, along with maintaining safe operations of the battery.2 The accurate determination of these internal states holds the key for the optimal performance of batteries. The effect of inaccurate state estimation leads to conservative use of batteries and at worst, may cause potentially unsafe situations leading to thermal runaway.2,3 This necessitates the development of accurate models that could predict different states of the battery more accurately, which at the same time are not as computationally complicated to be deployable in current BMS.3-5 In our past work, we have developed physics based fast-solving electrochemical models which are computationally efficient.6,7 We have also shown the development of optimal charging protocols8-11 using these models, and have experimentally shown that such an approach to could lead to a reduction in battery footprint by atleast 20%, or an enhancement in cycle life by 40%. We have also demonstrated the fast-solving models and the optimal operation of batteries, on low-cost microcontrollers with a low memory footprint, which could control the batteries in real time, mimicking an actual BMS. Based on a detailed technology to market analysis, this decrease in footprint by 20% has significant impact in EVs, HEVs and grid applications. CAPEX cost can be significantly reduced for microgrids, making it more attractive to install the batteries. This talk is aimed at the commercialization pathway of the developed next-generation BMS to the EV, consumer electronics and grid-scale market through a student-founded UW-based startup. The interplay between the fundamental depth in modeling, choice of numerical algorithms, application driven problem formulation and impact in industries will be presented. Acknowledgements The work presented herein was funded in part by the Advanced Research Projects Agency – Energy (ARPA-E), U.S. Department of Energy, under Award Number DE-AR0000275 along with the Clean Energy Institute (CEI) at the University of Washington (UW). References K. Kelly, M. Mihalic and M. Zolot, in Proceeding of the 17th Annual Battery Conference on Applications and Advances (2002).V. Pop, H. J. Bergveld, D. Danilov, P. P. L. Regtien and P. H. L. Notten, Battery management systems: accurate state-of-charge indication for battery powered applications, Springer Verlag (2008).J. Newman, K. E. Thomas, H. Hafezi, and D. R. Wheeler, J. Electrochem. Soc., 150, A176 (2003).M. Doyle, T. F. Fuller and J. Newman, J. Electrochem. Soc., 140, 1526 (1993).T.F. Fuller, M. Doyle and J. Newman, J. Electrochem. Soc., 141, 1 (1994).P.W.C. Northrop, V. Ramadesigan, S. De, and V. R. Subramanian, J. Electrochem. Soc., 158(12), A1461 (2011).P.W.C. Northrop, M. Pathak, D. Rife, S. De, S. Santhanagopalan and V. R. Subramanian, J. Electrochem. Soc., 162(6), A940 (2015).L.T. Biegler, A.M. Cervantes, and A. Wachter, Advances in simultaneous strategies for dynamic process optimization, Chemical Engineering Science, 57(4):575, 2002.B. Suthar, V. Ramadesigan, S. De, R. D. Braatz and V.R. Subramanian, Phy.Chem.Chem. Phy., 16(1), 277 (2014).B. Suthar, P.W.C. Northrop, R.D. Braatz and V.R. Subramanian, J. Electrochem. Soc., 161(11), F3144 (2014).M. Pathak, D. Sonawane, S. Santhanagopalan, R.D. Braatz and V.R. Subramanian, ECS Trans., 75(23), 51 (2017).

Direct Estimation of Parameters from Charge-Discharge Curves of Lithium-Ion Batteries Using Pseudo-2 Dimensional (P2D) Models

Lithium-ion batteries play a vital role in electric vehicles and energy storage systems. In order to monitor, predict and control the performance of lithium-ion batteries, intensive efforts by various researchers are being pursued to develop an electrochemical model-based battery management systems (BMS)1. The accuracy and predictability of the model used are of great importance to these systems, which heavily depends on the precision of the parameters needed for these models. Estimation of the model parameters is critical for the BMS to perform efficiently. Moreover, the parameters of a battery might change as the battery ages. Tracking those parameters is required to maximize the efficiency of the BMS. In addition to this, the parameters of the battery need to be estimated online. Estimating parameters for the lithium-ion batteries is challenging due to the complexity in solving the governing equations, and the possibility of degeneracy, with multiple sets of parameter values that might provide the same accuracy for fitting charge/discharge curves. Parameter estimation of various lithium-ion battery systems has been done for different models, including equivalent circuit model2, single particle model3 and pseudo 2D (P2D) model4,5. However, most of these models are built on known open circuit voltage curves for individual electrodes, along with geometric and transport base parameters for cathode/anode thicknesses, porosities etc. In this presentation, we propose a method to estimate various parameters of the P2D model, along with the open circuit potential of the cathode. The parameters are estimated using the experimental charge/discharge curves for two different types of cell formats. The accuracy of the parameters is validated by implementing a dynamic charging profile, and measuring the error in the voltage-time curves experimentally. We also attempt to address the possibility and relative importance/impact of estimating all the parameters needed for the P2D model using just charge/discharge curves. This will be facilitated using our past results in fast simulation of battery models6,7. Acknowledgements The authors would like to thank the United States Department of Energy (DOE) for the financial support for this work through the Advanced Research Projects Agency – Energy (ARPA-E) award #DE-AR0000275. References V. Ramadesigan, P. W. C. Northrop, S. De, S. Santhanagopalan, R. D. Braatz and V. R. Subramanian, J. Electrochem. Soc., 159, R31 (2012).Y. Hu, S. Yurkovich, Y. Guezennec and B. J. Yurkovich, Control Eng. Pract., 17, 1190 (2009).A. P. Schmidt, M. Bitzer, Á. W. Imre and L. Guzzella, J. Power Sources, 195, 5071 (2010).J. C. Forman, S. J. Moura, J. L. Stein and H. K. Fathy, J. Power Sources, 210, 263 (2012).S. Santhanagopalan, Q. Guo and R. E. White, J. Electrochem. Soc., 154, A198 (2007).V. R. Subramanian, V. Boovaragavan, and V. D. Diwakar, Electrochem. Solid St., 10, A255 (2007).M. T. Lawder, V. Ramadesigan, B. Suthar and V. R. Subramanian, Computers & Chemical Engineering, 82, 283 (2015).

Using Data Science to Improve Battery Performance: How Data Analytics Can Help Drive Design and Use Decisions

Batteries are complex electrochemical devices whose short-term and long-term operational lives are highly dependent on their construction, materials, and use cases. Models have been developed and used to improve the performance of batteries by providing different material design, model based control and model based battery management systems1. There is a recent trend in data-science-based approaches for modeling different devices and processes2,3. In this talk, we will review different data science methods and approaches in the context of parameter estimation, optimal control and mechanism elucidation for lithium-ion batteries. In particular, we will discuss opportunities and challenges in applying the same for battery models – for model reduction, data set training, online estimation and control. It has been demonstrated that using machine learning techniques can retain 99.9% of the accuracy of a complex model while improving solve time by four orders of magnitude, which can allow for the use of high accuracy models in a real-time control scenario3. We will be attempting to demonstrate that this is true for battery models as well, which typically have large cpu time requirements. Real-time model-based control has demonstrated significant improvements in time remaining, state of charge, and state of health, even using simple equivalent circuit models4. By reducing the order and solve time of higher complexity models, we believe that better predictability is possible. Acknowledgments The authors thank the United States Department of Energy (DOE) for the financial support for this work though the Advanced Research Projects Agency – Energy (ARPA-E) award #DEAR0000275.

Estimation of Parameters from Charge-Discharge Curves of Lithium-Ion Batteries Using P2D Model

Lithium-ion battery plays a vital role in electric vehicles and energy storage systems. In order to monitor, predict and control the status of lithium-ion batteries, model-based battery management systems (BMS) have been intensively studied and developed by researchers(1). The accuracy and predictability of the model used are of great importance to these systems, which heavily depends on the precision of the parameters needed in these models. Recently, Battery Informatics, Inc., a spin-off startup company from the Subramanian group proposed the concept of self-learning BMS. One of the objectives for self-learning BMS is to estimate parameters online as battery operates, where the ability to accurately and quickly estimate parameters that matches charge/discharge curves is essential. Estimating parameters for the lithium-ion batteries is challenging due to the complexity of the governing equations, and the possibility of multiple set of parameters that might provide the same accuracy of fitting discharge curves. Parameter estimation of various lithium-ion battery systems has been done for different models, including equivalent circuit model(2), single particle model(3), and pseudo 2D (P2D) model(4, 5). Most of these models are built on known open circuit voltage curves for individual electrodes, base parameters for cathode/anode thickness, porosities etc.. Recently, Qi et al (6)estimated open circuit potential of a single electrode together with other parameters using single particle model from a battery discharge curve. In this presentation, we will show the estimation of electrode open circuit potential together with other parameters of the P2D model. We will attempt to address the possibility and relative importance/impact of estimating all the parameters needed for the P2D model from charge-discharge curves. This will be facilitated using our past results in model reformulation(7) and parameter estimation for fade analysis (8). Acknowledgements The authors thank the United States Department of Energy (DOE) for the financial support for this work through the Advanced Research Projects Agency-Energy (ARPA-E) award #DEAR0000275.

Correlation between the model accuracy and model-based SOC estimation

State-of-charge (SOC) estimation is a core technology for battery management systems. Considerable progress has been achieved in the study of SOC estimation algorithms, especially the algorithm on the basis of Kalman filter to meet the increasing demand of model-based battery management systems. The Kalman filter weakens the influence of white noise and initial error during SOC estimation but cannot eliminate the existing error of the battery model itself. As such, the accuracy of SOC estimation is directly related to the accuracy of the battery model. Thus far, the quantitative relationship between model accuracy and model-based SOC estimation remains unknown. This study summarizes three equivalent circuit lithium-ion battery models, namely, Thevenin, PNGV, and DP models. The model parameters are identified through hybrid pulse power characterization test. The three models are evaluated, and SOC estimation conducted by EKF-Ah method under three operating conditions are quantitatively studied. The regression and correlation of the standard deviation and normalized RMSE are studied and compared between the model error and the SOC estimation error. These parameters exhibit a strong linear relationship. Results indicate that the model accuracy affects the SOC estimation accuracy mainly in two ways: dispersion of the frequency distribution of the error and the overall level of the error. On the basis of the relationship between model error and SOC estimation error, our study provides a strategy for selecting a suitable cell model to meet the requirements of SOC precision using Kalman filter.

Model Based Battery Management Systems (BMS) - from Theory to Commercialization

Battery Management Systems (BMS) are critical to the performance of lithium-ion batteries in any application, ranging from cell phones to EVs to giant batteries in an electric grid.1 An efficient BMS should be able to accurately predict the internal states of the battery such as the Sate-of-Charge (SoC), State-of-Health (SoH) using the voltage and current measurements from the battery, along with maintaining safe operations of the battery.2 The accurate determination of these internal states holds the key for the optimal performance of batteries. The effect of inaccurate state estimation leads to conservative use of batteries and at worst, may cause potentially unsafe situations leading to thermal runaway.2,3 This necessitates the development of accurate models that could predict different states of the battery more accurately, which at the same time are not computationally as complicated to be deployed in current BMS.3-5 In our past work, we have developed physics based reformulated models which are computationally extremely efficient.6,7 We have also shown the development of optimal charging protocols8-10using these models, and have experimentally shown that such an approach could lead to a reduction in battery footprint by atleast 20%, or an enhancement in cycle life by 40%. We have also demonstrated the reformulated models and the optimal operation of batteries on low-cost microcontrollers that could control the batteries in real time, mimicking an actual BMS. Based on a detailed technology to market analysis, this decrease in footprint by 20% has significant impact in EVs, HEVs and grid applications. CAPEX cost can be reduced for microgrids. This talk is aimed at the commercialization pathway of the development of next-generation BMS for grid-scale and EV Li-ion battery deployments through a student-founded start-up named Battery Informatics, Inc (Bii). The interplay between the fundamental depth in modeling, choice of numerical algorithms, application driven problem formulation and impact in industries will be presented. Acknowledgements The work presented herein was funded in part by the Advanced Research Projects Agency – Energy (ARPA-E), U.S. Department of Energy, under Award Number DE-AR0000275 along with the Clean Energy Institute (CEI) at the University of Washington (UW) and also the Washington Research Foundation. References K. Kelly, M. Mihalic and M. Zolot, in Proceeding of the 17th Annual Battery Conference on Applications and Advances (2002).V. Pop, H. J. Bergveld, D. Danilov, P. P. L. Regtien and P. H. L. Notten, Battery management systems: accurate state-of-charge indication for battery powered applications, Springer Verlag (2008).J. Newman, K. E. Thomas, H. Hafezi, and D. R. Wheeler, J. Electrochem. Soc., 150, A176 (2003).M. Doyle, T. F. Fuller and J. Newman, J. Electrochem. Soc., 140, 1526 (1993).T. F. Fuller, M. Doyle and J. Newman, J. Electrochem. Soc., 141(1) (1994).P. W. C. Northrop, V. Ramadesigan, S. De, and V. R. Subramanian, J. Electrochem. Soc., 158(12), A1461 (2011).P. W. C. Northrop, M. Pathak, D. Rife, S. De, S. Santhanagopalan and V. R. Subramanian, J. Electrochem. Soc., 162(6), A940 (2015).Biegler, L.T. Cervantes, A. M. Wachter, A., Advances in simultaneous strategies for dynamic process optimization, Chemical Engineering Science, 57(4), 575, (2002).B. Suthar, V. Ramadesigan, S. De, R. D. Braatz and V. R. Subramanian, Phy.Chem.Chem. Phy., 16(1), 277(2014).B. Suthar, P. W.C. Northrop, R. D. Braatz and V. R. Subramanian, J. Electrochem. Soc., 161(11), F3144 (2014).

Model Based Battery Management System (BMS) for Electric Transportation and Renewable Microgrids

Lithium ion batteries are a promising technology to reduce the dependence on fossil fuels for transportation and for the creation of resilient and reliable renewable microgrids. However, battery manufacturers overdesign lithium ion batteries used in electric vehicle to ensure safety and life due to the uncertainty of the internal states of the cell. Proactive battery management systems (BMS) and advanced sensing technologies offer an opportunity to significantly reduce the cost and weight of transportation batteries, and circumvent problems arising due to capacity fade and safety concerns. This talk will describe how multiscale electrochemical engineering models, mathematical model reformulation and the use of robust algorithms can alleviate some of these problems to help electrify the transportation industry by improving the range of variables that are predictable and controllable in a battery in real-timewithin an electric vehicle. The use of battery models in a BMS will be analyzed. The interplay between the fundamental depth in modeling, choice of numerical algorithms, and application driven problem formulation will be presented. The validity of implementation in a microcontroller environment for model predictive control (MPC) will be addressed and demonstrated. In addition, preliminary results on aggressive sizing and control strategies for batteries in renewable microgrids will be presented.Acknowledgements: I would like to acknowledge my current and past students and postdoctoral associates for the results presented.

Wavelet-Based Identification Method of Li-Ion Battery Model for Electric Vehicles

Online parameters identification is one of the major functions of model-based battery management system (BMS), which can be used to monitor the working status of battery, such as state of charge (SOC) and state of health (SOH). This paper proposed a wavelet-based identification method of Li-ion battery model for Electric Vehicles (EVs). The main idea is to decompose the measured terminal voltage and current data at multiple scales, and then recursive least squares (RLS) algorithm is used to extract the model parameters at a suitable scale. The proposed method is shown to have good robustness to measured noise and thus enhances the estimation accuracy by taking advantage of the noise removal ability and signal approximation properties of wavelet decomposition.

Comparison study on the battery models used for the energy management of batteries in electric vehicles

Battery model plays an important role in the simulation of electric vehicles (EVs) and states estimation of the batteries in the development of the model-based battery management system. To build a battery model with enough precision and suitable complexity, firstly this paper summarizes the seven representative battery models, which belong to the simplified electrochemical models or the equivalent circuit models. Then the model equations are built and the model parameters are identified with an online parameter identification method. The battery test bench is built and the experiment schedule is designed. Finally an evaluation is performed on the seven battery models by an experiment approach from the aspects of the estimation accuracy of the terminal voltages. To evaluate the effect of the number of RC networks on the model’s precision, the battery general equivalent circuit models (GECMs) with different RC networks are also discussed further. The results indicate the equivalent circuit model with two RC networks, the DP model, has an optimal performance.

Modeling for Lithium-Ion Battery used in Electric Vehicles

To improve and better the applicability of lithium-ion battery model in electric vehicles, a new electrochemical-polarization model was put forward for the real-time model-based battery management system and control applications by adding an extra RC network on the basis of the electrochemical model to describe the relaxation effect of the lithium-ion battery, and the open circuit voltage as a function of State of Charge defined by the Nernst model is used in the model to avoid a time-consuming, laborious and even error-prone experiment for specially determining open circuit voltage at several specified SoC values. The model parameters are identified by the least squares method with the experimental data of hybrid power pulse characteristic test on a LiFePO4 battery module. Experiments and simulation results show the new electrochemical-polarization model can simulate the dynamics of battery well. By using the proposed model and parameters identification approach, the time-consuming and complex experiments for model parameters’ identification and periodical calibration are avoided.