PhD Courses in Denmark

Active Fault-Tolerant Control: Theory and Applications (2024)

Doctoral School of Engineering and Science at Aalborg University

Organiser:        Associate Professor Zhenyu Yang, AAU Energy

Lecturers:         Professor Roozbeh Izadi-Zamanabadi, Dept of Electronics Systems, AAU.

                          Associate Professor Zhenyu Yang, AAU Energy, AAU

ECTS:                  4.0

Date:                  week 20, 13 May – 17 May 2024

Place:                 Aalborg University Esbjerg, Niels Bohrs Vej 8, 6700 Esbjerg

Deadline:           22 April 2024

Format:             hybrid (online & in person)

Max no. of participants: 30

 

  • As a part-time Professor at AAU, Dr. Izadi-Zamanabadi is also a Lead Control Expert at Danfoss A/S. He has over 30 years research and development experiences on FDD and FTC, as well as applications of advanced FDD and FTC in diverse industrial systems.

 

  • Dr. Yang has over 30 years research and development experiences in the FDD and FTC areas, including both traditional model-based methods and data-driven learning methods.
     

Description: A Fault Tolerant Control (FTC) system is referred to a controlled system that poses the capability to accommodate system component faults/failures automatically and is capable of maintaining overall system stability and acceptable performance in the event of such faults. An active FTC approach often consists of two integrated online functionalities, i.e., Fault Detection and Diagnosis (FDD) and Control Reconfiguration (CR). The FTC plays an essential rule in safe-critical systems, which now cover a wide range of engineering systems, from aircrafts, airspace systems and nuclear reactors, to the recent emerging energy systems, such as smart grid systems, offshore wind farms etc. The ultimate objective of applying FTC techniques is to cost-effectively increase engineering system’s reliability, safety availability and maintainability. This course covered the fundamental essentials and some latest results in active FTC research area, along with diverse application case studies. The course daily plan consists of about 4-hour lectures and 2-hour exercises every day, such as

 

  • o Kl.8.15-9.15 lecture-d-1
  • o Kl.9.30-10.30 lecture-d-2
  • o Kl.10.45-11.45 exercise-d-1
  • o Kl.11.45-12.30 lunch
  • o Kl.12.30-13.30 lecture-d-3
  • o Kl.13.45-14.45 lecture-d-4
  • o Kl.15.00-16.00 exercise-d-2
     
  • Day 1: Introduction & State-estimation-based FDD (Zhenyu Yang, 4-hour lectures & 2-hour exercises)
    • • Lec-1-1: General introduction of FTC & Active FTC
    • • Lec-1-2: Overview of FDD principles & methodologies
    • • Lec-1-3: Observer-based FDD (incl. single & multi-bank observer based methods)
    • • Lec-1-4: Kalman-Filter based FDD (incl. single & multi-bank KF based methods)
  • Day 2: FMEA and Structual Analysis (SA) method (Roozbeh Izadi-Zamanabadi, 4-hour lectures & 2-hour exercises)
    • • Lec-2-1: Failure Mode and Effect Analysis (FMEA)
    • • Lec-2-2: Introduction to Structural Analysis (SA)
    • • Lec-2-3: SA principles & computations
    • • Lec-2-4: SA applications & practical issues
  • Day 3: Data-driven modeling and FTC methods (Roozbeh Izadi-Zamanabadi, 4-hour lectures & 2- hour exercises)
    • • Lec-3-1: Introduction to data-driven approaches and methods
    • • Lec-3-2: statistical approach to feature extraction for anomaly detection based on latent projection methods
    • • Lec-3-3: Active data-driven estimation/FDD methods
    • • Lec-3-4: Data-driven FDD/FTC applications
  • Day 4: Unknown input observer & Machine-learning based FDD methods (Roozbeh Izadi- Zamanabadi & Zhenyu Yang, 4-hour lectures & 2-hour exercises)
    • • Lec-4-1 (Riz): Introduction to unknown-input observer (UIO) method
    • • Lec-4-2 (Riz): UIO extended for FDD
    • • Lec-4-3 (ZY): Parameter-estimation based FDD
    • • Lec-4-4 (ZY): Machine-learning based FDD
  • Day 5: Advanced FTC design & applications (Zhenyu Yang, 4-hour lectures & 2-hour exercises)
    • • Lec-5-1: Control Reconfiguration (CR) principles and integration with FDD
    • • Lec-5-2: H∞ control for robust reconfigurable control mixer design
    • • Lec-5-3: Eigenstructure assignment method for FDD/FTC design
    • • Lec-5-4: summary & wrap-up

 

Prerequisites: The participants need to have the fundamental knowledge about classical control theory (transfer-function based) and modern control theory (state-space based).

Form of evaluation: The course will be evaluated based on a mini-project report after the course. This mini-project expects some illustration that the student can apply some techniques/knowledge learned from this course into their current PhD project.

 

Price: 6000 DKK for PhD students outside of Denmark and 8000 DKK for the Industry excl. VAT
The Danish universities have entered into an agreement that allows PhD students at a Danish university (except Copenhagen Business School) the opportunity to free of charge take a subject-specific course at another Danish university. Guests at AAU Energy can free attend the course. 

Payment: A Online link for externals participants will be annonced after deadline for registration

Course literature: The course materials will be lecture slides, scientific papers and lecture notes, which will be distributed to participants before/along lectures. The following two books are recommended as further optional readings if the course participant wish to gain more knowledge in this subject, but they are not mandatory:

 

  • Jie Chen and R. Patton, “Robust model-based fault diagnosis for dynamic systems”, Kluwer, Boston, 1999.

 

 

  • R. Isermann, “Fault-diagnosis systems: An introduction from fault detection to fault tolerant”, Springer, 2006. ISBN 3-540-24112-4.
Back