Extended Kalman Filter Based Speed-Sensorless Load Torque and Inertia Estimations with Observability Analysis for Induction Motors

Abstract

This paper aims to introduce a novel extended Kalman filter (EKF) based estimator including observability analysis to the literature associated with the high performance speed-sensorless control of induction motors (IMs). The proposed estimator simultaneously performs the estimations of stator stationary axis components of stator currents and rotor fluxes, rotor mechanical speed, load torque including the viscous friction term, and reciprocal of total inertia by using measured stator phase currents and voltages. The inertia estimation is done since it varies with the load coupled to the shaft and affects the performance of speed estimation especially when the rotor speed changes. In this context, the estimations of all mechanical state and parameters besides flux estimation required for high performance control methods are performed together. The performance of the proposed estimator is tested by simulation and real-time experiments under challenging variations in load torque and velocity references; and in both transient and steady states, the quite satisfactory estimation performance is achieved.

Real-Time Speed Estimation for an Induction Motor: An Automated Tuning of an Extended Kalman Filter Using Voltage–Current Sensors

“…In fact, researchers focus on investigating the optimization of noise covariance matrices for induction motor speed estimation. This investigation can be categorized into two main parts: the first group of studies [ 10 , 11 ] uses artificial intelligence-based methods, which requires expert knowledge and involves a complex design procedure, whereas the second group [ 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 ] employs adaptive structures to eliminate the adverse effect under operating condition variations.…”

Compensation of Current Sensor Faults in Vector-Controlled Induction Motor Drive Using Extended Kalman Filters

“…Of course, in the event of failure of all CSs, the only way to restore the stator current comes down to its estimation using an open-loop observer [9], such as a Virtual Current Sensor (VCS), proposed in [12], or flux-based observer [13]. When it comes to speed-sensorless structures, extended Kalman filters (EKFs) are commonly presented [16][17][18][19][20][21][22][23][24][25][26]. Surprisingly, in the literature known to the authors, there is only one work [14] where an EKF was proposed for the current reconstruction.…”

“…However, the appropriate choice of Q and R matrices is crucial. They determine the quality and rate of the estimation process [20,28,29].…”

Rotor Resistance Estimator based on Virtual Current Sensor Algorithm for Induction Motor Drives

“…Spectral analysis techniques uses the analysis of characteristic harmonics in measured values of stator currents and/or voltages (Zai et al, 1992;Loron and Laliberte, 1993). Among the observer-based techniques, the extended Kalman filters (EKF) (Barut et al, 2012;Zerdali and Barut, 2018;Horváth and Kuslits, 2018) and extended Luenberger observers (ELO) are popular (Orlowska-Kowalska, 1989;Du and Brdys, 1993). In the case of EKF, a problem with linearisation of the extended mathematical model of IM in each numerical step appears.…”