Subsystem decomposition and distributed state estimation of nonlinear process networks

• Author / Creator

  Yin, Xunyuan

• The distributed framework has been recognized as a promising framework for handling the large size and the strong interaction of modern integrated chemical processes. Compared to distributed process control, its dual problem - distributed state estimation which is equally
  important, has received relatively less attention in the process control community. This thesis deals with distributed state estimation for nonlinear process networks as well as subsystem structure decomposition and configuration for distributed state estimation. The contributions of this thesis include the development of distributed state estimation methods for nonlinear process networks, the development of systematic approaches to properly decompose general nonlinear processes into subsystems for distributed estimation, and the
  application of the developed methods in different processes and output-feedback fault detection and isolation.

  First, two-time-scale nonlinear systems are taken into account. The system is decomposed into fast and slow subsystems based on the singular perturbation theory. Local observer-enhanced moving horizon state estimators are designed. A one-directional communication scheme is used. The convergence and boundedness of the estimation error is rigorously studied. A benchmark chemical process example is used to illustrate the proposed method. Then, attention is given to general nonlinear systems that can be divided into smaller subsystems. It is assumed that a decentralized state estimation scheme already exists for the system. And the aim is to form a distributed state estimation scheme based on the existing one without significant modifications. Compensators are designed for the subsystems and augmented estimators that communicate with each other are obtained to form a distributed network. Stability analysis is carried out. The convergence and ultimate boundedness of the estimation error dynamics can be ensured subject to reasonable assumptions. The proposed method is applied to three application examples and good state estimates are obtained. This distributed state estimation method is also utilized in the development of a distributed output-feedback fault detection and isolation mechanism for cascade nonlinear processes. The state estimation scheme provides state estimates that are further used for generating residual signals for the subsystems. A distributed fault detection and isolation mechanism is proposed and applied to a froth flotation process.

  With the development of distributed state estimation algorithm being completed, we further explore the decomposition and configuration of subsystem structure for distributed state estimation. A systematic procedure is proposed to address the considered subsystem decomposition problem. The procedure is based on the evaluation of physical closeness between state and output measurement variables. The proposed method is applied to the froth flotation process and subsystem models are configured. The decomposition result is consistent with physical topology and can be readily used for distributed state estimation design. Moreover, we consider systematic subsystem decomposition and distributed state estimation for a wastewater treatment plant. Based on an extension of the previously proposed method, the large-scale plant is decomposed into subsystems and local state estimators are developed. Good simulation results confirm the effectiveness of the systematic approach. Finally, to facilitate the synthesis of distributed state estimation and distributed control, a systematic approach on subsystem decomposition of process networks for simultaneous distributed estimation and control is presented.

• Subjects / Keywords

  • State estimation, subsystem decomposition, moving horizon estimation

• Graduation date

  Fall 2018

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/R31C1TX68

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.