Experimental evaluation of torque performance of voltage and current models using measured torque for induction motor drives

Abstract

In this paper, two kinds of observers are proposed to investigate torque estimation. The first one is based on a voltage model represented with a lowpass filter (LPF); which is normally used as a replacement for a pure integrator to avoid integration drift problem due to dc offset or measurement error. The second estimator used is an extended Kalman filter (EKF) as a current model, which puts into account all noise problems. Both estimation algorithms are investigated during the steady and transient states, tested under light load, and then compared with the measured mechanical torque. In all conditions, it will be shown that the torque estimation error for EKF has remained within narrower error band and yielded minimum torque ripples when compared to LPF estimation. This motivates the use of EKF observer in high performance control drives of induction machines for achieving improved torque response.