pith. sign in

arxiv: 2508.08578 · v2 · submitted 2025-08-12 · 📡 eess.SY · cs.SY

A Data-Driven Optimal Control Architecture for Grid-Connected Power Converters

Ruohan Leng , Linbin Huang , Huanhai Xin , Ping Ju , Xiongfei Wang , Eduardo Prieto-Araujo , Florian D\"orfler This is my paper

Pith reviewed 2026-05-18 23:49 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords data-driven controlDeePCgrid-connected convertersoptimal controladaptive controlrobust controlpower electronicsrenewable integration
0
0 comments X

The pith

Grid-connected power converters can achieve optimal control by learning grid behavior directly from measured input-output data instead of using simplified models.

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

This paper establishes that data-enabled predictive control can operate grid-connected power converters without relying on presumed models of the power grid. Traditional PID regulators tuned to simplified models often perform poorly or become unstable when the real grid is complex and variable. The authors introduce DeePConverters that use collected measurements to implicitly sense grid characteristics and optimize control actions in real time. A reader would care because these converters interface renewables, storage, and other devices into modern power systems where accurate grid models are hard to obtain and maintain.

Core claim

We employ a data-enabled predictive control (DeePC) to perform data-driven, optimal, robust, and adaptive control for power converters. We call the converters that are operated in this way DeePConverters. A DeePConverter can implicitly perceive the characteristics of the power grid from measured data and adjust its control strategy to achieve optimal, robust, and adaptive performance. The paper presents the modular configurations, generalized structure, control behavior specification, inherent robustness, detailed implementation, computational aspects, and online adaptation of DeePConverters, validated through high-fidelity simulations and hardware-in-the-loop tests.

What carries the argument

The DeePConverter, which applies data-enabled predictive control to use past input-output measurements for formulating and solving an online optimization problem that determines future control inputs.

If this is right

  • The converter maintains desired performance across grid variations without manual retuning of parameters.
  • Robustness to modeling errors and uncertainties arises inherently from the data-driven formulation.
  • Online adaptation updates the control law continuously as new measurements arrive.
  • The same architecture supports multiple objectives such as grid synchronization and active/reactive power regulation through modular configuration.

Where Pith is reading between the lines

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

  • Widespread adoption could simplify commissioning of new renewable plants by reducing dependence on detailed grid studies.
  • The approach may extend naturally to coordinated control of multiple converters sharing the same unknown grid.
  • Computational cost of solving the data-driven optimization at each step would need to remain compatible with typical converter sampling rates.

Load-bearing premise

That measured input-output data from the converter is sufficient to capture the relevant dynamics of the unknown, complex, and variable power grid for the purposes of DeePC optimization and online adaptation.

What would settle it

A hardware-in-the-loop experiment in which the DeePConverter loses stability or fails to regulate power when grid impedance or topology changes in a manner absent from the initial data collection window.

Figures

Figures reproduced from arXiv: 2508.08578 by Eduardo Prieto-Araujo, Florian D\"orfler, Huanhai Xin, Linbin Huang, Ping Ju, Ruohan Leng, Xiongfei Wang.

Figure 1
Figure 1. Figure 1: One-line diagram of a three-phase power converter with conventional [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Control block diagrams of the integral DeePConverters: Full and [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Control signal generation for three converter types: mapping observed trajectory matrix to [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Time-domain responses of a single-converter system with initial [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Time-domain responses of a single-converter system with [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: A two-area test system [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Time-domain responses of the two-area test system under four cases. [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 9
Figure 9. Figure 9: During the data-collection phase, band-limited white [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 9
Figure 9. Figure 9: Time-domain results of the HIL simulation: (a) Responses of Converter 1 during the data-collection period with injected white-noise signals; (b) [PITH_FULL_IMAGE:figures/full_fig_p010_9.png] view at source ↗
read the original abstract

Grid-connected power converters are ubiquitous in modern power systems, acting as grid interfaces of renewable energy sources, energy storage systems, electric vehicles, high-voltage DC systems, etc. Conventionally, power converters use multiple PID regulators to achieve different control objectives such as grid synchronization and voltage/power regulation, where the PID parameters are usually tuned based on a presumed (and often overly-simplified) power grid model. However, this may lead to inferior performance or even instabilities in practice, as the real power grid is highly complex, variable, and generally unknown. To tackle this problem, we employ a data-enabled predictive control (DeePC) to perform data-driven, optimal, robust, and adaptive control for power converters. We call the converters that are operated in this way DeePConverters. A DeePConverter can implicitly perceive the characteristics of the power grid from measured data and adjust its control strategy to achieve optimal, robust, and adaptive performance. We present the modular configurations, generalized structure, control behavior specification, inherent robustness, detailed implementation, computational aspects, and online adaptation of DeePConverters. High-fidelity simulations and hardware-in-the-loop (HIL) tests are provided to validate the effectiveness of DeePConverters.

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

2 major / 2 minor

Summary. The paper proposes DeePConverters, a data-driven control architecture for grid-connected power converters based on Data-enabled Predictive Control (DeePC). It argues that by using measured input-output trajectories, the converter can implicitly capture unknown and variable grid dynamics to achieve optimal, robust, and adaptive performance without relying on explicit models or PID tuning. The manuscript covers modular configurations, generalized structure, control specifications, robustness properties, implementation details, computational aspects, and online adaptation, with validation claimed via high-fidelity simulations and hardware-in-the-loop (HIL) tests.

Significance. If the central claims hold, the work could offer a practical model-free alternative for controlling converters interfacing renewables and storage in complex grids, potentially improving stability where traditional model-based methods fail due to grid variability. The emphasis on modular structure and online adaptation is a constructive contribution to the DeePC literature in power electronics, though the approach retains free parameters such as prediction horizon and data window lengths.

major comments (2)
  1. [Abstract] Abstract: The validation claim rests on 'high-fidelity simulations and HIL tests' yet provides no quantitative error metrics, performance indices, or comparison baselines against conventional PID or model-based controllers. This absence makes it impossible to assess whether the DeePC-based strategy delivers measurable improvements in robustness or adaptation under grid variations.
  2. [Online adaptation] Online adaptation section: The description of online adaptation does not specify the mechanism for refreshing the data window or ensuring persistency of excitation when the grid experiences abrupt changes (e.g., line switching or distant load steps). Without this, the Hankel-based predictor may lose accuracy, directly undermining the claim that local I/O data suffices to encode the effective grid impedance for stable QP solutions.
minor comments (2)
  1. [Implementation] Notation for the data matrices and behavioral model could be clarified with explicit definitions of the Hankel matrix dimensions and the role of the regularization parameters in the DeePC optimization.
  2. [Computational aspects] The manuscript would benefit from a dedicated table summarizing the free parameters (prediction horizon, data window length) and their tuning guidelines or sensitivity analysis.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback on our manuscript. We have addressed each major comment point by point below, agreeing where revisions are needed to improve clarity and completeness.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The validation claim rests on 'high-fidelity simulations and HIL tests' yet provides no quantitative error metrics, performance indices, or comparison baselines against conventional PID or model-based controllers. This absence makes it impossible to assess whether the DeePC-based strategy delivers measurable improvements in robustness or adaptation under grid variations.

    Authors: We agree that the abstract would be strengthened by including explicit quantitative metrics and direct comparisons. While the full manuscript contains simulation and HIL results illustrating performance under grid variations (including tracking behavior and adaptation), we did not tabulate error metrics or baselines against PID/model-based methods. In the revised version we will update the abstract with key indices such as RMS tracking error and settling time, and add a comparison subsection with quantitative results against conventional controllers. revision: yes

  2. Referee: [Online adaptation] Online adaptation section: The description of online adaptation does not specify the mechanism for refreshing the data window or ensuring persistency of excitation when the grid experiences abrupt changes (e.g., line switching or distant load steps). Without this, the Hankel-based predictor may lose accuracy, directly undermining the claim that local I/O data suffices to encode the effective grid impedance for stable QP solutions.

    Authors: We thank the referee for highlighting this point. The online adaptation section explains that the data window is refreshed with new input-output measurements to capture time-varying grid dynamics, but we acknowledge that the precise update rule, conditions for maintaining persistency of excitation, and handling of abrupt events (such as line switching) are not detailed enough. In the revision we will expand this section with the explicit data-window refresh algorithm, a discussion of persistency requirements, and additional analysis or pseudocode showing that the resulting QP remains well-posed and stable. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation applies established DeePC to new domain

full rationale

The paper presents a data-driven DeePC architecture for grid-connected converters, collecting measured I/O trajectories to construct the behavioral model and solve the resulting QP for control inputs. This follows the standard non-parametric DeePC formulation from prior literature without reducing any claimed prediction to a fitted parameter or self-referential definition within the paper's equations. Validation via high-fidelity simulations and HIL tests provides external checks independent of the derivation chain. Self-references to DeePC are to an established method with independent mathematical foundations (behavioral systems theory) and do not bear the load of proving the specific converter application. No load-bearing step equates outputs to inputs by construction.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that DeePC can be directly transferred to converter control with sufficient data richness; no new entities are postulated, but standard DeePC hyperparameters (horizons, regularization) function as free parameters.

free parameters (1)
  • prediction horizon and data window lengths
    Typical DeePC tuning parameters that must be chosen and affect performance; not specified in abstract.
axioms (1)
  • domain assumption Input-output data collected from the converter-grid system is persistently exciting and representative of future behavior.
    Fundamental requirement for DeePC to produce meaningful predictions; invoked implicitly when claiming implicit perception of grid characteristics.

pith-pipeline@v0.9.0 · 5767 in / 1296 out tokens · 31487 ms · 2026-05-18T23:49:06.554719+00:00 · methodology

discussion (0)

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

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Model-Free Power System Stability Enhancement with Dissipativity-Based Neural Control

    eess.SY 2025-10 unverdicted novelty 5.0

    Neural networks learn dissipativity matrices from data to create a model-free controller that improves transient stability in all-VSG power systems.

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages · cited by 1 Pith paper

  1. [1]

    Status of renewable energy grid integration in 2024,

    N. E. A. of China, “Status of renewable energy grid integration in 2024,” NEA Online Reports, 2025, available: https://www.nea.gov.cn/20250221/ e10f363cabe3458aaf78ba4558970054/c.html

    work page arXiv 2024

  2. [2]

    Communication from the commission to the european parliament, the european council, the council, the european economic and social committee and the committee of the regions,

    “Communication from the commission to the european parliament, the european council, the council, the european economic and social committee and the committee of the regions,”REPowerEU Plan, 2022

    work page 2022

  3. [3]

    Foundations and challenges of low-inertia systems,

    F. Milano, F. D ¨orfler, G. Hug, D. J. Hill, and G. Verbiˇc, “Foundations and challenges of low-inertia systems,” in2018 power systems computation conference (PSCC). IEEE, 2018, pp. 1–25

    work page 2018

  4. [4]

    Control of power converters in AC microgrids,

    J. Rocabert, A. Luna, F. Blaabjerg, and P. Rodriguez, “Control of power converters in AC microgrids,”IEEE Trans. Power Electron., vol. 27, no. 11, pp. 4734–4749, 2012

    work page 2012

  5. [5]

    Grid-synchronization stability of converter-based resources—an overview,

    X. Wang, M. G. Taul, H. Wu, Y . Liao, F. Blaabjerg, and L. Harne- fors, “Grid-synchronization stability of converter-based resources—an overview,”IEEE Open J. Ind. Appl., vol. 1, pp. 115–134, 2020

    work page 2020

  6. [6]

    Impedance-based stability criterion for grid-connected invert- ers,

    J. Sun, “Impedance-based stability criterion for grid-connected invert- ers,”IEEE Trans. Power Electron., vol. 26, no. 11, pp. 3075–3078, 2011

    work page 2011

  7. [7]

    Unified impedance model of grid-connected voltage-source converters,

    X. Wang, L. Harnefors, and F. Blaabjerg, “Unified impedance model of grid-connected voltage-source converters,”IEEE Trans. Power Electron., vol. 33, no. 2, pp. 1775–1787, 2017

    work page 2017

  8. [8]

    Overview of control and grid synchronization for distributed power generation systems,

    F. Blaabjerget al., “Overview of control and grid synchronization for distributed power generation systems,”IEEE Trans. Ind. Electron., vol. 53, no. 5, pp. 1398–1409, 2006

    work page 2006

  9. [9]

    A virtual synchronous machine implementation for distributed control of power converters in smartgrids,

    S. D’Arco, J. A. Suul, and O. B. Fosso, “A virtual synchronous machine implementation for distributed control of power converters in smartgrids,”Elect. Power Syst. Res., vol. 122, pp. 180–197, 2015

    work page 2015

  10. [10]

    Subsynchronous interaction between direct-drive PMSG based wind farms and weak AC networks,

    H. Liuet al., “Subsynchronous interaction between direct-drive PMSG based wind farms and weak AC networks,”IEEE Trans. Power Syst., vol. 32, no. 6, pp. 4708–4720, 2017

    work page 2017

  11. [11]

    Wind energy systems subsyn- chronous oscillations: Events and modeling,

    Y . Cheng, L. Fan, and J. Roseet al., “Wind energy systems subsyn- chronous oscillations: Events and modeling,”IEEE Power and Energy Society, Technical Report, PES-TR80, 2020

    work page 2020

  12. [12]

    A note on persistency of excitation,

    J. C. Willemset al., “A note on persistency of excitation,”Systems & Control Letters, vol. 54, no. 4, pp. 325–329, 2005

    work page 2005

  13. [13]

    Data-driven control based on the behavioral approach: From theory to applications in power systems,

    I. Markovsky, L. Huang, and F. D ¨orfler, “Data-driven control based on the behavioral approach: From theory to applications in power systems,” IEEE Control Syst. Mag., vol. 43, no. 5, pp. 28–68, 2023

    work page 2023

  14. [14]

    Data-enabled predictive con- trol: In the shallows of the DeePC,

    J. Coulson, J. Lygeros, and F. D ¨orfler, “Data-enabled predictive con- trol: In the shallows of the DeePC,” in2019 18th European Control Conference (ECC). IEEE, 2019, pp. 307–312

    work page 2019

  15. [15]

    Formulas for data-driven control: Stabilization, optimality, and robustness,

    C. De Persis and P. Tesi, “Formulas for data-driven control: Stabilization, optimality, and robustness,”IEEE Trans. Autom. Control, vol. 65, no. 3, pp. 909–924, 2019

    work page 2019

  16. [16]

    Data-driven model predictive control with stability and robustness guarantees,

    J. Berberich, J. K ¨ohler, M. A. Muller, and F. Allgower, “Data-driven model predictive control with stability and robustness guarantees,”IEEE Trans. Autom. Control, vol. 66, no. 4, pp. 1702–1717, 2021

    work page 2021

  17. [17]

    Robust data-enabled predictive control: Tractable formulations and performance guarantees,

    L. Huang, J. Zhen, J. Lygeros, and F. D ¨orfler, “Robust data-enabled predictive control: Tractable formulations and performance guarantees,” IEEE Trans. Autom. Control, vol. 68, no. 5, pp. 3163–3170, 2023

    work page 2023

  18. [18]

    Data-enabled predictive control for grid-connected power converters,

    L. Huanget al., “Data-enabled predictive control for grid-connected power converters,” in2019 IEEE 58th Conference on Decision and Control (CDC). IEEE, 2019, pp. 8130–8135

    work page 2019

  19. [19]

    Decentralized data- enabled predictive control for power system oscillation damping,

    L. Huang, J. Coulson, J. Lygeros, and F. D ¨orfler, “Decentralized data- enabled predictive control for power system oscillation damping,”IEEE Trans. Control Syst. Technol., vol. 30, no. 3, pp. 1065–1077, 2021

    work page 2021

  20. [20]

    Data-driven predictive current control for syn- chronous motor drives,

    P. G. Carletet al., “Data-driven predictive current control for syn- chronous motor drives,” in2020 IEEE Energy Conversion Congress and Exposition (ECCE). IEEE, 2020, pp. 5148–5154

    work page 2020

  21. [21]

    Adaptive robust data-driven building control via bilevel reformulation: An experimental result,

    Y . Lian, J. Shi, M. Koch, and C. N. Jones, “Adaptive robust data-driven building control via bilevel reformulation: An experimental result,”IEEE Trans. Control Syst. Technol., vol. 31, no. 6, pp. 2420–2436, 2023

    work page 2023

  22. [22]

    Data- enabled predictive control for quadcopters,

    E. Elokda, J. Coulson, P. N. Beuchat, J. Lygeros, and F. D ¨orfler, “Data- enabled predictive control for quadcopters,”International Journal of Robust and Nonlinear Control, vol. 31, no. 18, pp. 8916–8936, 2021

    work page 2021

  23. [23]

    Toward a data-driven template model for quadrupedal locomotion,

    R. T. Fawcett, K. Afsari, A. D. Ames, and K. A. Hamed, “Toward a data-driven template model for quadrupedal locomotion,”IEEE Robotics and Automation Letters, vol. 7, no. 3, pp. 7636–7643, 2022

    work page 2022

  24. [24]

    Data-driven control for power oscillation damping: Robustness and closed-loop identification,

    F. Braun, J. Poland, and M. Larsson, “Data-driven control for power oscillation damping: Robustness and closed-loop identification,” in2023 IEEE Belgrade PowerTech. IEEE, 2023, pp. 01–06

    work page 2023

  25. [25]

    Quadratic regularization of data-enabled predictive control: Theory and application to power converter experiments,

    L. Huang, J. Zhen, J. Lygeros, and F. D ¨orfler, “Quadratic regularization of data-enabled predictive control: Theory and application to power converter experiments,”19th IFAC Symposium on System Identification, vol. 54, no. 7, pp. 192–197, 2021

    work page 2021

  26. [26]

    Data-driven simulation and control,

    I. Markovsky and P. Rapisarda, “Data-driven simulation and control,” International Journal of Control, vol. 81, no. 12, pp. 1946–1959, 2008

    work page 1946

  27. [27]

    Identifiability in the behavioral setting,

    I. Markovsky and F. D ¨orfler, “Identifiability in the behavioral setting,” 2020, available online

    work page 2020

  28. [28]

    Handbook of linear data- driven predictive control: Theory, implementation and design,

    P. C. N. Verheijen, V . Breschi, and M. Lazar, “Handbook of linear data- driven predictive control: Theory, implementation and design,”Annual Reviews in Control, vol. 56, p. 100914, 2023

    work page 2023

  29. [29]

    Robust output-feedback mpc with integral action,

    C. Løvaaset al., “Robust output-feedback mpc with integral action,” IEEE Trans. Autom. Control, vol. 55, no. 7, pp. 1531–1543, 2010

    work page 2010

  30. [30]

    How many grid-forming converters do we need? a perspective from small signal stability and power grid strength,

    H. Xinet al., “How many grid-forming converters do we need? a perspective from small signal stability and power grid strength,”IEEE Trans. Power Syst., pp. 1–13, 2024

    work page 2024

  31. [31]

    Modeling of a virtual synchronous machine- based grid-interface converter for renewable energy systems integration,

    I. Cvetkovicet al., “Modeling of a virtual synchronous machine- based grid-interface converter for renewable energy systems integration,” in2014 IEEE 15th Workshop on Control and Modeling for Power Electronics (COMPEL). IEEE, 2014, pp. 1–7

    work page 2014

  32. [32]

    Osqp: An operator splitting solver for quadratic programs., 2020, 12,

    B. Stellato, G. Banjac, P. Goulart, A. Bemporad, and S. B. OSQP, “Osqp: An operator splitting solver for quadratic programs., 2020, 12,”Math. Program. Comput., vol. 12, pp. 1–36, 2020

    work page 2020