(6df) Developing Electrochemical-Based Predictive Battery Health Monitoring for Future Battery Management Systems

Saeed Khaleghi Rahimian, Technical Lead of Modeling Team, Battery Technology Group, SERES (formerly SF Motors), Santa Clara, CA

Research Interests:

Primary Research: Develop High-fidelity Electrochemical-based Battery Life Models

Due to global concerns about the environmental impacts of fossil fuels, there has been a significant growth in Electric Vehicles (EVs) and Energy Storage Systems (ESSs) market. Li-ion batteries due to high energy and power densities are the dominant technology in EVs and they have a huge potential for use in off-grid power supply systems in combination with renewable photovoltaics and wind power. However, these batteries are prone to unexpected failures and safety issues. A battery management system (BMS) is a key element in EVs and the smart grid to make the batteries operate within the proper voltage and temperature interval, guarantee the safety and prolong their service life. To fulfill this goal, the BMS requires an accurate estimation of the battery state of charge (SOC), state of power (SOP) and state of health (SOH). The accurate and reliable online battery state estimation is a challenging issue since these states are not measurable using readily available sensors. The SOH, which depends on the battery capacity and power fade, predicts the number of times the battery can be charged and discharged before its life is terminated, thus it can help prevent catastrophic hazards and premature failure. To optimize the battery performance while maintaining its safety and reliability, a physics-based battery life model is significantly more beneficial compared to the popular data-driven or semi-empirical equivalent circuit models. Nonetheless, due to highly complex nature of degradation mechanisms in Li-ion cells [1], there is a considerable lack of high-fidelity life models that can predict the cell behavior under a variety of cycling and storage conditions.

Past/Current Works:

I am currently leading a battery modeling group to build and validate full-order/reduced-order electrochemical-based life models for commercial Li-ion cells with Ni-rich cathodes and Si/graphite composite anodes. We initially incorporated the Li plating side reaction on anode particles into the well-known Newman electrochemical model to study the impact of battery pack cooling strategy on nonuniform Li plating under fast charge conditions [2]. As part of the offline SOH estimation, implemented in the existing BMS, we built a control-oriented physics-based calendar life model by modeling the anode Solid Electrolyte Interphase (SEI) formation [3] as well as a cathode parasitic side reaction [4]. The model was able to predict the cell capacity fade and resistance rise data in a wide range of SOC and temperature storage conditions [5]. I am also working on a novel online SOH estimation approach that can track the loss of electrodes’ active materials and Li⁺ inventory by fitting the cell open circuit voltage at a few SOCs [6].

Future Direction

As an assistant professor, I would like to continue working on physics-based life modeling of Li-ion batteries to develop an advanced SOH estimation algorithm for BMS future generations. To achieve this goal, my plan is to incorporate the other critical aging phenomena such as active materials loss, electrodes’ microstructure changes and electrolyte decomposition with gas generation into the life model. This life model is crucial to determine the battery dominant degradation mechanisms during its lifetime. Thus, it would help us to optimize the operation conditions (e.g. fast charge) to enhance the battery durability while avoiding catastrophic failures. The model can also be integrated with a simplified battery pack thermal simulation to improve the battery thermal management system. The other application of the model is to predict the cell remaining useful life, which can be greatly useful for decision making on second-life of EV batteries [7]. This type of modeling would be computationally expensive; however, the simulation can be run remotely with the advanced cloud technology [8].

Secondary Research: Develop Efficient Numerical Global Optimization Methods

Global optimization (GO) possesses a vast application in different fields of science and engineering [9]. GO aims at finding the best solution or a set of local solutions with high quality of a problem subject to a set of constraints which define a feasible region. Homotopy continuation methods are promising GO techniques that extend the convergence domain of locally convergent methods (e.g. Newton-Raphson) to seek all roots of a nonlinear system of equations [10].

Past/Current Works:

In parallel to my electrochemical engineering research, I also worked on a new numerical technique that significantly enhanced the conventional homotopy continuation methods in terms of finding all solutions of a single (multiple) nonlinear equation(s) [11-12]. Recently, I came across a great mathematical paper by Decker and Keller [13], which was the motivation to improve the method robustness by applying the appropriate numerical continuation method [14]. However, the computation time increased dramatically with the number of equations (~ 2ⁿ, n= number of equations) and thus, it was limited to small systems (n<10). By using a novel idea, I was able to reduce the computation time significantly (~ n²), that allowed me to solve large systems up to 30 equations [15].

Future Direction

I have a strong interest to apply the new homotopy method to solve the nonlinear systems of equations arising in physics, chemistry and biological sciences. In particular, I would like to find all stationary points of potential energy functions, which play a crucial role in determining the rates of chemical and physical transformations. My next plan is to combine the new method with a locally convergent optimization routines (e.g. Nelder-Mead) to locate the global optimal solution(s) of objective functions with several local optimums. I am highly ambitious to develop a novel GO method to locate the global minima of potential energy functions, which is extremely useful in prediction of the structure of proteins, clusters, crystals and biomolecules. This would help to accelerate the new material discovery [16] and the structure-based computer-aided drug design [17].

Teaching Interests:

I believe a major part of my role as a teacher of engineering students is to help them prepare to be successful professional engineers. A student embarking on a career path, whether in academia or industry, should have a comprehensive knowledge base, a wide range of problem-solving skills, and the ability to communicate and collaborate with others. This involves not only imparting knowledge directly related to course materials, but also fostering the development of secondary skills which will help them approach real-world problems.

My teaching interests are quite broad at both the undergraduate and graduate levels. At the undergraduate level, I am interested in teaching Reaction Kinetics, Transport Phenomena, Thermodynamics, Numerical Methods in Chemical Engineering, and Introduction to Electrochemical Energy Systems. At the graduate level, I am primarily interested in teaching courses in Advanced Reaction Engineering, Advanced Numerical Methods applied in Chemical Engineering, and Electrochemical Energy Systems Modeling and Optimizations.

References

[1] M. Palacín et al., Why do batteries fail?, Science, 351 (2016).

[2] M. Forouzan et al., Non-Uniform Li Plating Behavior of Li-Ion Cells in a Battery Pack with Bottom, Symmetric Top-Bottom, or Side Cooling Thermal Management Systems during Fast Charging, 235^th ECS Meeting (2019).

[3] M. Pinson et al., Theory of SEI Formation in Rechargeable Batteries: Capacity Fade, Accelerated Aging and Lifetime Prediction, J. Electrochem. Soc.,160 (2013).

[4] J. Kim et al., Prospect and Reality of Ni‐Rich Cathode for Commercialization, Adv. Energy Mater. (2017).

[5] S. Khaleghi Rahimian et al., A Reduced-Order Physics-Based Calendar Life Model for Li-Ion Cells, submitted to AIChE Annual Meeting (2019).

[6] S. Khaleghi Rahimian et al., A Novel Online State of Health Estimation of Li-Ion Cells, submitted to AIChE Annual Meeting (2019).

[7] M. Baumann et al., Could-connected Battery Management for Decision Making on Second-Life of Electric Vehicle Batteries, 13^th International Conference on Ecological Vehicles and Renewable Energies (2018).

[8] T. Kim et al., Cloud-Based Battery Condition Monitoring and Fault Diagnosis Platform for Large-Scale Lithium-Ion Battery Energy Storage Systems, Energies, 11 (2018).

[9] C. Floudas et al., A Review of Recent Advances in Global Optimization, J Glob Optim, 49 (2009).

[10] M. Kuno et al., Computing All Real Solutions to Systems of Nonlinear Equations with a Global Fixed-point Homotopy, Ind. Eng. Chem. Res., 27 (1988).

[11] S. Khaleghi Rahimian, et al., A New Homotopy for Seeking all Real Roots of a Nonlinear Equation, Comp. & Chem. Eng., 35 (2011).

[12] S. Khaleghi Rahimian et al., A Robust Homotopy Continuation Method for Seeking All Real Roots of Unconstrained Systems of Nonlinear Algebraic and Transcendental Equations, Ind. Eng. Chem. Res., 50 (2011).

[13] D. Decker et al., Multiple Limit Point Bifurcation, J. Mathematical Analysis and Applications, 75 (1980).

[14] S. Khaleghi Rahimian et al., A Global Homotopy Continuation Algorithm to Solve Systems of Nonlinear Equations, submitted to Journal of Computational and Applied Mathematics.

[15] S. Khaleghi Rahimian et al., Finding All Solutions of Systems of Nonlinear Equations with an Efficient Global Homotopy Continuation Method, submitted to AIChE Annual Meeting (2019).

[16] A. Pulido et al., Functional Materials Discovery using Energy–Structure–Function Maps, Nature, 543 (2017).

[17] G. Sliwoski et al., Computational Methods in Drug Discovery, Pharmacological Reviews, 66 (2013).