States and parameters estimation in induction motor using Bayesian techniques

This paper addresses the problem of rotor speed, flux and parameters estimation of induction motor on the basis of a three-order electrical model. Thus, we propose to use a particle filtering (PF) to estimate states and parameters for an induction motor. It is assumed that only the voltages stator currents are measurable. In addition, the rotor resistance and magnetizing inductance, which vary with the motor temperature and magnetization level, can also be estimated within the same framework. Hence, the objective of this work is to estimate three states (the rotor speed, the rotor flux, and the stator flux) and two parameters (the rotor resistance and the magnetizing inductance). Simulation analysis demonstrates that the Bayesian algorithm can well estimate the states/parameters under disturbs of the noise, and it provides efficient accuracies for the states estimation. In addition, detailed case studies show that Bayesian algorithm has advantages over Unscented Kalman filter (UKF) for highly nonlinear estimation problems. Evaluation of the methods was performed by using Root Mean Square Error.