Resource Efficient Sensorless Control of Low-Power Permanent Magnet Synchronous Machines
Automation, Mechatronics, Control Systems, Intelligent Technical Systems, Robotics
Final Report Abstract
In the last decades, Permanent Magnet Synchronous Motors (PMSMs) have gained wide applicability in industrial, automotive and domestic applications due to their power density, high dynamic behaviour and robustness. In order to control such machines, discrete sensors like encoders and resolvers are required for rotor position sensing, which increases system cost and complexity. For this reason, sensorless respectively self-sensing techniques are being developed, with back-Electromotive Force (EMF) based techniques focusing on middle-to-high speed operation, and anisotropy-based ones for standstill and low-speed. Within the latter category, techniques are based on the exploitation of current ripples inherently introduced by a Pulse-Width-Modulation (PWM) inverter. By applying a numerical differentiation, or by oversampling the current, information about the inductance and, consequently, the position can be obtained. For the case of low-power machines, it is important to apply resource-efficient techniques that allow the implementation in cost- and space-critical applications. The approach presented in this research project aims at providing a contribution to the toolbox of self-sensing techniques by proposing a hybrid analog/digital signal processing method capable of providing similar performance to well-known techniques at lower computational expenses. Based on a preceding work introducing the Integrator-based Direct Inductance Measurement (IDIM) technique for single-phase electromagnetic actuators, the current ripple is processed by a fast-resettable integrator circuit synchronized to the switching states of the inverter. The conducted research allows the application of the IDIM technique onto PMSMs first by analysing mathematically the current ripple in three-phase machines under the consideration of back-EMF, stator resistance and eddy currents. Afterwards, a machine excitation strategy involving different PWM patterns and current measurement topologies is developed. Then, a position estimator is synthesized and validated in simulation, that takes as input the integrated current ripple. Finally, experiments show that the noise power of the estimated inverse inductance of the approach is comparable to the one of the computation-intensive oversampling approaches. Moreover, position estimates on three different machines with the new approach show a reliable position estimation when compared to a setup with a high-resolution encoder.
Publications
-
An Integrated Fast Resettable Integrator Circuit based on Switched Capacitors for Sensorless Control of PMSMs, IKMT 2022 – Innovative Small Drives and Micro Motor Systems, 13. ETG/GMM Symposium, Linz, Austria, 2022.
König, N., Stanitzki, A., Nienhaus, M. & Grasso, E.
-
Analysis of Static Eccentricity on the Position Estimation of Zero-Sequence Voltage Based Sensorless Techniques. 2023 IEEE International Symposium on Sensorless Control for Electrical Drives (SLED), 1-7. IEEE.
Klein, Carsten; König, Niklas; Palmieri, Marco; Nienhaus, Matthias & Grasso, Emanuele
-
Sensorless Control of a Two-Phase Linear PMSM exploiting PWM-Induced Current Ripples. 2023 IEEE International Symposium on Sensorless Control for Electrical Drives (SLED), 1-6. IEEE.
Summa, Joshua; König, Niklas; Nienhaus, Matthias & Grasso, Emanuele
-
Analysis of PWM-induced current ripples in electromagnetic actuators for position self-sensing, PhD thesis, Saarland University, 2024
König, N.
-
Position Estimation for PMSMs at any speed using the zero-sequence voltage and modified space vector modulation, PhD thesis, Saarland University, 2024
Schuhmacher, K.
