Authors:

Suhani Mathur

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Permanent Magnet Synchronous Motors (PMSMs) are central to modern electrification systems, powering electric vehicles, industrial automation, robotics, and aerospace applications. Their widespread use is driven by their high efficiency, torque density, and precise controllability. This research investigates a hybrid approach that combines physics-based modeling with adaptive neural learning to estimate four key parameters: stator resistance, d- and q-axis inductances and permanent magnet flux linkage. The estimation is implemented within a high-fidelity simulation framework built in MATLAB/Simulink, which emulates real-world motor dynamics under field-oriented control (FOC) using a behaviorally accurate model representative of a digital twin. Leveraging this environment, an Adaline neural network is trained using tapped voltage, current, and rotor speed signals to learn parameter relationships through real-time linear regression. Initial results show that estimator performance is highly sensitive to the operating regime: clean, steady- state intervals enable convergence toward reference values, while transient conditions introduce signal noise and reduce reliability. These early insights highlight the importance of data segmentation, physical modeling, and estimator interpretability in data-driven motor control. While still in progress, this work establishes a foundation for scalable parameter identification that avoids the need for intrusive signal injection or offline calibration. Future development will explore nonlinear models, incorporate thermal and mechanical dynamics, and extend toward a full digital twin architecture capable of real-time estimation and diagnostics. Ultimately, this research supports the development of smarter, self- monitoring electric drive systems with applications in electric vehicles, robotics, and industrial automation. Keywords: Computer Engineering; Neural Networks; Digital Twin; Motor Control

Source:

Purdue University / 2025

Topics:

No topics listed

Co-authors:

Suhani Mathur