A

Iterative Moving Horizon Algorithm for Rotor Angle Estimation of Nonlinear IPMSMs

International Conference on Electrical Machines (ICEM), pp. 1–8

Abstract

This paper presents an iterative Moving Horizon Estimation (MHE) algorithm for sensorless rotor angle reconstruction of highly nonlinear Interior Permanent Magnet Synchronous Machines (IPMSMs). In contrast to conventional approaches utilizing constant inductance models, the proposed method explicitly incorporates the machine’s intrinsic nonlinearities, specifically the current dependencies of differential inductances and flux linkages formulated in the stator-fixed αβ-coordinate system. This physics-aware modeling approach minimizes parameter mismatch relative to the physical machine, ensuring robust estimation even under conditions of deep magnetic saturation. To circumvent the prohibitive computational cost of the resulting nonlinear optimization problem, a recursive linearization scheme is derived that transforms the estimation task into a Quadratic Program (QP). Analytical expressions for the Hessian matrix and the gradient vector are therefore formulated. The performance of the algorithm is validated through a comprehensive simulation study. The estimator is examined dynamically and statically across the entire base speed range under diverse d- and q-current trajectories. Results demonstrate that the proposed algorithm achieves consistently low angular tracking errors across the entire operating envelope. These results demonstrate the high accuracy and robustness of the proposed estimator, particularly in operating regions dominated by significant magnetic saturation and cross-coupling effects.

Authors 3

  1. RWTH Aachen University

    Affiliation as printed

    Institute of Electrical Machines RWTH Aachen University

  2. RWTH Aachen University

    Affiliation as printed

    Institute of Electrical Machines RWTH Aachen University

  3. RWTH Aachen University

    Affiliation as printed

    Institute of Electrical Machines RWTH Aachen University

Cited by 0 stored of 0

References 13

13 results