pith. sign in

arxiv: 2604.10813 · v2 · pith:A24RM5J3new · submitted 2026-04-12 · 📡 eess.SY · cs.SY

System Identification of Lithium-Ion Battery Equivalent Circuit Models Using Ensemble Kalman Inversion

Farzaneh Barat , Sara Wilson , Huijeong Kim , Huazhen Fang This is my paper

Pith reviewed 2026-05-10 15:22 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords system identificationensemble Kalman inversionlithium-ion batteriesequivalent circuit modelsparameter estimationelectro-thermal modelsnonlinear dynamics
0
0 comments X

The pith

Ensemble Kalman inversion identifies parameters in lithium-ion battery models with accurate results and rapid convergence.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper applies ensemble Kalman inversion to estimate unknown parameters in lithium-ion battery equivalent circuit models that include coupled electrical and thermal effects. It evolves an ensemble of candidate parameter sets through repeated sampling and update steps that use local approximations to search for the values that best match observed data. This approach is tested on two different circuit models through both computer simulations and laboratory experiments. A reader would care because accurate parameters are needed to predict battery behavior in applications like electric vehicles, where mismatches can affect efficiency or safety estimates. The method is presented as handling the nonlinearity and large parameter counts that make traditional identification difficult.

Core claim

Ensemble Kalman inversion performs maximum a posteriori parameter estimation by evolving an ensemble of samples through successive local Gaussian approximations and Kalman-type updates, enabling an iterative search that combines Monte Carlo sampling with incremental corrections. When applied to two equivalent circuit models with coupled electro-thermal dynamics, this produces accurate parameter estimates and rapid convergence, as confirmed by both simulation studies and real experimental data.

What carries the argument

Ensemble Kalman inversion, which evolves an ensemble of parameter samples via Monte Carlo sampling combined with Kalman-type updates to perform iterative maximum a posteriori estimation.

If this is right

  • Accurate parameter estimates are obtained for nonlinear models with coupled electro-thermal dynamics.
  • The iterative process converges rapidly in both simulated and experimental settings.
  • The same procedure can be applied to additional battery models beyond the two tested here.
  • The combination of sampling and updates provides empirical stability when identifying many parameters at once.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method's stability might allow it to track slowly changing parameters over the life of a battery if run periodically on operating data.
  • Similar ensemble-based inversion could be tried on other complex engineering systems such as motors or power converters that also have nonlinear coupled dynamics.
  • If the approach scales well with model size, it could reduce the need for manual tuning of battery model parameters in control software.

Load-bearing premise

The local approximations used in each iteration remain reliable for the battery's nonlinear multi-physics behavior, and the measurements excite all relevant parameters without significant unknown disturbances.

What would settle it

A controlled simulation where all true parameter values are known in advance, followed by checking whether the final estimates from the method match those true values within a small error after the reported number of iterations.

Figures

Figures reproduced from arXiv: 2604.10813 by Farzaneh Barat, Huazhen Fang, Huijeong Kim, Sara Wilson.

Figure 1
Figure 1. Figure 1: The TheveninT model, which couples the Thevenin [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The NDCT model, which couples the NDC submodel [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: We now examine system identification for the NDCT model, which contains 12 parameters—three more than the TheveninT model. Despite this increased complexity, the proposed approach delivers effective performance. As re￾ported in Table II, the estimated parameters closely match the nominal values, with most relative percentage errors below 1%. The estimates of κ1 and κ2 are less accurate, again due to the lo… view at source ↗
Figure 3
Figure 3. Figure 3: Boxplots of the ensembles for the parameters during the iterations in identifying the TheveninT model. The central [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of the measured and predicted voltage [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Boxplots of the ensembles for the parameters during the iterations in identifying the NDCT model. The central red [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Comparison of the measured and predicted voltage [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
read the original abstract

System identification remains an intriguing challenge for lithium-ion batteries, as many models are nonlinear, exhibit multi-physics coupling, and involve a large number of parameters. In this paper, we address this challenge using the ensemble Kalman inversion (EnKI) method for battery system identification. EnKI performs maximum a posteriori parameter estimation through successive local Gaussian approximations, enabling an iterative and incremental search for unknown parameters. The search combines Monte Carlo sampling with Kalman-type updates to evolve an ensemble of samples, thereby offering empirical stability and the ability to handle strongly nonlinear models. We validate the proposed approach on two equivalent circuit models with coupled electro-thermal dynamics, through both simulation and experiments. The results demonstrate that the proposed approach achieves accurate parameter estimation with rapid iterative convergence, and it shows strong potential for application to other battery models.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes the use of ensemble Kalman inversion (EnKI) for system identification of lithium-ion battery equivalent circuit models (ECMs) incorporating coupled electro-thermal dynamics. The method is validated on two ECMs using both simulation data with known ground-truth parameters and experimental data from physical battery cells, with claims of accurate parameter estimation and rapid convergence.

Significance. If the central claims hold, the work provides a practical iterative method for parameter estimation in complex, nonlinear battery models that could enhance battery management systems. The ensemble-based approach offers advantages in handling nonlinearity and providing empirical stability, which is valuable for multi-physics systems.

major comments (3)
  1. [Experimental validation] The assessment of 'accurate parameter estimation' in experiments is based solely on convergence of terminal voltage residuals and iterative behavior, without providing quantitative error metrics for the estimated parameters, baseline comparisons (e.g., to least-squares methods), or independent verification of parameter values against ground truth. This weakens the claim since voltage fit alone may not guarantee unique or physically accurate parameters due to nonlinear coupling.
  2. [Method description and assumptions] The reliance on local Gaussian approximations in EnKI for strongly nonlinear battery dynamics is not sufficiently justified; no analysis is provided on the validity of these approximations or sensitivity to ensemble size and initialization, which is load-bearing for the convergence claims.
  3. [Results and discussion] No discussion of identifiability or posterior concentration is included, leaving open the possibility that multiple parameter sets fit the data equally well under the chosen excitation profiles.
minor comments (2)
  1. [Abstract] The abstract claims 'strong potential for application to other battery models' but provides no specific examples or discussion of limitations for broader applicability.
  2. [Notation] Ensure consistent use of symbols for parameters across equations and text to avoid confusion in the ECM descriptions.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments, which help improve the clarity and rigor of our work on ensemble Kalman inversion for battery system identification. We address each major comment point by point below, indicating planned revisions where appropriate.

read point-by-point responses

  1. Referee: The assessment of 'accurate parameter estimation' in experiments is based solely on convergence of terminal voltage residuals and iterative behavior, without providing quantitative error metrics for the estimated parameters, baseline comparisons (e.g., to least-squares methods), or independent verification of parameter values against ground truth. This weakens the claim since voltage fit alone may not guarantee unique or physically accurate parameters due to nonlinear coupling.

    Authors: We agree that experimental validation lacks ground-truth parameters, limiting direct quantitative error metrics. Our simulation results do include comparisons against known true values, showing low estimation errors. In experiments, terminal voltage residual convergence is a common proxy in the battery identification literature when ground truth is unavailable. To strengthen the manuscript, we will add a least-squares baseline comparison in the simulation studies, report ensemble-derived parameter uncertainty statistics, and include an explicit discussion of experimental limitations without independent verification. This addresses the concern without overstating experimental claims. revision: partial

  2. Referee: The reliance on local Gaussian approximations in EnKI for strongly nonlinear battery dynamics is not sufficiently justified; no analysis is provided on the validity of these approximations or sensitivity to ensemble size and initialization, which is load-bearing for the convergence claims.

    Authors: We acknowledge the need for stronger justification of the local Gaussian approximations central to EnKI. The method relies on ensemble-estimated covariances for iterative updates, which empirically handle nonlinearity in our electro-thermal models. We will revise the method description to include a short theoretical note on the approximation's validity for this class of problems, supported by references to prior EnKI applications in nonlinear systems. Additionally, we will add sensitivity analysis results for ensemble sizes (e.g., 50–200 members) and varied initializations, demonstrating consistent convergence behavior. revision: yes

  3. Referee: No discussion of identifiability or posterior concentration is included, leaving open the possibility that multiple parameter sets fit the data equally well under the chosen excitation profiles.

    Authors: We will incorporate a dedicated discussion on parameter identifiability. This will analyze the final ensemble covariance as a proxy for posterior concentration and evaluate how the multi-step current and temperature excitation profiles excite the electro-thermal dynamics to reduce ambiguity. We will also reference existing battery identifiability studies to contextualize our results and note any remaining non-uniqueness risks. revision: yes

Circularity Check

0 steps flagged

No circularity: EnKI parameter estimation applied to external battery data with independent validation

full rationale

The paper applies ensemble Kalman inversion to estimate ECM parameters from measured voltage and temperature data. In simulation sections, ground-truth parameters are known a priori and used to compute estimation errors, providing an external benchmark. In experiments, the method is run on physical cell data and evaluated by convergence behavior plus terminal-voltage residuals; these metrics are not constructed from the estimated parameters themselves. No equation or claim reduces a reported result to a fitted quantity by definition, and no load-bearing step relies on a self-citation chain that itself assumes the target result. The derivation chain therefore remains self-contained against external data.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard assumptions of the ensemble Kalman inversion algorithm and the chosen equivalent-circuit battery structures; no new free parameters, axioms, or entities are introduced beyond those already standard in the field.

axioms (1)
  • domain assumption Ensemble Kalman inversion performs maximum a posteriori estimation through successive local Gaussian approximations
    Explicitly stated in the abstract as the mechanism enabling iterative search for unknown parameters.

pith-pipeline@v0.9.0 · 5442 in / 1069 out tokens · 28764 ms · 2026-05-10T15:22:49.656430+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.