Understanding the concept of faults in systems, their mathematical parametrisation, and an introduction in Fault Detection and Isolation (FDI) algorithms. The familiarity of essential mathematical tools for designing observers for the detection/estimation of faults in systems. A brief review of the most popular methods used for FDI (Model based, Signal based, Knowledge based, Hybrid, Active methods). The introduction of the essential tools for model-based methods for FDI via the design of observers for the detection and diagnosis of faults (in actuators and sensors). An introduction in Lyapunov theory tools for the stabilisation of closed loop systems, and the analysis and synthesis of controllers via sufficient conditions. The design and analysis of fault tolerant control systems. The general purpose of the course includes the analysis and use of mathematical formulas, description and preparation of sketches, simple drawings and diagrams, the interpretation of results from various sources and measurements, and the necessary corrective actions when required.

Module material for theory and lab: The material for the module is uploaded on a weekly basis in the platform moodle. This can be found at URL . For acess to the material do not hesitate to contact me via email at gkladis [AT] puas [DOT] gr .

Description / Outline: 

The goal of the module Monitoring and Detection of faults in Dynamical Systems is to provide the necessary tools to the students so that they are able to design and implement algorithms to detect and diagnose faults in systems. The systems investigated are in state space form and are subject to modeling inacurracies, noise in sensors and faults occurring in actuators/sensors, in the structure of the system (and their combinations). The architecture of the algorithms used for the fault detection and diagnosis includes appropriately designed observers. Their design is performed utilising the error dynamics of the estimated state whilist using tools from Lyapunov theory where the stability of the system can be guaranteed. Observers are an important and essential part of all modern automatic control systems since their goal is the parameter and state estimation, even in the case that the system is subject to disturbances, uncertainty, noise sensors, and possibly faults. In any modern industrial environment, the need for detection, diagnosis, prediction and resolution of faults is imperative, whereby an automation engineer must be able to overcome these challenges in such an environment.


[1] Isermann, R. (2006). Fault-diagnosis systems: An introduction from fault detection to fault tolerance. Berlin, Germany: Springer.
[2] C. Edwards, T. Lombaerts, and H. Smaili, Fault Tolerant Flight Control: A Benchmark Challenge. London: Springer-Verlag, 2010. (Chapters 1-5)
[3] Zhang Ke, Jiang Bin, and Shi Peng. Observer-Based Fault Estimation and Accommodation for Dynamic Systems. Springer Berlin / Heidelberg, 2013. (chapters 1-4)
[4] Paraskevopoulos, P.N.: Modern Control Engineering. Marcel Dekker, NewYork, (2002). (Chapters 2,5,10)
[5] H. K. Khalil, Nonlinear Systems, third edition ed. New Jersey: Prentice Hall, Inc, 2002. (Chapter 1-3)

21/11/2017 - 12:00

31/10/2017 - 12:00

10/10/2017 - 12:00

16/01/2017 - 16:45

Ενότητα 12η: Διαμερισμένη στρατηγική για εντοπισμό και διάγνωση σφαλμάτων σε σμήνη οχημάτων

16/01/2017 - 16:30

Ενότητα 11η: Εντοπισμός και Διάγνωση σφαλμάτων σε μη γραμμικά συστήματα

16/01/2017 - 16:15

Ενότητα 10η: Εντοπισμός και Διάγνωση σφαλμάτων σε Lipschitz μη γραμμικά συστήματα

13/12/2016 - 12:30

Ενότητα 8η: Εντοπισμός και Διάγνωση Χρονικά Μεταβαλλόμενων Σφαλμάτων (I)
Ενότητα 9η: Εντοπισμός και Διάγνωση Χρονικά Μεταβαλλόμενων Σφαλμάτων (II)

29/11/2016 - 12:30

Ενότητα 6η: Εντοπισμός και Διάγνωση Χρονικά Αμετάβλητων Σφαλμάτων (I)
Ενότητα 7η: Εντοπισμός και Διάγνωση Χρονικά Αμετάβλητων Σφαλμάτων (II)

15/11/2016 - 16:15

Ενότητα 5η : Εισαγωγή στον Εντοπισμό και Διάγνωση Σφαλμάτων - Η model based τεχνική