Trainee Project
Title:
Data-driven fault-tolerant control (FTC) with application to the Eco-sûr platform
Dates:
2024/11/04 - 2025/04/30
Student:
Supervisor(s):
Description:
The design of automatic controls based on the modeling of systems by physical principles and laws is becoming an important issue with the growing complexity of modern industrial systems (aerospace systems, transport, intelligent power grids, buildings, etc.). On the other hand, technological advances and the omnipresence of sensors and digital communication networks (embedded systems, cyber-physical systems, connected objects, internet, etc.) are accompanied by a vast amount of data available in real time and off-line, with the major challenge of analyzing and exploiting this data to optimize decisions and develop new, more effective applications through data-driven learning. This use of data is currently strongly boosted by the emergence of artificial intelligence techniques in automatic control. It is in this context that the proposed topic seeks to harness the benefits of the data-driven revolution for the design and implementation of optimal controls for fault-tolerant control and dependability of dynamic systems.
The first phase of this research internship will consist of a state-of-the-art review of recent work and results on various approaches to data-driven optimal and sub-optimal control. It will then focus and/or develop work on optimal control techniques in one or other of the promising paradigms highlighted in the state-of-the-art for fault-tolerant control objectives. Particular attention will be paid to behavioral system theory techniques, which provide an appropriate mathematical framework for direct data-driven analysis and synthesis.
The application part of the work will be devoted to simulations in the MATLAB/Simulink environment on aspects of optimal building control (e.g., thermal comfort control, air quality (CO2) control, demand response management control, ...) and fault tolerance of heating, ventilation and air-conditioning (HVAC) equipment. The results will be validated on CRAN's newly-built Eco-Sûr platform, dedicated to control issues for building energy efficiency. This platform is equipped with numerous measurement systems providing data on a wide range of variables relating to indoor and outdoor climatic conditions in buildings, with a view to monitoring and controlling them for optimum energy consumption.
The first phase of this research internship will consist of a state-of-the-art review of recent work and results on various approaches to data-driven optimal and sub-optimal control. It will then focus and/or develop work on optimal control techniques in one or other of the promising paradigms highlighted in the state-of-the-art for fault-tolerant control objectives. Particular attention will be paid to behavioral system theory techniques, which provide an appropriate mathematical framework for direct data-driven analysis and synthesis.
The application part of the work will be devoted to simulations in the MATLAB/Simulink environment on aspects of optimal building control (e.g., thermal comfort control, air quality (CO2) control, demand response management control, ...) and fault tolerance of heating, ventilation and air-conditioning (HVAC) equipment. The results will be validated on CRAN's newly-built Eco-Sûr platform, dedicated to control issues for building energy efficiency. This platform is equipped with numerous measurement systems providing data on a wide range of variables relating to indoor and outdoor climatic conditions in buildings, with a view to monitoring and controlling them for optimum energy consumption.
Keywords:
data-based control, optimal control, fault tolerance, optimization, energy efficiency
Department(s):
| Control Identification Diagnosis |
Publications: