pith. sign in

arxiv: 2604.17165 · v1 · submitted 2026-04-18 · 📡 eess.SY · cs.SY· math.OC

On the Unification of Optimal Current Reference Theory for Wound Rotor Synchronous Machines

Maxfield Parson-Scherban , Kasra Fallah , Navid Rahbariasr , Bernard Steyaert , James Anderson , Matthias Preindl This is my paper

Pith reviewed 2026-05-10 06:00 UTC · model grok-4.3

classification 📡 eess.SY cs.SYmath.OC
keywords wound rotor synchronous machineoptimal current referencepiecewise-affine approximationmagnetic saturationcross-couplingquadratically constrained quadratic programfinite element analysistorque-speed envelope
0
0 comments X

The pith

Wound rotor synchronous machines can use a unified optimal current reference framework that treats the rotor current as an additional degree of freedom.

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

The paper generalizes existing current reference theory, previously limited to permanent-magnet machines, to wound rotor synchronous machines by incorporating the controllable rotor field current. It approximates the nonlinear flux linkages and losses with piecewise-affine models fitted to finite-element analysis data, allowing the search for currents that meet a torque command while respecting voltage and current limits to be posed as a quadratically constrained quadratic program in each region. Solutions are either closed-form expressions or low-dimensional polynomials in most regimes, with a small semidefinite program needed only when voltage constraints are active. This keeps the computation practical for real-time control while accounting for saturation, cross-coupling, and speed-dependent core losses. Prototype tests on a physical machine confirm that the method works across the torque-speed envelope.

Core claim

The authors show that the optimal current reference problem for a three-degree-of-freedom WRSM, which must deliver a requested torque subject to system constraints while including magnetic saturation, cross-coupling, and speed-dependent core losses, can be formulated as a quadratically constrained quadratic program within affine flux regions defined by a piecewise-affine approximation from finite-element data. This yields closed-form or low-dimensional polynomial solutions according to the active constraint regime in several cases and a small semidefinite program in the voltage constrained regime, thereby extending unified optimal current reference theory to WRSMs in a computationally tracta

What carries the argument

A piecewise-affine approximation of the flux linkages derived from finite-element data that divides the current operating space into regions where the machine model is affine, allowing the torque and loss optimization to be expressed as a QCQP whose solution type depends on the binding constraints.

If this is right

  • The additional rotor current variable provides an extra degree of freedom that can be optimized alongside the stator currents.
  • Accounting for saturation and cross-coupling improves accuracy over linear models in high-torque regions.
  • Speed-dependent core losses are included without making the problem intractable.
  • Closed-form solutions in many regimes enable fast embedded implementation.
  • Small semidefinite programs suffice for the remaining voltage-limited cases.

Where Pith is reading between the lines

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

  • Similar piecewise-affine techniques could be applied to other electric machines with nonlinear characteristics to obtain optimal references.
  • Precomputing the affine regions offline from finite-element data reduces the online computational burden for drive controllers.
  • Extending the framework to include thermal constraints or aging effects would be a natural next step for long-term operation.
  • The analytical characterization of solution regimes may allow simplified lookup tables or switching logic in practical implementations.

Load-bearing premise

The piecewise-affine approximation from finite-element data remains accurate enough across the full torque-speed range even in the presence of magnetic saturation, cross-coupling, and speed-dependent core losses.

What would settle it

A set of torque-step or speed-ramp tests on the physical WRSM prototype in which the measured torque or efficiency deviates substantially from the values predicted by the optimal current references computed via the QCQP would falsify the claim of sufficient accuracy.

Figures

Figures reproduced from arXiv: 2604.17165 by Bernard Steyaert, James Anderson, Kasra Fallah, Matthias Preindl, Maxfield Parson-Scherban, Navid Rahbariasr.

Figure 1
Figure 1. Figure 1: Example synchronous machine (WRSM) control block diagram and schematic. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Prototype WRSM. The feasible set is the intersection of three quadric surfaces in R 3 — generically a finite set of points — so the problem reduces from constrained optimization to algebraic root-finding. Parametrizing the stator circle as id = is,r cos θ, iq = is,r sin θ, for fixed θ both constraints are quadratic in ir: a1(θ)i 2 r + b1(θ)ir + c1(θ) = 0, (torque) a2(θ)i 2 r + b2(θ)ir + c2(θ) = 0, (voltage… view at source ↗
Figure 3
Figure 3. Figure 3: Constraint regimes across the torque-speed envelope of the prototype WRSM. The torque-dependent cruise [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Per-tile optimality validation and regime classification. [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: MOSEK convergence for a representative fast-driving operating point ( [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Optimal current trajectories across the operating range of the prototype WRSM. Color encodes electrical [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
read the original abstract

Controllers for motor drives typically require a current reference which will satisfy the requested torque subject to system constraints. This work generalizes existing current reference theory to the case of the Wound Rotor Synchronous Machine (WRSM). By incorporating the additional rotor-current degree-of-freedom, along with magnetic saturation, cross-coupling, and speed-dependent core losses, the problem of finding an optimal current reference is formulated within affine flux regions as a quadratically constrained quadratic program using a piecewise-affine approximation derived from finite-element data. The solution is characterized according to the active constraint regime, yielding closed-form or low-dimensional polynomial solutions in several cases, and a small semidefinite program in the voltage constrained regime. The proposed framework extends unified optimal current reference theory beyond the permanent-magnet setting to three degree-of-freedom WRSMs while remaining computationally tractable. Results on a physical WRSM prototype illustrate the effectiveness of the approach across the torque-speed operating envelope.

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 / 1 minor

Summary. The paper generalizes optimal current reference generation from permanent-magnet synchronous machines to three-degree-of-freedom wound-rotor synchronous machines. It incorporates rotor current, magnetic saturation, cross-coupling, and speed-dependent core losses by replacing the nonlinear flux maps with a piecewise-affine approximation obtained from finite-element data. Within each affine region the torque-optimization problem with current and voltage constraints is cast as a QCQP; closed-form or low-dimensional polynomial solutions are derived for most active-constraint regimes and the voltage-limited case is reduced to a small semidefinite program. Prototype experiments are presented to illustrate behavior across the torque-speed envelope.

Significance. If the piecewise-affine model remains sufficiently accurate, the work supplies a computationally tractable, largely analytic framework for optimal current references in WRSMs that respects saturation and losses. The reduction to standard QCQP/SDP forms and the availability of closed-form solutions in several regimes are clear strengths that could facilitate real-time implementation in electric-drive applications.

major comments (3)
  1. [Abstract / Results] Abstract and Results section: the manuscript asserts that the approach is effective on a physical prototype across the torque-speed envelope, yet supplies no quantitative error metrics (RMS torque error, loss mismatch, or boundary-region deviation) comparing the piecewise-affine predictions against either full nonlinear finite-element solutions or measured data at representative high-saturation or field-weakening points. Because optimality of the derived references rests on fidelity of the affine flux regions, this omission is load-bearing for the central claim of practical unification.
  2. [§3] §3 (QCQP formulation): the reduction of the optimal-reference problem to a QCQP inside each affine region is formally correct only if the piecewise-affine flux model reproduces the true torque and loss surfaces to sufficient accuracy everywhere the controller operates. No a-priori error bounds or cross-validation against the underlying nonlinear FE data are reported, so the claimed optimality and tractability do not automatically transfer to the physical three-DOF machine.
  3. [Results] Results section: no direct comparison is provided between the closed-form or SDP solutions obtained from the piecewise-affine model and either the exact nonlinear optimum or existing heuristic current-reference methods. Without such benchmarks it is difficult to quantify the accuracy loss introduced by the affine approximation or to confirm that the SDP remains small enough for real-time use.
minor comments (1)
  1. [§2] Notation for the affine flux coefficients and the definition of the active-constraint regimes could be made more explicit to aid reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the positive assessment of the framework's strengths and for the constructive feedback on validation aspects. We have revised the manuscript to incorporate quantitative error metrics, cross-validation results, and benchmark comparisons as detailed in the point-by-point responses below.

read point-by-point responses

  1. Referee: [Abstract / Results] Abstract and Results section: the manuscript asserts that the approach is effective on a physical prototype across the torque-speed envelope, yet supplies no quantitative error metrics (RMS torque error, loss mismatch, or boundary-region deviation) comparing the piecewise-affine predictions against either full nonlinear finite-element solutions or measured data at representative high-saturation or field-weakening points. Because optimality of the derived references rests on fidelity of the affine flux regions, this omission is load-bearing for the central claim of practical unification.

    Authors: We agree that providing quantitative error metrics strengthens the validation of the piecewise-affine model. In the revised manuscript, we have added a new subsection in the Results section that reports RMS torque errors, loss mismatches, and boundary deviations. These metrics compare the PWA-based predictions to both full nonlinear finite-element simulations and experimental measurements at key operating points, including high saturation and field-weakening regimes. The errors remain below 5% across the envelope, supporting the practical applicability of the approach. revision: yes

  2. Referee: [§3] §3 (QCQP formulation): the reduction of the optimal-reference problem to a QCQP inside each affine region is formally correct only if the piecewise-affine flux model reproduces the true torque and loss surfaces to sufficient accuracy everywhere the controller operates. No a-priori error bounds or cross-validation against the underlying nonlinear FE data are reported, so the claimed optimality and tractability do not automatically transfer to the physical three-DOF machine.

    Authors: The referee correctly notes the dependence on model fidelity. While the original manuscript relied on the prototype experiments for implicit validation, we have now included explicit cross-validation results and a-priori error bounds derived from the FE data fitting process. A new figure and table in §3 and the Results section quantify the approximation error over the current space, confirming that the QCQP solutions remain near-optimal for the physical machine. revision: yes

  3. Referee: [Results] Results section: no direct comparison is provided between the closed-form or SDP solutions obtained from the piecewise-affine model and either the exact nonlinear optimum or existing heuristic current-reference methods. Without such benchmarks it is difficult to quantify the accuracy loss introduced by the affine approximation or to confirm that the SDP remains small enough for real-time use.

    Authors: We have addressed this by adding benchmark comparisons in the revised Results section. Specifically, we compare the PWA-derived closed-form and SDP solutions to numerically optimized references using the full nonlinear model, as well as to conventional heuristic methods such as maximum torque per ampere with field weakening. The accuracy loss is quantified (typically under 3% in torque), and timing benchmarks confirm the SDP solves in milliseconds on standard embedded processors, validating real-time feasibility. revision: yes

Circularity Check

0 steps flagged

No significant circularity: derivation applies standard QCQP optimization to external FEM-derived PWA model

full rationale

The paper formulates the WRSM current-reference problem as a QCQP inside each affine flux region obtained from finite-element data, then derives closed-form solutions, low-dimensional polynomials, or a small SDP according to active constraints. These steps follow directly from standard convex optimization theory applied to the stated model; the PWA pieces are treated as given external inputs rather than fitted or predicted inside the derivation itself. No self-definitional reductions, fitted inputs renamed as predictions, load-bearing self-citations, or ansatz smuggling are exhibited in the abstract or described chain. The central claim of tractable extension to three-DOF WRSMs therefore remains independent of its own outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the accuracy of a piecewise-affine flux model extracted from finite-element analysis and on the assumption that the operating space can be partitioned into affine regions where the QCQP formulation holds.

axioms (1)
  • domain assumption Magnetic flux linkage can be represented by a piecewise-affine function of current in distinct operating regions derived from finite-element data.
    Invoked to convert the nonlinear machine model into a QCQP that admits closed-form or low-dimensional solutions.

pith-pipeline@v0.9.0 · 5488 in / 1297 out tokens · 37881 ms · 2026-05-10T06:00:33.730085+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

28 extracted references · 28 canonical work pages

  1. [1]

    Energy-Efficiency Policy Opportunities for Electric Motor-Driven Systems,

    P. Waide and C. U. Brunner, “Energy-Efficiency Policy Opportunities for Electric Motor-Driven Systems,”IEA Energy Papers, vol. 2011, no. 7, Apr. 2011

    work page 2011

  2. [2]

    Global EV Outlook 2025 – Analysis,

    “Global EV Outlook 2025 – Analysis,” https://www.iea.org/reports/global-ev-outlook-2025, May 2025

    work page 2025

  3. [3]

    Expansion of operating limits for permanent magnet motor by current vector control considering inverter capacity,

    S. Morimoto, Y . Takeda, T. Hirasa, and K. Taniguchi, “Expansion of operating limits for permanent magnet motor by current vector control considering inverter capacity,”IEEE Transactions on Industry Applications, vol. 26, no. 5, pp. 866–871, Sep. 1990

    work page 1990

  4. [4]

    Simultaneous IDA-passivity-based control of a wound rotor synchronous motor,

    C. Batlle, A. Doria-Cerezo, and G. Espinosa-Perez, “Simultaneous IDA-passivity-based control of a wound rotor synchronous motor,” in2008 47th IEEE Conference on Decision and Control, Dec. 2008, pp. 3187–3191

    work page 2008

  5. [5]

    Optimal current waveforms for brushless permanent magnet motors,

    N. Moehle and S. Boyd, “Optimal current waveforms for brushless permanent magnet motors,”International Journal of Control, vol. 88, no. 7, pp. 1389–1399, Jul. 2015

    work page 2015

  6. [6]

    Maximum torque-per-current control of induction motors via semidefinite programming,

    ——, “Maximum torque-per-current control of induction motors via semidefinite programming,” in2016 IEEE 55th Conference on Decision and Control (CDC), Dec. 2016, pp. 1920–1925

    work page 2016

  7. [7]

    Sliding mode control of an isolated wound rotor synchronous generator,

    A. D `oria-Cerezo, E. Fossas, R. Mu ˜noz-Aguilar, and V . Utkin, “Sliding mode control of an isolated wound rotor synchronous generator,” in2009 European Control Conference (ECC), Aug. 2009, pp. 2728–2733

    work page 2009

  8. [8]

    Optimal State Reference Computation With Constrained MTPA Criterion for PM Motor Drives,

    M. Preindl and S. Bolognani, “Optimal State Reference Computation With Constrained MTPA Criterion for PM Motor Drives,”IEEE Transactions on Power Electronics, vol. 30, no. 8, pp. 4524–4535, Aug. 2015

    work page 2015

  9. [9]

    A unified theory for optimal feedforward torque control of anisotropic synchronous machines,

    H. Eldeeb, C. M. Hackl, L. Horlbeck, and J. Kullick, “A unified theory for optimal feedforward torque control of anisotropic synchronous machines,”International Journal of Control, vol. 91, no. 10, pp. 2273–2302, Oct. 2018

    work page 2018

  10. [10]

    Nonlinear model predictive torque control and setpoint computation of induction machines for high performance applications,

    T. Englert and K. Graichen, “Nonlinear model predictive torque control and setpoint computation of induction machines for high performance applications,”Control Engineering Practice, vol. 99, p. 104415, Jun. 2020

    work page 2020

  11. [11]

    Electrical Motor Drivelines in Commercial All-Electric Vehicles: A Review,

    J. de Santiago, H. Bernhoff, B. Ekerg ˚ard, S. Eriksson, S. Ferhatovic, R. Waters, and M. Leijon, “Electrical Motor Drivelines in Commercial All-Electric Vehicles: A Review,”IEEE Transactions on V ehicular Technology, vol. 61, no. 2, pp. 475–484, Feb. 2012

    work page 2012

  12. [12]

    Wound Rotor Salient Pole Synchronous Machine Drive for Electric Traction,

    C. Rossi, D. Casadei, A. Pilati, and M. Marano, “Wound Rotor Salient Pole Synchronous Machine Drive for Electric Traction,” inConference Record of the 2006 IEEE Industry Applications Conference F orty-First IAS Annual Meeting, vol. 3, Oct. 2006, pp. 1235–1241

    work page 2006

  13. [13]

    Applying a Measurement-Based Iron Loss Model to an Efficiency Optimized Torque Control of an Electrically Excited Synchronous Machine,

    S. M ¨uller and N. Parspour, “Applying a Measurement-Based Iron Loss Model to an Efficiency Optimized Torque Control of an Electrically Excited Synchronous Machine,” in2020 International Conference on Electrical Ma- chines (ICEM), vol. 1, Aug. 2020, pp. 779–785

    work page 2020

  14. [14]

    Artificial neural network based optimal feedforward torque control of electrically excited synchronous machines

    N. Monzen and C. Hackl, “Artificial neural network based optimal feedforward torque control of electrically excited synchronous machines.”

    work page

  15. [15]

    Applying semidefinite relaxation to optimal efficiency reference generation for wound rotor synchronous machines,

    M. Parson-Scherban, B. Steyaert, E. Swint, W. W. Pennington, and M. Preindl, “Applying semidefinite relaxation to optimal efficiency reference generation for wound rotor synchronous machines,” inProceedings of the 2026 American Control Conference (ACC), New Orleans, Louisiana, USA, 2026, to appear. 11 APREPRINT- APRIL21, 2026

    work page 2026

  16. [16]

    Effects of Magnetic Saturation on Optimal Efficiency Refer- ence Generation for the Wound Rotor Synchronous Machine,

    M. Parson-Scherban, B. Steyaert, and M. Preindl, “Effects of Magnetic Saturation on Optimal Efficiency Refer- ence Generation for the Wound Rotor Synchronous Machine,” inIECON 2024 - 50th Annual Conference of the IEEE Industrial Electronics Society, Nov. 2024, pp. 1–6

    work page 2024

  17. [17]

    Extended MTPA and Flux-Weakening Control with Total Copper Loss Minimisation and Transient Torque Compensation for Wound-Rotor Synchronous Machines,

    R. Breda, G. Andrioli, S. Calligaro, and R. Petrella, “Extended MTPA and Flux-Weakening Control with Total Copper Loss Minimisation and Transient Torque Compensation for Wound-Rotor Synchronous Machines,” in 2023 IEEE Energy Conversion Congress and Exposition (ECCE), Oct. 2023, pp. 4829–4836

    work page 2023

  18. [18]

    Optimal current setpoint computation for externally excited synchronous machines,

    J. Reinhard, K. L ¨ohe, and K. Graichen, “Optimal current setpoint computation for externally excited synchronous machines,” in2022 IEEE Conference on Control Technology and Applications (CCTA), Aug. 2022, pp. 1319– 1326

    work page 2022

  19. [19]

    Generic loss minimization for nonlinear synchronous machines by ana- lytical computation of optimal reference currents considering copper and iron losses,

    C. Hackl, J. Kullick, and N. Monzen, “Generic loss minimization for nonlinear synchronous machines by ana- lytical computation of optimal reference currents considering copper and iron losses,” in2021 22nd IEEE Inter- national Conference on Industrial Technology (ICIT), vol. 1, Mar. 2021, pp. 1348–1355

    work page 2021

  20. [20]

    Piecewise Affine Modeling of Wound-Rotor Synchronous Machines for Real-Time Motor Control,

    B. W. Steyaert, E. Swint, W. W. Pennington, and M. Preindl, “Piecewise Affine Modeling of Wound-Rotor Synchronous Machines for Real-Time Motor Control,”IEEE Transactions on Industrial Electronics, vol. 70, no. 6, pp. 5571–5580, Jun. 2023

    work page 2023

  21. [21]

    Inductance Calculations of Permanent-Magnet Syn- chronous Machines Including Flux Change and Self- and Cross-Saturations,

    K. J. Meessen, P. Thelin, J. Soulard, and E. A. Lomonova, “Inductance Calculations of Permanent-Magnet Syn- chronous Machines Including Flux Change and Self- and Cross-Saturations,”IEEE Transactions on Magnetics, vol. 44, no. 10, pp. 2324–2331, Oct. 2008

    work page 2008

  22. [22]

    Optimal Setpoint Computation for Constrained Torque Control of PMSMs,

    T. Englert and K. Graichen, “Optimal Setpoint Computation for Constrained Torque Control of PMSMs,” in 2018 European Control Conference (ECC), Jun. 2018, pp. 2671–2677

    work page 2018

  23. [23]

    Convex Optimization – Boyd and Vandenberghe,

    “Convex Optimization – Boyd and Vandenberghe,” https://web.stanford.edu/˜boyd/cvxbook/

    work page

  24. [24]

    K. Zhou, J. C. Doyle, and K. Glover,Robust and Optimal Control. USA: Prentice-Hall, Inc., 1996

    work page 1996

  25. [25]

    A Remark on the Rank of Positive Semidefinite Matrices Subject to Affine Constraints,

    A. I. Barvinok, “A Remark on the Rank of Positive Semidefinite Matrices Subject to Affine Constraints,” Jan. 2001

    work page 2001

  26. [26]

    Real Time Core Loss Estimation for the Wound Rotor Synchronous Machine,

    B. Steyaert, E. Swint, W. W. Pennington, and M. Preindl, “Real Time Core Loss Estimation for the Wound Rotor Synchronous Machine,” in2023 IEEE 14th International Symposium on Diagnostics for Electrical Machines, Power Electronics and Drives (SDEMPED), Aug. 2023, pp. 246–250

    work page 2023

  27. [27]

    Gurobi Optimizer Reference Manual,

    Gurobi Optimization, LLC, “Gurobi Optimizer Reference Manual,” 2026. [Online]. Available: https://www.gurobi.com

    work page 2026

  28. [28]

    V ersion 10.0., 2024

    MOSEK ApS,The MOSEK Optimization Toolbox for MATLAB Manual. V ersion 10.0., 2024. [Online]. Available: https://docs.mosek.com/latest/toolbox/index.html A Polynomial Coefficients for Torque-Only Regime: Cruise In the cruise regime the problem reduces to (2). KKT stationarity givesi(µ) =−A(µ) −1b(µ)withA(µ) := 2(Reff +µQ τ)andb(µ) := [b r, b d −µψ q, b q +µ...

    work page 2024