FAULT DIAGNOSIS AND RECONFIGURATION IN FLIGHT CONTROL SYSTEMS

Transcription

1 FAULT DIAGNOSIS AND RECONFIGURATION IN FLIGHT CONTROL SYSTEMS by CHINGIZ HAJIYEV Istanbul Technical University, Turkey and FIKRET CALISKAN Istanbul Technical University, Turkey Kluwer Academic Publishers Boston/Dordrecht/London

2 Contents List of Figures List of Tables Preface xiii xix xxi 1. INTRODUCTION Introduction Fault Detection and Isolation Redundancy Performance of FDI Robustness Reconfiguration and Fault Tolerant Flight Control Systems Self-repairing Flight Control Systems Fault Detection and Dynamic Reconfiguration Analytical Redundancy Game Theory and Real-time Parameter Insensitive Disturbance Attenuating Redesign Techniques for Reconfiguration Sensor/Actuator Fault Diagnosis Sensor Fault Detection in Flight Control Systems Integrated Sensor/Actuator FDI and Reconfigurable Control for Fault-Tolerant Flight Control System Design 20 References GENERAL THEORY OF OBSERVERS Introduction Theory and Design of Observers Approaches to Observer Design Parameter Insensitive Observers Robustness of Observer-Based Controllers 39

3 Vlll An Observer Adjustment Procedure Robust Observers: Doyle-Stein Observer Failure Insensitive Luenberger Observer Conclusion 54 References LINEAR KALMAN FILTERING Introduction The Optimum Linear Discrete Kalman F :!ter Optimum Discrete Kalman Filter Stability Discrete Kalman Filter - Correlated System and Measurement Noise General Discrete Kalman Filtering Divergence in the Kalman Filler. Suboptimal Kalman Filters The Methods of Numerical Stabilization of Kalman Filters. Square Root Filtering Diagnostics of Kalman Filtering Algorithms Monitoring Conditions for the Scalar Filter Diagnostics of the Square Root Filter Summary 83 References NONLINEAR ESTIMATION AND SYSTEM IDENTIFICATION Introduction Linearized Kalman Filter The Extended Kalman Filter Derivation of EKF Equations for Nonlinear Flight Dynamics Derivation of the Mathematical Model of the Aircraft Longitudinal Motion Design of EKF to Estimate the Aircraft Longitudinal Motion EKF and System Identification An Example of Parameter Identification Stopping Rules Formation and Faults Detection in Parametric Identification Problems Generation of Stopping Rules Fault Detection in the Kalman Filter Computation of the Covariance matrix of the Discrepancy Between Two Successive Estimates An Experimental Example of Parameter Identification Conclusion and Discussion 119 References 120

4 IX 5. INTRODUCTION TO MONITORS, DETECTORS, AND DIAGNOSTICANS Introduction Model-Free Methods Model-based fault diagnosis Monitors The parity space approach Parameter identification approach Dedicated observer approach Detectors Hypothesis Testing Neymann-Pearson Criterion Diagnosticians Summary 151 References AN OVERVIEW OF ALGORITHMS FOR FAILURE DETECTION AND ISOLATION Introduction 6.2. Failure Detection Techniques Sensor Level Failures Failures in Computing Subsystem and Lanes Actuator and Control Surface Level Failures Robustness of Failure Detection Process to Parameter Changes in Dynamic Systems Summary 183 References THE INNOVATION APPROACH TO FAULT DETECTION Introduction The Innovation Sequence The Methods for Checking the Statistical Characteristics of Innovation Sequence The Appropriateness Criterions of Innovation Sequence to the White Noise The Criterions, Checking the Mean of the Innovation Sequence The Classical Criterions, Checking the Covariance Matrices of the Innovation Sequence Some New Methods for Checking the Covariance Matrices of Innovation Sequence Checking the Covariance Matrices of the Innovation Sequence via the Statistics of Relation Between Two Quadratic Forms Check of the Generalized Variance of the Normalized Innovation Sequence Simultaneous Checking the Mean and Covariance Matrices of the Innovation Sequence 211

5 7.5. Conclusion and Discussion 220 References SENSOR FAULT DETECTION AND ISOLATION IN FLIGHT CONTROL SYSTEMS BASED ON INNOVATION APPROACH Introduction Innovation Sequence Application to Aircraft Sensor Fault Detection Simulation of Aircraft Dynamics via Kalman Filter Fault Detection by Verifying the Innovation Mean Fault Detection by Verifying the Innovation Covariance Comparison of Checking Covariance Matrix Algorithms Fault Detection via Spectral Norm of Innovation Matrix Sensor Fault Isolation Algorithm Aircraft Application of Sensor Fault Isolation Algorithm Conclusion and Discussion 254 References ROBUST KALMAN FILTERS FOR FDI Introduction Robust Kalman Filter Insensitive to Actuator Failures Simulation of RKF based on Doyle-Stein condition Robust Kalman Filter Insensitive to Abnormal Measurements Robust Kalman Filter Insensitive to Sensor Failures Actuator Fault Diagnosis based on Robust Kalman Filter Kalman Filter for Actuator Parameter Estimation Actuator Fault Detection Algorithm An illustrative example Conclusion and Discussion 280 References FLIGHT CONTROL RECONFIGURATION Introduction An Automatic Redesign Approach for Restructurable Control Systems Introduction The Automatic Redesign Procedure Development of the Automatic Redesign Procedure Solution of the Optimization Problem 291

6 CH. HAJIYEV AND F. CALISKAN XI Control Systems Insensitivity to Variations in Dynamic System Parameters Reconfiguration Based on Min-max Type Criteria The Min-max Principle Parameter Insensitive, Disturbance Attenuating Redesign Simulation of the Reconfiguration Algorithm The EKF Based Reconfigurable Control Identification of the Control Distribution Matrix The EKF Based Reconfigurable Control Algorithm Fault-Tolerant Flight Control System Based on Innovation Approach Conclusion and Discussion 324 References 324 Index 327 About the Authors 341