Identification of block cascade models of nonlinear systems

(1994) Identification of block cascade models of nonlinear systems. Masters thesis, King Fahd University of Petroleum and Minerals.


Download (3MB) | Preview

Arabic Abstract


English Abstract

In this thesis, the structure identification problem of block cascade models composed of linear dynamic and static nonlinear blocks is addressed. Two main approaches are considered, the structure identification in the time domain and the structure identification in the frequency domain. In the time domain analysis, cross-correlation techniques are used for the structure identification of nonlinear systems when the input to these systems is a white Gaussian signal. The nonlinear systems considered in this part are those that can be described by a model consisting of a linear dynamic block in cascade with a static nonlinear element followed by another linear dynamic block. This model is referred to as a Wiener-Hammerstein model. Two cross-correlation functions, named as the linear and nonlinear cross-correlation functions, were used to identify the structure of this model. From these two functions, a structure identification criterion that covers all the subclasses of the Wiener-Hammerstein model is presented. In the frequency domain analysis, the structure identification of the Wiener-Hammerstein model is performed utilizing the bispectrum of the system output. Using a zero-mean stationary white Gaussian process as an input to this model, a structure identification criterion based on the bispectrum of the output sequence only is presented and applied to the above Wiener-Hammerstein model as well as to a more general class of nonlinear systems that can be represented by parallel branches of block cascades in parallel with a linear system. The case treated here is for branches of the same type. The mixed type branches case, however, needs further investigation. It is found from the simulation results that the structure identification criterion is still effective in characterizing the system structure in both situations.

Item Type: Thesis (Masters)
Subjects: Mechanical
Department: College of Engineering and Physics > Mechanical Engineering
Committee Advisor: Moustafa, Kamal A. F.
Committee Members: Khulief, Yehia A. and Ahmed, Munir and Emara-Shabaik, H.E.
Depositing User: Mr. Admin Admin
Date Deposited: 22 Jun 2008 14:04
Last Modified: 01 Nov 2019 14:00