Fault diagnosis and robust control: vehicle application - 5AMAC515

Goal(s)

This course is an introduction to the detection and localisation of sensor and actuator faults for linear dynamic systems (vehicle, aircraft, DC motor,car shock absorber,tank...).We focus our attention about the Model-based Fault Detection and Isolation Techniques: parity space approach and observer based approach. A reliable fault diagnosis system should be able to distinguish faults from system disturbances, uncertain parameters and measurement noise. More precisely, the fault diagnosis system must be robust to the disturbances and uncertainties while remaining sensitive to faults. To ensure robust diagnosis and control of the system, we present different tools: H?, LMI, PI observer, observer bank, parity space....Each approach is illustrated by numerical example or real application.

Content(s)

Content:

• Fault detection and isolation
  • Parity space approach
  • Observer-based approach
• Robust control
  • Analysis of uncertain systems
  • H?control
  • Introduction to Linear Matrix Inequalities
  • Diagnosis by H? synthesis

Audience/prerequisites: the course is intended for master students in applied science with a major in control. A basic knowledge on linear control theory(state space representation)is required.

• Linear systems
• State space approach
• Frequential approach
• Optimisation