Correlating Battery Internal Mechanisms with Equivalent Circuit Models via EIS Features
ECS Meeting Abstracts · 07.07.2026 · MA2026-01 (1) · 102-102
DOI: 10.1149/ma2026-011102mtgabs ↗Kurz gefasst
Fachartikel in der Zeitschrift ECS Meeting Abstracts, erschienen am 07.07.2026. Verfasst von Cheng-Hsi Tien und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Battery modeling is critical for performance analysis and relies primarily on two major approaches, the pseudo-two-dimensional (P2D) model and the equivalent circuit model (ECM). While ECMs are widely adopted for real-time applications due to their efficiency, their lack of direct physical interpretation limits the observability of internal electrochemical states. In contrast, P2D models capture detailed transport and reaction processes but are computationally expensive. To bridge this gap, this research presents a feature-based mapping framework using Electrochemical Impedance Spectroscopy (EIS) to correlate P2D physical parameters directly with ECM circuit elements. Figure 1 summarizes the overall workflow used to connect the two model domains. The research integrates COMSOL simulations for the P2D model with systematic parameter sweeps for the ECM. For the physics-based domain, the P2D model incorporates geometric, transport, and kinetic parameters. As detailed in Table 1, we examine critical physical variables including electrolyte volume fractions ( ε l ), solid diffusion coefficients ( D s ), electrical conductivities ( σ ), and reaction rates ( K ) across both positive and negative electrodes. The ECM domain employs a 1RCPE architecture (Table 2), defined by ohmic resistance ( R 0 ), charge-transfer resistance ( R 1 ), constant phase element parameters ( Q 1 , n 1 ), and the Warburg diffusion element ( W ). To link the two parameter spaces, this study establishes a unified feature representation defined in Table 3. A vector of geometric descriptors is obtained and extracted from each Nyquist response, including the high-frequency intercept ( R 0,hf ), semicircle attributes (width W sc , height H sc , aspect ratio AR sc ), the imaginary turning point ( Z" turn ), and the low-frequency tail length ( L tail ). These descriptors constitute a unified feature representation, allowing for direct quantitative correlation between the two distinct parameter spaces. Given that P2D physical parameters and ECM circuit elements possess disparate units and magnitudes, direct comparison is infeasible. To resolve this, we implement a normalized sensitivity analysis based on relative parameter perturbation. Specifically, characteristic points are first identified from the complex impedance data to compute the six geometric descriptors. Each P2D or ECM parameter ( θ ) was then perturbed slightly around its nominal level, and the resulting change in every feature ( f ) is expressed in percentage form. The relative variation is defined in Eq. (1).Here, i is the index of the impedance features, j is the index of the model parameters, and the subscript “ 0 ” denotes the nominal value. The normalized sensitivity was first defined as the ratio of these relative changes, as shown in Eq. (2), and in practice was evaluated more robustly by fitting a linear relation across multiple sweep points, which is expressed in Eq. (3). The superscript k denotes the k -th perturbation level. This procedure yields a dimensionless sensitivity matrix that is used to compare how the P2D and ECM parameter spaces act on the common impedance-feature domain. These quantitative indices constitute the basis for the visualization in Figures 2 and 3, where the Sankey flow width is directly proportional to the normalized sensitivity magnitude ( S ij ), visually decomposing the complex parameter interactions. Sensitivity mapping reveals clear structural alignments between the two domains. As shown in Figure 2, negative-electrode electrolyte volume fraction ( ε l,neg ) and negative-electrode reaction rate constants ( K neg ) exhibit strong linkages to the semicircle geometry features ( W sc , H sc ). Figure 3 mirrors this pattern, where ECM parameters R 1 and n 1 target these exact same descriptors, quantitatively confirming that electrode reaction kinetics dictate the mid-frequency arc geometry. In contrast, the low-frequency response is governed primarily by mass transport. Specifically, the P2D sensitivity map identifies positive-electrode solid diffusion coefficients ( D s,pos ) and positive-electrode maximum ionic concentration ( c s,pos max ) as the primary drivers of the tail length ( L tail ) and turning point. This specific dependency aligns robustly with the Warburg element ( W ), which exclusively governs the low-frequency tail, establishing a direct physical link for mass transport phenomena. By systematically mapping these sensitivities, the proposed framework validates that specific circuit elements are verifiable proxies for underlying physicochemical processes. Moreover, the framework offers a practical pathway for model reduction and hybrid modeling. By imbuing standard circuit elements with P2D physical fidelity, it facilitates advanced parameter estimation and low-order surrogate construction. This synthesis yields physics-informed ECMs, enabling real-time mechanistic diagnosis that combines computational efficiency with deep electrochemical insight. Figure 1
Zitieren
APA
Tien, C., Ko, C., & Chen, K. (2026). Correlating Battery Internal Mechanisms with Equivalent Circuit Models via EIS Features. ECS Meeting Abstracts, MA2026-01(1), 102-102. https://doi.org/10.1149/ma2026-011102mtgabs
Vancouver
Tien C, Ko C, Chen K. Correlating Battery Internal Mechanisms with Equivalent Circuit Models via EIS Features. ECS Meeting Abstracts. 2026;MA2026-01(1):102-102. doi:10.1149/ma2026-011102mtgabs
BibTeX
@article{Tien2026,
author = {Tien, Cheng-Hsi and Ko, Chi-Jyun and Chen, Kuo-Ching},
title = {Correlating Battery Internal Mechanisms with Equivalent Circuit Models via EIS Features},
journal = {ECS Meeting Abstracts},
year = {2026},
volume = {MA2026-01},
number = {1},
pages = {102--102},
doi = {10.1149/ma2026-011102mtgabs}
}
EUVIMED