Showing papers by "Hong Wang published in 1995"

Advanced Adaptive Control

Abstract: Index: 1. Introduction 1.1 Introduction 1.2 Adaptive Control Schemes 1.3 The Context of Adaptive Control 1.4 Book Outline 2. Preliminaries 2.1 Matrices 2.2 Eigenvalues, Eigenvectors and Eigenrows 2.3 Time-Varying Matrices 2.4 Norms and Inner Products 2.5 Dynamic System Representations 2.6 Least Squares Estimation 2.7 Technical Estimation 2.8 A Design Example 3. Artificial Neural Networks: Aspects of Modelling and Learning 3.1 Introduction 3.2 Network Approximation 3.3 Multi-Layer Perceptrons 3.4 Associative Memory Networks 3.5 Polynomial Neural Networks 3.6 The Curse of Dimensionality 3.7 Supervised Learning 3.8 Instantaneous Learning Rules 3.9 Weight Convergence 3.10 A Geometric Interpretation of the LMS Algorithm 3.11 Gradient Noise 4. Fuzzy Modelling and Control Systems 4.1 Fuzzy Systems 4.2 Neurofuzzy Systems 4.3 Adaptive Fuzzy Models 4.4 Adaptive Fuzzy Control Systems 5. Neural Network and Fuzzy Logic Based Adaptive Control 5.1 Introduction 5.2 Matching Systems 5.3 Mismatching Systems 5.4 General Systems 5.5 Fuzzy Logic Based Control 5.6 Conclusions 6. Sup Controllers and Self-Tuning Sup Regulators 6.1 Introduction 6.2 External Input Spaces 6.3 Performance Index 6.4 Sup Regulators and Output Performance 6.5 Sup Controllers and Output Performance 6.6 Sup Controllers for Non-Minimum Phase Plants 6.7 The Self-Tuning Sup Regulator Algorithm 6.8 Convergence of the Self-Tuning Algorithm 6.9 An Example 6.10 Conclusions 7. Mean Controllers and Self-Tuning Mean Regulators 7.1 Introduction 7.2 Input Spaces and Performance Index 7.3 Mean Regulators and Output Performance 7.4 Mean Controllers and Output Performance 7.5 Stability of the Closed Loop Systems 7.6 The Self-Tuning Mean Regulator Algorithm 7.7 Conclusions 8. MRAPC for Time Delay Systems 8.1 Introduction 8.2 The Smith Predictive Control 8.3 The Modified Smith Predictor Control 8.4 MRAPC Using Parametric Optimization Theory 8.5 MRAPC Using Lyapunov Stability Theory 8.6 Conclusions 9. Rule-based Adaptive Control Systems Design 9.1 Introduction 9.2 Systems Representation 9.3 Meta Identification Rules 9.4 Model Adjusting Rules 9.5 Controller Tuning via Rule Based Method 9.6 Simulation and Application 9.7 Conclusions 10. Adaptive Control of Singular Systems 10.1 Introduction 10.2 System Representation 10.3 Parameter Estimation 10.4 Preliminary Feedback Design 10.5 Adaptive Control Design 10.6 Simulation Results 10.7 Application to The Control of A Gas Turbine 10.8 Conclusions Contents

A direct adaptive neural network control for unknown nonlinear systems and its application

Abstract: Presents a direct adaptive neural network control strategy for unknown nonlinear systems which are described by an unknown NARMA model. Taking the neural network as a neuro model of the system, control signals are directly obtained by minimizing either the instant difference or the cummulative differences between a setpoint and the output of the neuro model. An application to a flow rate control system is studied and desired results are obtained.

On the application of neural networks modelling to a wet end chemical process in paper making

Abstract: This paper presents a novel approach and its application to the modelling of non-linear unknown systems via B-spline neural networks. The system is assumed to be described by a NARMAX model, which is subjected to coloured noise. A new regression formula is constructed and it has been shown that the generalised RLS algorithm can be directly applied to train the weights and produce a desirable estimate on the various modes of the noise. The selection of the knots distribution of B-spline neural networks is also considered. An algorithm has been developed which produces an optimal knot distribution for an input axis using training data. Finally, the method developed is applied to build up a local model for a paper machine wet end, which represents the relationship between added chemicals (rosin and alum) and the sizing of the resulting paper. Desired results are obtained.

A parity space approach to robust fault detection in generalized state space systems

Abstract: Many sophisticated approaches have been proposed for fault detection in dynamic systems using standard model representations. In this paper, redundancy relations for fault detection in systems represented by a generalized state space description are presented. Conditions for the existence of a redundancy relation and the simplest form this can take for the generalized system are obtained. The design of the redundancy relation in the presence of model uncertainties is also considered and conditions for the existence of a perfectly robust solution obtained. For cases where this does not exist a procedure is defined for calculating an optimally robust solution. A simulation of a three-shaft gas turbine is used to demonstrate the potential of the method.

Adaptive feedback control for unevenly sampled systems

Abstract: This paper presents an adaptive control algorithm for a new type of discrete-time system that is unknown and is sampled unevenly. Such systems are obtained by unevenly sampling the unknown continuous time plant with two sequential sampling periods (T1 and T2), and are therefore referred to as unknown S(T1, T2) systems. This paper discusses their representation and stability when that system incorporates a known time delay. A generalized causal adaptive one-step-ahead controller is developed and it is shown to be able to stabilize the unknown S(T1, T2) system with respect to a uniformly bounded signal. The application to a typical S(T1, T2) system: the cross-directional basis weight control system in a paper making process, is described. Desirable results are obtained.

Minimum structured processes modelling: a rule-based approach

Abstract: This paper presents a simple method for minimum structured process modelling via a rule-based technique. Under the assumption that the processes are stable, step input signals are applied and step responses data are collected. Six different types of minimum structured models are then defined. They are Monotone Response (MR), Undershoot Monotone Response (UMR), Oscillatory Response (OR), Undershoot Oscillatory Response (UOR), Odd Undershoot Monotone Response (OUMR), and Odd Undershoot Oscillatory Response (OUOR). A classifier is built which, for a given step response, generates the information about the type of response. A proposed minimum structured model is then obtained. Different minimum structured models are used to achieve the initial fitting for the given response. This leads to the rough tuning of model parameters. Finally, a rule-based fine tuner is constructed and used to find out the accurate parameters of the proposed model. Desirable results are obtained when the method is applied to the modelling of the machine direction weight profile in a paper-making process and the speed control system of a hydraulic turbine generator.

A direct adaptive neural network control for unknown nonlinear systems

Abstract: Presents a direct adaptive neural network control strategy for unknown nonlinear systems which are described by an unknown NARMA model. Taking the neural network as a neuro model of the system, control signals are directly obtained by minimizing either the instant difference or the cummulative differences between a setpoint and the output of the neuro model. An application to a flow rate control system is studied and desired results are obtained.