Bayesian methods for time‐varying state and parameter estimation in induction machines. (19th August 2014)

• Record Type: Journal Article
• Authors: Mansouri, Majdi M.; Mohamed‐Seghir, Moustafa M.; Nounou, Hazem N.; Nounou, Mohamed N.; Abu‐Rub, Haitham A.
• Is Part Of: International journal of adaptive control and signal processing. Volume 29:Number 7(2015:Jul.)
• Page Start: 905
• Page End: 924
• Publication Date: 2014-08-19
• Subjects: Adaptive control systems -- Periodicals; Adaptive signal processing -- Periodicals; 629.836
• DOI: 10.1002/acs.2511
• Languages: English
• ISSNs: 0890-6327

Abstract

<abstract abstract-type="main" id="acs2511-abs-0001"> <title>Summary</title> <p id="acs2511-para-0001">This paper addresses the problem of nonlinear time‐varying state and parameter estimation of induction machines (IMs) on the basis of a third‐order electrical model. The objectives of this paper are threefold. The first objective is to propose the use of an improved particle filter (IPF) with better proposal distribution for nonlinear and non‐Gaussian state and parameter estimation. The second objective is to extend the state and parameter estimation techniques (i.e., extended Kalman filter (EKF), unscented Kalman filter (UKF), particle filter (PF), and IPF) to better handle nonlinear and non‐Gaussian processes without <italic>a priori</italic> state information, by utilizing a time‐varying assumption of statistical parameters. In this case, the state vector to be estimated at any instant is assumed to follow a Gaussian model, where the expectation and the covariance matrix are both random. The third objective is to compare the performances of EKF, UKF, PF, and IPF in estimating the states of the power process model representing the IM (i.e, the rotor speed, the rotor flux, the stator flux, the rotor resistance, and the magnetizing inductance) and their abilities to estimate some of the key system parameters, which are needed to define the IM process model. The results show that the IPF provides a significant improvement over the PF because, unlike the PF, which depends on the choice of sampling distribution used to estimate the posterior distribution, the IPF yields an optimum choice of the sampling distribution, which also accounts for the observed data. This conclusion is also supported by the experimental results. Copyright © 2014 John Wiley & Sons, Ltd.</p> </abstract>