Showing papers by "Ron J. Patton published in 2001"

Robust fault detection using Luenberger-type unknown input observers-a parametric approach

Abstract: A new parametric observer-based approach for robust fault detection in multivariable linear systems with unknown disturbances is proposed. The residual is generated through utilizing a Luenberger function observer. By using a parametric solution to a class of generalized Sylvester matrix equations, a parametrization is proposed for the residual generator on the basis of a Luenberger function observer. By further properly constraining the design parameters provided in the Luenberger observer design, the effect of the unknown disturbance is decoupled from the residual signal. The proposed approach provides all the degrees of freedom and is demonstrated to be simple and effective.

Interpretation of Trained Neural Networks by Rule Extraction

Abstract: The paper focuses on the problem of rule extraction from neural networks, with the aim of transforming the knowledge captured in a trained neural network into a familiar form for human user. The ultimate purpose for us is to develop human friendly shells for neural network based systems. In the first part of the paper it is presented an approach on extracting traditional crisp rules out of the neural networks, while the last part of the paper presents how to transform the neural network into a set of fuzzy rules using an interactive fuzzy operator. The rules are extracted from ordinary neural networks, which have not a structure that facilitate the rule extraction. The neural network trained with the well known Iris data set was considered as benchmark problem.

Robust fault diagnosis in a chemical process using multiple-model approach

Abstract: Presents a robust model-based technique for the detection and isolation of sensor faults in a chemical process. The diagnosis system is based on the robust estimation of process outputs. A dynamic non-linear model of the process under investigation is obtained by a procedure exploiting Takagi-Sugeno (T-S) multiple-model fuzzy identification. The combined identification and residual generation schemes have robustness properties with respect to modelling uncertainty, disturbance and measurement noise, providing good sensitivity properties for fault detection and fault isolation. The identified system consists of a fuzzy combination of T-S models to detect changing plant operating conditions. Residual analysis and geometrical tests are then sufficient for fault detection and isolation, respectively. The procedure presented is applied to the problem of detecting and isolating faults in a benchmark simulation of a tank reactor chemical process.

Actuator fault diagnosis in a continuous stirred tank reactor using identification techniques

Abstract: This paper presents a model-based procedure for the detection and isolation of actuator faults in a chemical process. The diagnosis system is based on the estimation of process outputs. A dynamic Multi-input, multiple output (MIMO) process of the process under investigation is obtained by identification procedures, exploiting both Auto Regressive exogenous and Takagi-Sugeno (T-S) fuzzy input-output systems. Fuzzy systems are exploited to cope with different process working conditions. Residual analysis and geometrical tests are then used for fault detection and isolation, respectively. The proposed designs were evaluated using a benchmark simulation of a Continuous Stirred Tank Reactor.

Fault-tolerant control systems design using the linear matrix inequality method

Abstract: This paper discusses an approach to 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) forms, a fault-tolerant control systems design problem is formulated and solved via a linear matrix inequality (LMI)-based synthesis approach, to recover the convexity of the design problem whilst considering the robust performance and robust stability against faults and uncertainties simultaneously, a constrained optimisation approach is used. The simulation results of a design example (a longitudinal motion flight control problem for an unmanned aircraft in the case of suffering battle damage on its wing) shows that the robust stability and satisfactory performance are achievable.

Fuzzy model approximation and its SVD reduction

Abstract: This paper is concerned with the argument that the identification of Takagi Sugeno (T-S) fuzzy models from training data should involve an important feature between data fitness and model complexity. One hand a (T-S) fuzzy model with a large number of fuzzy rules may encounter the risk of having an approximation capable of fitting training data well. On the other hand it may be difficult to run this fuzzy model structure due to heavy computational cost. In order to facilitate the development of a balance between these requirements, a Higher Order Singular Value Decomposition (HOSVD) based T-S fuzzy model reduction is introduced using the well-known Yam SVD fuzzy rule-based approximation technique.

Identification and fault diagnosis of an industrial gas turbine prototype model.

Abstract: This paper addresses a model-based procedure exploiting analytical redundancy for the detection and isolation of faults of a power plant. The residual generation is performed by means of output observers and Kalman filters in connection with the uncertainty affecting the measurements acquired from the monitored system. The model of the process under investigation required to design observers and filters is obtained by identification. The proposed fault detection and isolation tool has been tested on a simulated model of an industrial gas turbine prototype.