Application of Regression based Speed Estimation for Sensorless Vector Controlled IM Drive

Abstract

This paper presents a new approach to estimate the speed of vector-controlled three-phase induction motor drive using polynomial regression. The speed estimation via polynomial regression does not require any gain adjustment as well as PI controllers. The identification method is relying on the learning algorithm to enable an estimator fit between the reactive power (Q = Vsq.isd − Vsd.isq) and speed (ωr). The regressor quantity (Q) explicitly dependent on the voltage and current vectors and thus computation of flux is not required. The formulation of polynomial regression based speed estimation technique is simply realizable as the intensive computation of differential/integral is not required which makes the approach efficient and less time consuming. The Simulation results validates precise tracking of rotor speed for high speed, low speed and reversal from motoring to regenerative mode is achieved through this approach.