Stanford Energy Control Lab Members Present at the 245th ECS Meeting

Title: Long-Term Calendar Aging across Commercial Lithium-Ion Cell Chemistries - Part II: Modeling and Early Prediction

Authors: Cui, X., Stroebl, F., Lam, V., Uppaluri, M., Chueh, W.C., Onori, S.

Abstract:

Lithium-ion (Li-ion) batteries are widely used in applications such as mobility, stationary grid systems, and various consumer and commercial systems. In electric vehicles (EVs), batteries often remain idle, with only about 10% utilization. Moreover, in stationary battery energy storage systems (BESS) designed for peak shaving, resting periods tend to cause more aging effects than the operational cycles. Consequently, quantifying the effect of calendar aging is crucial. However, current calendar aging models are inadequate for accurately modeling and predicting the diverse aging behaviors of commercial Li-ion batteries. In this study, we analyze a new, long-term Li-ion battery calendar aging dataset, which includes eight different commercial battery types. We compare and validate three calendar aging models—semi-empirical, symbolic regression, and Long Short-Term Memory Network (LSTM). We examine the transferability of these models across various commercial cell types, offering insights into their generalizability. Additionally, we introduce a novel trajectory-to-trajectory approach for early prediction of calendar aging lifetime. For the first time, we investigate the model's ability to interpolate and extrapolate aging behavior under different storage temperatures. This research provides practical insights for optimizing battery management strategies in conditions dominated by calendar aging.

Title: A Comparative Study of Numerical Methods for Lithium-Ion Battery Electrochemical Modeling

Authors: Lucero, J., Xu, L., Cooper, J., Allam, A., Onori, S.

Abstract:

Lithium-ion battery (LIB) with high specific energy density and long cycle life is one widely used energy storage technology. In order to ensure longevity and safety of the battery system, a battery management system (BMS) that relies on battery model is required. With the development of BMS technique, more attention has been given to electrochemical models that provide insights into the battery internal states. In general, battery electrochemical models consist of partial differential equations (PDEs) describing the thermodynamic and electrochemical process inside the cell, and those PDEs can only be solved numerically due to their complexity.

Existing battery simulation software utilize numerical methods such as finite difference method (FDM), finite volume method (FVM), and finite element method (FEM) to solve PDEs in battery models. However, there are little discussions for the selection of a specific numerical methods. In our recent study, we compare two spatial discretization methods commonly used to numerically solve the governing PDEs of LIB electrochemical models, namely FDM and FVM, in terms of model accuracy and mass conservation guarantee. First, we provide the mathematical details of the spatial discretization process for both FDM and FVM to solve the battery single particle model (SPM), this result has not been shown in any publication before. Second, we propose a new Hermite extrapolation based FVM scheme leading to higher accuracy when compared to the normally used linear extrapolation based FVM scheme. Then, SPM parameters are identified using experimental data, and a comprehensive parameter sensitivity analysis is conducted under different current input profiles to study parameter identifiability. Finally, model accuracy and mass conservation analysis of the FDM and FVM schemes are presented. Our study shows that the FVM scheme with Hermite extrapolation leads to accurate and robust control-oriented battery model while guaranteeing mass conservation and high accuracy. Also, the proposed new FVM scheme can be extended to solve other battery models, such as SPM model with electrolyte and Doyle-Fuller-Newman model. Moreover, we provide findings of mass conservation analysis for FDM and FVM schemes, which we hope can facilitate BMS model selection.

Title: A Data-Driven Approach for Predicting the Lifetime of Lithium-Metal Batteries

Authors: Uppaluri, M., Ma, W., Ha, S., Aliahmad, N., Saatchi, A., Littau, K., Onori, S.

Abstract:

Lithium-metal batteries (LMBs) are promising candidates for accelerating the implementation of electric vehicles in our society. LMBs comprise of a lithium-metal anode that possesses the highest theoretical specific capacity (3860 mAh/g) and the lowest redox potential (-3.04V vs standard hydrogen electrode). LMBs undergo repeated plating and stripping of lithium ions on the anode surface. The non-homogenous stripping process leads to the isolation of metallic lithium from the anode surface, forming dead-lithium. The parasitic reaction between the lithium-metal anode and the electrolyte leads to the formation of the solid electrolyte interphase (SEI) layer. These phenomena lead to the reduction of available lithium and a rise in cell impedance, causing a rapid reduction of lifetime for these batteries.

In this talk, we present a diverse dataset collected from different LMBs of different capacities and voltage ranges, that includes features from the initial formation cycle, rest voltages and CV charging during cycling. A data-driven model from the LMB cycling data is presented to predict the remaining useful lifetime (RUL) of these cells. Insights from the model can be used to determine the bad cells in a batch, improve the designs of the cells and derive optimal charging protocols for their implementation in the BMS.

Title: On-Board Diagnostics for Li-Ion Batteries Using Electrochemical Impedance Spectroscopy (EIS)-Based Health Estimation Models

Authors: Thatipamula, S., Khan, M.A., Onori, S.

Abstract:

There exists a growing need for standardized On-Board Diagnostics (OBD) for electric vehicles to provide accurate health metrics and guarantees to both consumers and manufacturers.

Previous work has shown the powerful LIB capacity-based State-of-Health (SoH) estimation capability of Electrochemical Impedance Spectroscopy (EIS) measurements and data-driven models. Since EIS measurements are dependent not only on SoH but also State-of-Charge (SoC) and temperature [4], it is important that the measurements are conducted after the cell reaches an equilibrium, and that these other variables are also tracked. Although EIS measurements are somewhat quicker than some traditional capacity-determination experimental methods, the time taken for such measurements is not insignificant. Therefore, building a pipeline to determine the EIS frequency measurements most important for SoH estimation is an important task in developing a suitable EIS-based OBD. By exploring a frequency range between 0.1 and 200 Hz, we study EIS measurements related to diffusion and charge transfer processes in LIB operation, with each process being partially distinguishable due to their varying timescales.

In this work, using EIS data collected from 5Ah LIBs with an NMC-111 cathode and a graphite anode at various SoCs (0, 25, 50, 75 and 100%) and cell lifetime (0, 10, 20, 40 and 90 days) as input features, we develop sequential, data-driven health estimation models for LIBs. The 22 cells used for this analysis have been aged over a period of 90 days in two different ways: either through active cycling at different C-rates (0.2 and 1C) and temperatures (0, 25 and 40⁰C), or passive “calendar aging” where the cells are left without use at a specific temperature and SOC. Using feature attribution techniques (Shapley values, feature occlusion, etc.), we find the most influential of the frequency ranges in the EIS measurements that relate strongly with cell performance degradation. To develop a streamlined and efficient SoH estimation framework, we formulate an optimization problem to find the EIS experimental design in terms of frequency ranges that delivers maximum accuracy in SoH estimation at different points in the lifetime of the cell. These streamlined and optimized EIS experimental designs and SoH estimation models can be used directly in the development of rapid and efficient on-board diagnostic tools.

Title: A Data-Driven Modeling Framework for Lithium Metal Battery Health Estimation

Authors: Ma, W., Ha, S., Pozzato, G., Saatchi, A., Aliahmad, N., Littau, K., Onori, S.

Abstract:

The Achilles' heel of LMB, despite its high theoretical energy density, has been its limited lifespan and variability in aging trajectories due to coupling of multiple failure mechanisms. Effective health estimation could enable timely maintenance actions before failure and a rapid LMB development iteration. Data-driven methods capture a wide spectrum of failure mechanisms and incorporates aging behavior variability. Recent work around data-driven LMB health estimation rely on empirical correlations observed specific to the datasets, and lack adaptability for a wide variety of experimental conditions and cell behaviors, especially those with a flat discharge capacity retention followed by abrupt decline.

In this work, we cycled 36 LMB cells across various temperatures and charge/discharge rates. The examined cells exhibit significant cell-to-cell variabilities even under identical experimental conditions, and consistent abrupt capacity fade behaviors. We extracted an extensive inventory of non-invasive features based on cycle-to-cycle variation of LMB external voltage-current measurements throughout aging, and managed to accurately track LMB capacity fade with quantified confidence intervals using these features combined with machine learning tools. The identified features provide not only reliable and robust LMB health monitoring performance, but also starting points for further first-principal modeling and experimental investigation to demystify LMB aging mechanisms. The abundance of discovered features, along with an automatic and time-efficient correlation mining pipeline, enables the generalizability of our proposed framework toward a wide variety of LMB cells.

Title: Long-Term Calendar Aging across Commercial Lithium-Ion Cell Chemistries- Part I – Exploratory Data Analysis

Authors: Lam, V., Cui, X., Stroebl, F., Uppaluri, M., Onori, S., Chueh, W.C.

Abstract:

In this study, calendar aging is investigated across different cell chemistries, temperatures, and state of charges in a dataset spanning over 10 years. By tracking the capacity and resistance over the course of degradation, we analyze the similarities and difference in calendar aging seen across chemistries and testing conditions. We parameterize degradation curves using standard semi-empirical models to assess Arrhenius dependence, t^x fitting, and time dependent non-linearity of calendar aging across different chemistries, testing conditions, and cell to cell variability. Through this process we demonstrate the complexity of calendar aging and the difficulty of its modeling, calling for the need for more robust prediction models.