pith. sign in

arxiv: 2606.28507 · v1 · pith:2O3M7T25new · submitted 2026-06-26 · 📡 eess.SY · cs.SY· math.OC

Rapid and robust parameter estimation for electrochemical battery models via BOLT: A batch-optimized local-to-global technique

Feng Guo , Luis D. Couto , Keivan Haghverdi , Khiem Trad , Grietus Mulder This is my paper

Pith reviewed 2026-06-30 01:24 UTC · model grok-4.3

classification 📡 eess.SY cs.SYmath.OC
keywords parameter estimationbattery modelselectrochemical modelsoptimizationlithium-ion batteriesmodel calibrationsingle-particle modeltrust-region methods
0
0 comments X

The pith

BOLT estimates electrochemical battery model parameters to 12 mV accuracy using roughly 20,000 evaluations in under 9 seconds per run.

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

The paper introduces BOLT as a workflow that initializes many candidate parameter vectors, refines them in parallel with trust-region local optimization, accelerates each model run with just-in-time compilation, and retains only those vectors that produce consistent voltage predictions across multiple operating conditions. Experiments on a grouped single-particle model of a commercial NMC cell show that the resulting parameter sets match measured voltages with an average mean absolute error of 12.4 mV while requiring far fewer model calls and exhibiting lower run-to-run variability than particle-swarm or genetic-algorithm baselines. Synthetic tests with a known true parameter vector confirm that BOLT recovers the reference values to within 0.6 percent relative error even when 1–3 mV of voltage noise is added. A reader would care because repeated, reliable calibration is a bottleneck for battery-management systems, digital twins, and second-life screening that must run on modest hardware.

Core claim

BOLT achieves a favorable trade-off among voltage-response accuracy, computational efficiency, and repeated-run stability by combining diversified candidate initialization, batch-parallel trust-region reflective local refinement, JIT-accelerated model evaluation, and multi-condition consistency screening within a single calibration workflow.

What carries the argument

BOLT (Batch-Optimized Local-to-Global Technique), which launches parallel local trust-region refinements on a diversified set of initial guesses and then screens survivors for voltage consistency across operating conditions.

If this is right

  • BOLT(32) produces 12.4 mV average MAE over five operating conditions while using only 20,636 model calls and 8.97 seconds per run.
  • The method recovers a known reference parameter vector with mean absolute relative error below 0.6 percent in synthetic data.
  • Performance remains stable under 1–3 mV voltage noise perturbations.
  • The workflow is intended for BMS parameter updating, control-oriented digital twins, and second-life battery screening.

Where Pith is reading between the lines

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

  • If the consistency screen generalizes, BOLT could be applied to full-order electrochemical models without changing the outer workflow.
  • The batch structure suggests the same pattern could accelerate parameter fitting for other expensive physics simulators that lack analytic gradients.
  • Frequent on-device recalibration becomes feasible once the per-run cost drops below ten seconds.

Load-bearing premise

The grouped single-particle model together with the five chosen operating conditions and noise levels are representative enough that consistency screening will accept good parameter vectors and reject poor ones for other cells or chemistries.

What would settle it

Running BOLT on a different lithium-ion chemistry or on a higher-fidelity electrochemical model and measuring whether the selected parameters still reproduce measured voltages to within 12 mV under previously unseen current profiles.

Figures

Figures reproduced from arXiv: 2606.28507 by Feng Guo, Grietus Mulder, Keivan Haghverdi, Khiem Trad, Luis D. Couto.

Figure 1
Figure 1. Figure 1: Schematic illustration of the SPM considered in this study. [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Workflow of the proposed BOLT framework. Initial parameter vectors are ran [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Operating conditions considered in this study. (a) Measured terminal-voltage [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Effect of the total candidate number N on the MAE distribution of the BOLT method over 50 repeated runs. As shown in [PITH_FULL_IMAGE:figures/full_fig_p018_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Synthetic recoverability and noise robustness of BOLT(32) over the five operating [PITH_FULL_IMAGE:figures/full_fig_p021_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Overall MAE distributions of BOLT(32), PSO2, PSO5, GA2, and GA5 over 50 [PITH_FULL_IMAGE:figures/full_fig_p027_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Condition-wise MAE distributions of BOLT(32), PSO2, PSO5, GA2, and GA5 [PITH_FULL_IMAGE:figures/full_fig_p030_7.png] view at source ↗
read the original abstract

Accurate and efficient parameter estimation is essential for applying electrochemical battery models in simulation, state estimation, control, and repeated model updating. However, conventional optimization methods, such as particle swarm optimization (PSO) and genetic algorithms (GA), often require many model evaluations and show considerable run-to-run variability, limiting their use in time-sensitive calibration scenarios. This study proposes a Batch-Optimized Local-to-Global Technique (BOLT) for rapid and robust parameter estimation of electrochemical battery models. BOLT combines diversified candidate initialization, batch-parallel trust-region reflective (TRF) local refinement, JIT-accelerated model evaluation, and multi-condition consistency screening within a unified calibration workflow. Comparative experiments based on a grouped single-particle model and measured data from a commercial 18650 NMC lithium-ion cell show that BOLT achieves a favorable trade-off among voltage-response accuracy, computational efficiency, and repeated-run stability. BOLT(32) achieves an average mean absolute error of \(12.4 \pm 0.1\) mV over five operating conditions, requiring only \(20636 \pm 3081\) model calls and \(8.97 \pm 1.20\) s per run. Synthetic-data validation with a known parameter vector in the grouped SPM formulation further shows that BOLT recovers the reference parameter vector under model-consistent conditions and remains robust under 1--3 mV voltage-noise perturbations, with the mean parameter absolute relative error below \(0.6\%\). These results indicate that BOLT provides a practical calibration framework for BMS parameter updating, control-oriented battery digital twins, and second-life battery screening.

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

1 major / 1 minor

Summary. The manuscript proposes BOLT, a Batch-Optimized Local-to-Global Technique for rapid parameter estimation of electrochemical battery models. It integrates diversified candidate initialization, batch-parallel trust-region reflective (TRF) local optimization, JIT-accelerated evaluations, and multi-condition consistency screening. On a grouped single-particle model fitted to voltage data from a commercial 18650 NMC cell, BOLT(32) reports an average MAE of 12.4 ± 0.1 mV across five operating conditions, using 20636 ± 3081 model calls and 8.97 ± 1.20 s per run, with lower run-to-run variability than PSO and GA baselines. Synthetic-data recovery tests recover a known parameter vector with mean absolute relative error below 0.6% under 1–3 mV noise.

Significance. If the reported performance holds, BOLT supplies a practical, low-variability calibration workflow suitable for repeated BMS updates and control-oriented digital twins. The quantitative reporting of means and standard deviations over five runs, together with explicit synthetic recovery metrics, strengthens the efficiency and stability claims relative to standard global optimizers.

major comments (1)
  1. [Abstract] Abstract and the description of the multi-condition consistency screening: the central robustness claim rests on this screening step, yet its reliability is demonstrated only under the grouped SPM with the five chosen operating conditions and 1–3 mV noise. The synthetic recovery validates the workflow under model-consistent conditions but does not test whether the screening discards valid vectors or retains poor ones when the underlying model omits electrolyte dynamics, temperature effects, or when applied to different cells/chemistries.
minor comments (1)
  1. [Abstract] Abstract: the notation BOLT(32) appears without an immediate definition of the batch size or other parameter; this should be stated explicitly on first use.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive comment on the validation scope of the multi-condition consistency screening. We agree that the reported robustness is demonstrated specifically within the grouped SPM and the chosen conditions/noise levels, and we will revise the manuscript to clarify these boundaries.

read point-by-point responses

  1. Referee: [Abstract] Abstract and the description of the multi-condition consistency screening: the central robustness claim rests on this screening step, yet its reliability is demonstrated only under the grouped SPM with the five chosen operating conditions and 1–3 mV noise. The synthetic recovery validates the workflow under model-consistent conditions but does not test whether the screening discards valid vectors or retains poor ones when the underlying model omits electrolyte dynamics, temperature effects, or when applied to different cells/chemistries.

    Authors: We agree that the robustness claims for the multi-condition consistency screening are validated only under the grouped SPM formulation, the five operating conditions, and 1–3 mV synthetic noise. The synthetic recovery experiments confirm recovery of a known parameter vector under model-consistent conditions but do not address model mismatch (e.g., omitted electrolyte dynamics or temperature dependence) or transfer to other cell chemistries. In the revised manuscript we will (i) temper the abstract wording to specify that the reported stability and accuracy apply to the grouped SPM on the tested NMC data, and (ii) add an explicit limitations paragraph that states the current validation scope and identifies broader model classes and cell types as future work. No new experiments are required for this clarification. revision: yes

Circularity Check

0 steps flagged

No circularity: performance metrics are direct experimental measurements

full rationale

The paper's central results (MAE of 12.4 mV, model call counts, run times, and synthetic recovery errors) are obtained by running BOLT on measured cell data and synthetic trajectories generated from the grouped SPM. These quantities are not algebraically forced by the fitted parameters or by any self-citation chain; they are counted or averaged from the optimization runs themselves. The multi-condition consistency screening is a procedural filter whose correctness depends on modeling assumptions, but the reported numbers do not reduce to those assumptions by definition. No self-definitional loops, fitted-input predictions, or load-bearing self-citations appear in the derivation of the performance claims.

Axiom & Free-Parameter Ledger

1 free parameters · 0 axioms · 1 invented entities

The contribution is an engineering workflow rather than a mathematical derivation; it relies on standard optimization routines and the assumption that the grouped SPM is adequate for the tested conditions.

free parameters (1)
  • batch size = 32
    BOLT(32) variant uses a batch of 32 candidates; this hyperparameter is chosen by the authors to achieve the reported trade-off.
invented entities (1)
  • BOLT workflow no independent evidence
    purpose: Unified calibration procedure combining initialization, batch TRF, JIT, and consistency screening
    The named technique is the paper's main contribution; no external falsifiable prediction is supplied beyond the reported experiments.

pith-pipeline@v0.9.1-grok · 5849 in / 1363 out tokens · 34524 ms · 2026-06-30T01:24:56.161695+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

42 extracted references

  1. [1]

    Xiong, W

    R. Xiong, W. Sun, Q. Yu, F. Sun, Research progress, challenges and prospects of fault diagnosis on battery system of electric vehicles, Ap- plied Energy 279 (2020) 115855

    2020

  2. [2]

    Y. Wang, X. Zhang, K. Liu, Z. Wei, X. Hu, X. Tang, Z. Chen, System identification and state estimation of a reduced-order electrochemical model for lithium-ion batteries, Etransportation 18 (2023) 100295

    2023

  3. [3]

    Y. Li, X. Gao, X. Feng, X. Han, J. Du, L. Lu, M. Ouyang, Parameter- independent error correction for potential measurements by reference electrode in lithium-ion batteries, Journal of Energy Chemistry 67 (2022) 34–45. 37

    2022

  4. [4]

    Z.P.Cano, D.Banham, S.Ye, A.Hintennach, J.Lu, M.Fowler, Z.Chen, Batteries and fuel cells for emerging electric vehicle markets, Nature energy 3 (4) (2018) 279–289

    2018

  5. [5]

    Haghverdi, D

    K. Haghverdi, D. L. Danilov, G. Mulder, L. D. Couto, F. Guo, R.-A. Eichel, Review of the safe operating area (soa) for batteries: Evolv- ing concepts and structured frameworks, Journal of Power Sources 652 (2025) 237483

    2025

  6. [6]

    Q. Qi, W. Liu, Z. Deng, J. Li, Z. Song, X. Hu, Battery pack capacity estimation for electric vehicles based on enhanced machine learning and field data, Journal of Energy Chemistry 92 (2024) 605–618

    2024

  7. [7]

    Z. M. Konz, B. M. Wirtz, A. Verma, T.-Y. Huang, H. K. Bergstrom, M. J. Crafton, D. E. Brown, E. J. McShane, A. M. Colclasure, B. D. McCloskey, High-throughput li plating quantification for fast-charging battery design, Nature Energy 8 (5) (2023) 450–461

    2023

  8. [8]

    X. Hu, L. Xu, X. Lin, M. Pecht, Battery lifetime prognostics, Joule 4 (2) (2020) 310–346

    2020

  9. [9]

    Y.Gao, C.Liu, S.Chen, X.Zhang, G.Fan, C.Zhu, Developmentandpa- rameterization of a control-oriented electrochemical model of lithium-ion batteries for battery-management-systems applications, Applied Energy 309 (2022) 118521

    2022

  10. [10]

    Dubarry, D

    M. Dubarry, D. Howey, B. Wu, Enabling battery digital twins at the industrial scale, Joule 7 (6) (2023) 1134–1144

    2023

  11. [11]

    Naseri, S

    F. Naseri, S. Gil, C. Barbu, E. Cetkin, G. Yarimca, A. Jensen, P. G. Larsen, C. Gomes, Digital twin of electric vehicle battery systems: Com- prehensive review of the use cases, requirements, and platforms, Renew- able and Sustainable Energy Reviews 179 (2023) 113280

    2023

  12. [12]

    J. Song, R. C. Ihuaenyi, J. Lim, Z. Wang, W. Li, R. Fang, A. K. Ghamsari, H. Xu, Y. M. Lee, J. Zhu, A microstructural electrochemo- mechanical model of high-nickel composite electrodes towards digital twins to bridge the particle and electrode-level characterizations, En- ergy & Environmental Science 18 (7) (2025) 3129–3147. 38

    2025

  13. [13]

    Hafner, J

    D. Hafner, J. Pasukonis, J. Ba, T. Lillicrap, Mastering diverse control tasks through world models, Nature (2025) 1–7

    2025

  14. [14]

    C. Zhu, J. Han, G. Fan, K. Li, B. Guo, X. Zhang, K. Liu, Reduced- order multi-particle model of lithium-ion batteries capturing electro- chemical heterogeneity at subzero temperatures, Applied Energy 410 (2026) 127488

    2026

  15. [15]

    H. A. A. Ali, L. H. Raijmakers, K. Chayambuka, D. L. Danilov, P. H. Notten, R.-A. Eichel, A comparison between physics-based li-ion battery models, Electrochimica Acta 493 (2024) 144360

    2024

  16. [16]

    M. Mama, E. Solai, T. Capurso, A. Danlos, S. Khelladi, Comprehen- sive review of multi-scale lithium-ion batteries modeling: From electro- chemical dynamics up to heat transfer in battery thermal management system, Energy Conversion and Management 325 (2025) 119223

    2025

  17. [17]

    Miguel, G

    E. Miguel, G. L. Plett, M. S. Trimboli, L. Oca, U. Iraola, E. Bekaert, Re- view of computational parameter estimation methods for electrochemi- cal models, Journal of Energy Storage 44 (2021) 103388

    2021

  18. [18]

    Y. Li, G. Liu, W. Deng, Z. Li, Comparative study on parameter identi- fication of an electrochemical model for lithium-ion batteries via meta- heuristic methods, Applied Energy 367 (2024) 123437

    2024

  19. [19]

    S. Ding, Y. Li, H. Dai, L. Wang, X. He, Accurate model parameter identification to boost precise aging prediction of lithium-ion batteries: A review, Advanced Energy Materials 13 (39) (2023) 2301452

    2023

  20. [20]

    X. Cai, C. Zhang, Z. Chen, L. Zhang, D. U. Sauer, W. Li, Characteri- zation and quantification of multi-field coupling in lithium-ion batteries under mechanical constraints, Journal of Energy Chemistry 95 (2024) 364–379

    2024

  21. [21]

    T. F. Fuller, M. Doyle, J. Newman, Simulation and optimization of the dual lithium ion insertion cell, Journal of the electrochemical society 141 (1) (1994) 1

    1994

  22. [22]

    S. G. Marquis, V. Sulzer, R. Timms, C. P. Please, S. J. Chapman, An asymptotic derivation of a single particle model with electrolyte, Journal of The Electrochemical Society 166 (15) (2019) A3693. 39

    2019

  23. [23]

    B. S. Haran, B. N. Popov, R. E. White, Determination of the hydro- gen diffusion coefficient in metal hydrides by impedance spectroscopy, Journal of Power Sources 75 (1) (1998) 56–63

    1998

  24. [24]

    F. Guo, L. D. Couto, G. Mulder, K. Trad, G. Hu, O. Capron, K. Haghverdi, A systematic review of electrochemical model-based lithium-ion battery state estimation in battery management systems, Journal of Energy Storage 101 (2024) 113850

    2024

  25. [25]

    Ecker, T

    M. Ecker, T. K. D. Tran, P. Dechent, S. Käbitz, A. Warnecke, D. U. Sauer, Parameterization of a physico-chemical model of a lithium-ion battery: I. determination of parameters, Journal of The Electrochemical Society 162 (9) (2015) A1836

    2015

  26. [26]

    J. Lu, T. Wu, K. Amine, State-of-the-art characterization techniques for advanced lithium-ion batteries, Nature Energy 2 (3) (2017) 1–13

    2017

  27. [27]

    Wang, J.-N

    Q.-K. Wang, J.-N. Shen, Z.-F. Ma, Y.-J. He, Decoupling parameter estimation strategy based electrochemical-thermal coupled modeling method for large format lithium-ion batteries with internal tempera- ture experimental validation, Chemical Engineering Journal 424 (2021) 130308

    2021

  28. [28]

    Rojas, L

    C. Rojas, L. Oca, I. Lopetegi, U. Iraola, J. Carrasco, A critical look at efficient parameter estimation methodologies of electrochemical models for lithium-ion cells, Journal of Energy Storage 80 (2024) 110384

    2024

  29. [29]

    Z. Deng, H. Deng, L. Yang, Y. Cai, X. Zhao, Implementation of reduced- order physics-based model and multi-parameters identification strategy for lithium-ion battery, Energy 138 (2017) 509–519

    2017

  30. [30]

    M. A. Rahman, S. Anwar, A. Izadian, Electrochemical model parameter identification of a lithium-ion battery using particle swarm optimization method, Journal of Power Sources 307 (2016) 86–97

    2016

  31. [31]

    Xiong, L

    R. Xiong, L. Li, Z. Li, Q. Yu, H. Mu, An electrochemical model based degradation state identification method of lithium-ion battery for all- climate electric vehicles application, Applied energy 219 (2018) 264–275

    2018

  32. [32]

    F. Guo, L. D. Couto, G. Thenaisie, Efficiency and optimality in electro- chemical battery model parameter identification: A comparative study 40 of estimation techniques, in: 2024 10th International Conference on Op- timization and Applications (ICOA), IEEE, 2024, pp. 1–6

    2024

  33. [33]

    A. Tian, K. Dong, X.-G. Yang, Y. Wang, L. He, Y. Gao, J. Jiang, Physics-based parameter identification of an electrochemical model for lithium-ion batteries with two-population optimization method, Applied Energy 378 (2025) 124748

    2025

  34. [34]

    Huang, N

    S. Huang, N. Lin, Z. Wang, Z. Zhang, S. Wen, Y. Zhao, Q. Li, A novel data-driven method for online parameter identification of an electro- chemical model based on cuckoo search and particle swarm optimization algorithm, Journal of Power Sources 601 (2024) 234261

    2024

  35. [35]

    M. Kim, H. Chun, J. Kim, K. Kim, J. Yu, T. Kim, S. Han, Data- efficient parameter identification of electrochemical lithium-ion battery model using deep bayesian harmony search, Applied Energy 254 (2019) 113644

    2019

  36. [36]

    Wang, Y.-B

    B.-C. Wang, Y.-B. He, J. Liu, B. Luo, Fast parameter identification of lithium-ion batteries via classification model-assisted bayesian optimiza- tion, Energy 288 (2024) 129667

    2024

  37. [37]

    Zhang, Y

    W. Zhang, Y. Xie, H. He, Z. Long, L. Zhuang, J. Zhou, Multi-physics coupling model parameter identification of lithium-ion battery based on data driven method and genetic algorithm, Energy 314 (2025) 134120

    2025

  38. [38]

    F. Guo, L. D. Couto, Comparative performance analysis of numerical discretization methods for electrochemical model of lithium-ion batter- ies, Journal of Power Sources 650 (2025) 237365

    2025

  39. [39]

    F. Guo, L. D. Couto, A control-oriented simplified single particle model with grouped parameter and sensitivity analysis for lithium-ion batter- ies, Journal of Power Sources 649 (2025) 237309

    2025

  40. [40]

    F. Guo, L. D. Couto, Cpg-spmt: Control-oriented parameter-grouped single particle model with thermal effects for lithium-ion batteries, Com- puter Physics Communications (2026) 110075

    2026

  41. [41]

    M. A. Branch, T. F. Coleman, Y. Li, A subspace, interior, and con- jugate gradient method for large-scale bound-constrained minimization problems, SIAM Journal on Scientific Computing 21 (1) (1999) 1–23. 41

    1999

  42. [42]

    F. Guo, L. D. Couto, K. Trad, G. Mulder, K. Haghverdi, G. Thenaisie, Optimizing parameter estimation for electrochemical battery model: A comparative analysis of operating profiles on computational efficiency and accuracy, Journal of Power Sources 665 (2026) 239044. 42

    2026