Showing papers by "Ron J. Patton published in 1999"

Active fault-tolerant flight control systems design using the linear matrix inequality method

Abstract: This paper discusses the issues of robust control law design for fault-tolerant systems. Based on the assumption that the effects of faults can be expressed in linear-fractional-transformation (LFT...

Robust pole assignment in descriptor systems via proportional plus partial derivative state feedback

Abstract: The problem of eigenvalue assignment with minimum sensitivity in multivariable descriptor linear systems via proportional plus partial derivative state feedback is considered. Different from the purely proportional state feedback case, the number of finite closed-loop eigenvalues is required to be equal to the system dimension. Based on a result in the perturbation theory of generalized eigenvalue problem of matrix pairs, the closed-loop eigenvalue sensitivity measures in terms of the closed-loop normalized right and left eigenvectors are established. By combining these measures and a recently proposed general parametric eigenstructure assignment result for descriptor linear systems via proportional plus derivative state feedback, the robust pole assignment problem is converted into an independent minimization problem. A simple algorithm in sequential order is obtained for solution to the problem of robust pole assignment in descriptor linear systems via proportional plus partial derivative state feedback...

Fault Diagnosis of Non-Linear Dynamic Systems

Abstract: The majority of model-based fault diagnosis methods are based on linear system models. For non-linear systems, the fault diagnosis problem has been traditionally approached in two steps. Firstly, the model is linearized at an operating point, and then robust techniques are applied to generate residual signals which are insensitive to model parameter variations within a small neighborhood of the operating point. The robustness issue is tackled using techniques developed for linear system models. The strategy only works well when the linearization does not cause a large mismatch between linear model and non-linear behavior, the residual has been designed to be robust enough to tolerate small model perturbations around the operating point, and the system closely operates around the operating point specified. However, for systems with high nonlinearity and a wide dynamic operating range, the linearized approach fails to give satisfactory results. A linearized model is an approximate description of the non-linear system dynamics around the operating point. However, when the system operating range becomes wider, the linearized model is no longer able to represent the system dynamics. One solution is to use a large number of linearized models corresponding to a range of operating points. However, this would involve a large number of FDI systems corresponding to all operating points. This is not very practical for real-time application.

Non-linear dynamic systems fault detection and isolation using fuzzy observers:

Abstract: This paper presents a novel fault detection and isolation scheme for non-linear dynamic systems. This scheme utilizes a fuzzy observer to generate the diagnostic residual signal for fault detection and isolation. This fuzzy observer, based upon the idea of Takagi-Sugeno fuzzy models, comprises a number of locally linear observers and the final state estimate is a fuzzy fusion of all local observer outputs. To ensure good estimation performance, the eigenvalues of the fuzzy observer are assigned in a pre-defined region in the complex plane. The stability as well as eigenvalue constraint conditions for the fuzzy observer design are presented and solved in the linear matrix inequality framework in this paper. Finally, the paper demonstrates the application of fuzzy observers in detecting and isolating intermittent faults in the induction motor of a railway traction system using a real-time test-rig implementation.

Robust Residual Generation Using Unknown Input Observers

Abstract: The generation of robust residuals is the most important task in model-based fault diagnosis techniques. As pointed out in Section 2.11, the disturbance de-coupling based approaches are the dominant approaches for robust residual generation. For those approaches, uncertain factors in system modeling are considered to act via an unknown input (or disturbance) on a linear system model. Although the unknown input vector is unknown, its distribution matrix is assumed known. Based on the information given by the distribution matrix, the unknown input (disturbance) can be de-coupled from the residual. Robust FDI is thus achievable using the disturbance de-coupled residual. This chapter focuses on the robust residual generation problem via unknown input observers. The principle of the unknown input observer (UIO) is to make the state estimation error de-coupled from the unknown inputs (disturbances). In this way, the residual can also be de-coupled from each disturbance, as the residual is defined as a weighted output estimation error. This approach was originally proposed by Watanabe and Himmelblau (1982) who considered the robust sensor fault detection and isolation problem for the system with modeling uncertainty. Later, Wunnenberg & Frank (Wunnenberg and Frank, 1987; Frank and Wunnenberg, 1989; Wunnenberg, 1990) generalized this approach for detecting and isolating both sensor and actuator faults by considering the case when unknown inputs also appear in the output equation. In parallel with this development, a robust scheme for diagnosing actuator faults via UIOs is proposed by Chen and Zhang (1991). A very important contribution of the paper by Chen and Zhang (1991) was to demonstrate the robust FDI approach via to a realistic chemical process system example. Note that Viswanadham and Srichander (1987) and Phatak and Viswanadham (1988) also studied the actuator fault detection and isolation problem via UIOs, however they failed to consider robustness issues. Many other investigators have considered the use of UIOs for robust FDI: e.g. Hou and Muller (1991), Hou and Muller (1994b), Frank and Seliger (1991), Seliger and Frank (1991a), Keller, Nowakowski and Darouach (1992), Chang and Hsu (1993a), Ragot, Maquin and Kratz (1993), Saif and Guan (1993), Wang and Daley (1993), Chen and Patton (1994b), Shields (1994), Yu, Shields and Mahtani (1994b), Yu and Shields (1996) and Hwang, Chang and Hsu (1997).

A multiple model approach to fault-tolerant control using Takagi-Sugeno fuzzy modelling: Real application to an induction motor drive system

Abstract: This paper investigates the development of a Supervisory Control System with qualitative tasks at the upper level and simple quantitative models at the lower level to control complex non-linear systems. A new quantitative approach for the stability of non-linear fuzzy inference systems using Takagi-Sugeno (T-S) fuzzy models is presented. The necessary conditions for the assignability of eigenvalues to a region in the s-plane and the necessary conditions to guarantee the stability of fuzzy models for both T-S fuzzy observers and T-S fuzzy controllers together are derived. The problem is solved via the Linear Matrix Inequality (LMI) method. The paper demonstrates the feasibility of this system architecture through application to a fault-tolerant design for a railway traction system using DSP in a hardware test-rig.

Basic Principles of Model-Based Fault Diagnosis

Abstract: The model-based approach to fault diagnosis in automated processes has been receiving considerable attention since the beginning of 1970s, both in a research context and also in the domain of application studies on real processes. There are a great variety of methods in the literature, based on the use of mathematical models of the monitored processes and modern control theory.

Fault diagnosis of the 3 tank system using fuzzy multiple inference modelling

Abstract: This paper proposes a new approach for detecting and isolating faults in a non-linear dynamic process using the multiple model approach. The necessary conditions for the assignability of eigenvalues to a region in the s-plane and the necessary conditions to guarantee the stability of fuzzy models for Takagi-Sugeno (T-S) fuzzy observers are derived. The paper is structured in two stages. The first is to generate residual signals based on a comparison between the actual and estimated states and in the second stage the Takagi-Sugeno fuzzy observer is reconfigured to identify not only the faults but also the degree or severity of each fault. The techniques developed are demonstrated using a laboratory three tank system.

Supervisory fault tolerant system using fuzzy multiple inference modelling

Abstract: This paper presents a novel approach to integrating quantitative and qualitative information in fault-diagnosis. This paper investigates the development of a supervisory control scheme for a non-linear system with qualitative tasks at the upper level and a lower level comprising quantitative model based non-linear control. A new quantitative approach for the stability of non-linear fuzzy inference systems using Takagi-Sugeno (T-S) fuzzy models is presented. The necessary conditions for the assignability of eigenvalues to a region in the s-plane and the necessary conditions to guarantee the joint stability of fuzzy models for combined T-S fuzzy observers and T-S fuzzy controllers together are derived.

Robust Residual Generation by the Assignment of Observer Eigenstructure

Abstract: In Chapter 3, various approaches for generating robust residual via unknown input observers have been introduced. The underlying principle of these approaches is to make the state estimation error be independent of disturbances (or unknown inputs). The residual is defined as the (weighted) output estimation error which is a linear transformation of the state estimation error. The residual generated by UIOs is also independent of disturbances, if the disturbance term does not appear in the output equation or the disturbance term in the output equation has been nulled. In model-based FDI, the state estimation is not necessarily needed, because the required information is the diagnostic signal - residual. Hence, it is not necessary to de-couple the state estimation error from disturbances in model-based FDI. A direct approach to design disturbance de-coupled residuals is then required. In this approach, the residual itself is de-coupled from disturbances, however the state estimation error may not be. It can be expected that existing conditions for such a direct approach could be relaxed compared with those required for UIOs.

Robust Residual Generator Design Via Multi-Objective Optimization and Genetic Algorithms

Abstract: To ensure reliable operation of control systems, hard faults in system components are not tolerable and must be detected before they actually occur. Hopefully, faults are detected during the maintenance stage. However, the situation is different for soft (incipient) faults. Their effect on the system is very small and almost unnoticeable during their incipient stage. They may develop slowly to cause very serious effects on the system, although these incipient faults may be tolerable when they first appear. Hence, the most important issue of reliable system operation is to detect and isolate incipient faults as early as possible. An early indication of incipient faults can give the operator enough information and time to take proper measures to prevent any serious consequence on the system.

Fault diagnosis and fault-tolerant estimation of a rail traction system via fuzzy observers:

Abstract: This paper studies the fault diagnosis and fault-tolerant estimation problems of a rail traction system. The main idea is to use a bank of fuzzy observers to detect faults and to isolate failed sensors. Once the failed sensor is isolated, the fuzzy observer driven by healthy sensors provides reliable torque and flux estimates which are used for the control purpose. The concept and design procedures for fuzzy observers are presented in this paper. A fuzzy observer is able to produce accurate state estimations and generate robust residuals for non-linear dynamic systems such as a rail traction system. The fault diagnosis and fault-tolerant estimation scheme proposed for the rail traction system is demonstrated using simulations.