PhD defence by Glenn Andreasen


02.02.2021 kl. 13.00 - 16.00


“Data-driven Fault Tolerant Control of Refrigeration Systems”

Supermarket refrigeration systems are complex dynamic systems with many interacting subsystems. These systems vary dependent on refrigerant, size, system configurations e.g. with/without air condition, ejectors, etc.  Furthermore, these systems are exposed to changing system conditions such as weather or day and night operation. These factors are often unknown prior to the design of control-oriented applications such as fault detection, fault diagnosis, and fault tolerant control solutions. To cope with these factors, data-driven methods are an appealing approach for design of control-oriented applications.

The main focus of this work is to obtain information of the unknown supermarket refrigeration system by utilizing data-driven methods (i.e. system identification), and applying these findings in fault detection, fault diagnosis, and fault tolerant control solutions.

One of the challenges for data-driven methods, addressed in this thesis, is to derive a data-driven method that is suitable for industrial controller units with limited memory and computational power. The data-driven method applied in this work is utilized for addressing two different fault faults: refrigerant leakage and faulty heating of an ambient temperature sensor.

Assessment Committee
Professor, Steven X. Ding, University of Duisburg-Essen, Germany
Automation Engineer, Christoph Josef Backi, BASF SE, Germany
Associate Professor, Zhenyu Yang, Aalborg University (Chairman)

Professor, Jakob Stoustrup, Aalborg University, (main supervisor)
Professor, Roozbeh Izadi-Zamanabadi, Aalborg University & Lead expert, Innovation lab, Danfoss, (industrial supervisor)

Associate Professor, Kirsten M. Nielsen, Aalborg University


Automation and Control Section, Department of Electronic Systems



Registration Deadline

01.02.2021 kl. 12.00

Register at