Sensorless Flying Start of Synchronous Reluctance Drive Based on Opposite Voltage Vector Pulses

Read full text
Authors: Luka Pravica, Filip Jukić, Damir Sumina, Igor Erceg
Cited by: 0
References: 26

Abstract

Synchronous reluctance machines (SynRMs) are increasingly used in applications such as electric drives and energy conversion systems. In addition to standard control algorithms, certain applications, especially those that consider a priori rotating SynRMs, require a flying start procedure where the rotor speed and position should be estimated before the SynRM is connected to the converter. This synchronization between the converter and the SynRM minimizes the inrush current and prevents torque disturbances. In this paper, a novel method of flying start for SynRMs is proposed based on the application of symmetrical and opposite active voltage pulses together with two zero-voltage vectors and a generalized second-order integrator with a frequency-locked loop (SOGI-FLL). By applying such a voltage pattern to the terminals of a rotating SynRM, phase currents are induced whose sampled counterparts follow the sine waveform. Based on the derived machine model in the stationary$\alpha \beta$reference frame for such discontinuous operation of the SynRM, the induced$\beta$-axis current is inherently free of DC offset and is directly used as input to the SOGI-FLL to estimate the rotor angle and speed. Moreover, the voltage pattern based on the opposite vector pulses followed by the zero-voltage vectors also reduces the current distortion and torque ripple, which improves the estimation accuracy and reduces the mechanical stress. The proposed method was first verified by simulations in MATLAB/Simulink and then experimentally validated in a laboratory setup, confirming the practical feasibility and robustness of the method.

Fields of Study & Subjects

Sensorless Control of Electric Motors Electric Motor Design and Analysis Control Systems in Engineering

Publication details

Journal / source
IEEE Transactions on Industry Applications
Volume
62
Issue
5
Pages
7272–7282
Publication date
27 January 2026