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 · 2026

Read On the Unification of Optimal Current Reference Theory for Wound Rotor Synchronous Machines by Maxfield Parson-Scherban, Kasra Fallah, Navid Rahbariasr, Bernard Steyaert, James Anderson, Matthias Preindl (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-17165.txt.

Introduction

Parson-Scherban and five colleagues extend optimal current reference theory to wound rotor synchronous machines, adding a rotor-current degree of freedom plus magnetic saturation, cross-coupling, and speed-dependent core losses. They recast the optimal reference search within affine flux regions as a quadratically constrained quadratic program built from finite-element data. Solutions are closed-form or low-dimensional polynomials in several constraint regimes, with a small semidefinite program in the voltage-limited case. Tests on a physical prototype confirm the approach across the torque-speed envelope.

Catalog record

CREATORMaxfield Parson-Scherban, Kasra Fallah, Navid Rahbariasr, Bernard Steyaert, James Anderson, Matthias Preindl
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37046
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.17165

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

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 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. 1 INTRODUCTION Electric motors account for a substantial share of global electricity consumption [ 27 , 10 ] , making efficiency-oriented motor control an important practical problem. In electric vehicle applications, this problem is particularly challenging because traction motors must operate efficiently over a wide range of torque and speed conditions, rather than near a single nominal operating point. This motivates the study of real-time current control strategies that explicitly account for efficiency and feasibility across the operating envelope. A central component in this setting is the optimal current reference generator: for a given torque request and operating speed, it selects stator and rotor current references that satisfy the requested torque, respect current and voltage limits, and minimize electrical losses. This is inherently a constrained decision problem. For a fixed torque-speed pair, the set of feasible current vectors is generally non-unique and, in general, infinite [ 17 ] . As illustrated in Fig. 1 , current reference generation therefore acts as an optimization layer within the control loop: the reference generator selects the steady-state target, while the lower-level current controller drives the machine to that operating point. The steady-state efficiency of the closed-loop system consequently depends on the quality of this upstream decision. Figure 1: Example synchronous machine (WRSM) control block diagram and schematic. Optimization-based formulations of motor control problems have received growing attention in both the motor drives and control communities [ 2 , 14 , 15 , 6 ] . For the Permanent Magnet Synchronous Machine (PMSM), the optimal current reference problem is by now well understood: it admits a Quadratically Constrained Quadratic Programming (QCQP) formulation, and analytical characterizations are available for the relevant operating regimes, including field-weakening [ 17 , 22 , 7 , 9 ] . The Wound Rotor Synchronous Machine (WRSM) has recently re-emerged as a promising alternative for traction applications [ 5 ] . By replacing permanent magnets with an externally excited rotor winding, the WRSM reduces dependence on rare-earth materials, improves safety under certain fault scenarios, and enables active flux control over a broad operating range [ 24 ] . This additional flexibility, however, significantly complicates the current reference generation problem. In contrast to the PMSM case, the rotor current introduces a third control degree-of-freedom, which enlarges the feasible set and makes the structure of the optimal solution substantially more involved, especially in the presence of magnetic saturation, cross-saturation, and speed-dependent core losses. Existing approaches to WRSM current reference generation are incomplete in complementary ways. Numerical optimization methods can provide high-quality solutions offline [ 19 ] , but their computational cost and runtime variability can limit direct use in embedded real-time implementations. Learning-based surrogates trained on such solutions can improve runtime performance [ 16 ] , but they generally do not preserve analytical structure or provide certifiable optimality properties, and may require retraining when machine characteristics change. On the analytical side, the WRSM current reference problem has been formulated as a QCQP in [ 21 , 20 ] , and KKT-based analysis of selected operating regimes appear in [ 3 , 23 ] . However, existing analytical treatments either neglect saturation and cross-saturation, or impose assumptions that exclude practically relevant magnetic coupling effects. To the best of our knowledge, a general analytical treatment of the WRSM reference generation problem that simultaneously accounts for saturation, cross-saturation, and core losses across the relevant operating regimes is not yet available. This paper addresses that gap. Using a Piecewise-Affine (PWA) approximation of the nonlinear flux map obtained from finite-element data, we formulate the WRSM current reference generation problem as a QCQP on each affine region. We then characterize the solution structure according to the binding constraints at optimality. This regime-wise decomposition yields closed-form or low-dimensional polynomial solution procedures in several practically relevant cases, and a small semidefinite programming (SDP) relaxation in the voltage-constrained regime. Across the tested operating envelope, the resulting method matches the solution quality of a state-of-the-art nonlinear solver while achieving substantially lower runtime. A comparison with related work is reported in Table 1 . Contributions. The main contributions of this work are as follows. \bullet WRSM reference-generation formulation. We formulate the optimal current reference generation problem for the WRSM as a QCQP on each region of a PWA magnetic model, incorporating saturation, cross-saturation, and a quadratic core-loss model with speed-dependent iron losses. \bullet Regime-wise analytical solution structure. We characterize the active-constraint regimes of the problem and derive closed-form or low-dimensional polynomial solution procedures for four of the resulting cases not handled in [ 7 ] . The remaining voltage-constrained case is handled through a small SDP relaxation, which is observed to be exact throughout the tested operating envelope. \bullet Computationally efficient online reference generation. The proposed regime-aware solution strategy exploits problem structure to avoid relying exclusively on generic nonlinear optimization at runtime, making it well suited for real-time reference generation. \bullet Validation on a prototype WRSM. Using finite-element data from a prototype WRSM, we show that the proposed framework reproduces the optimal reference solutions obtained by a state-of-the-art nonlinear solver while reducing median solve time by approximately 95%, with median runtime below 1 ms. Table 1: Comparison of Related Work Work Machine Closed form Flux offset Cross- coupling Core loss SDP † [ 22 ] PMSM ✓ ✓ \times \times — [ 7 ] + [ 12 ] PMSM ✓ ✓ ✓ ✓ — [ 23 ] WRSM ✓ \times \times \times \times [ 21 ] WRSM \times ✓ ✓ ✓ ✓ Proposed WRSM ✓ ✓ ✓ ✓ ✓ † SDP only applicable to WRSM (3-DOF) voltage-constrained regime. 2 Problem Formulation The WRSM can be thought of as a dynamic system which takes a current vector i=[i_{r},i_{d},i_{q}]^{\top} as an input and produces some torque T_{p} at a certain electrical speed \omega_{e} as an output. The i_{d} and i_{q} elements of the current vector represent the direct-axis and quadrature-axis currents respectively and importantly fully represent the balanced 3-phase AC waveform fed to the motor via the Park-Clarke transform. The WRSM, unlike the PMSM, has one additional current input to control the flux induced by the rotor: i_{r} . The constraints on i are a result of the limitations of the coils through which

Read the complete text