Fault tolerant control for linear parameter varying systems: An improved robust virtual actuator and sensor approach

Area of Science:

• Control Systems Engineering
• Systems Theory
• Applied Mathematics

Background:

• Linear Parameter Varying (LPV) systems are widely used to model complex dynamic systems.
• Fault-Tolerant Control (FTC) is crucial for maintaining system performance and safety under component failures.
• Existing FTC methods often struggle with parameter-dependent system matrices.

Purpose of the Study:

• To develop a robust FTC methodology for discrete-time LPV systems.
• To design virtual sensors and actuators capable of reconfiguration.
• To guarantee Input-to-State Stability (ISS) despite parameter variations and faults.

Main Methods:

• Utilizing Linear Matrix Inequalities (LMIs) for synthesizing a fault-reconfiguration block.
• Developing a single reconfiguration block robust to diverse fault types and magnitudes.
• Applying the methodology to parameter-dependent input and output matrices in LPV models.

Main Results:

• Successfully synthesized a virtual actuator and virtual sensor reconfiguration block.
• Demonstrated robustness to parameter-dependent matrices in LPV systems.
• Achieved superior fault mitigation performance compared to existing literature approaches in experimental validation.

Conclusions:

• The proposed FTC methodology offers a robust and effective solution for LPV systems with parameter-dependent dynamics.
• The developed virtual sensor and actuator approach provides enhanced fault tolerance and stability.
• Experimental results confirm the superiority of this method in mitigating sensor and actuator faults.