Event-triggered fault detection for a class of discrete-time linear systems using interval observers

Area of Science:

• Control Systems Engineering
• Systems Theory
• Signal Processing

Background:

• Fault detection (FD) is crucial for system reliability and safety.
• Existing methods often incur high communication costs.
• Event-triggered mechanisms offer a potential solution to reduce communication load.

Purpose of the Study:

• To develop a novel event-triggered fault detection scheme for discrete-time linear systems.
• To improve robustness against disturbances and enhance fault sensitivity.
• To reduce the information communication burden while guaranteeing FD performance.

Main Methods:

• An event-triggered interval observer is proposed to generate residuals, considering disturbances and event errors.
• l1 and H∞ performances are introduced to enhance residual robustness and fault sensitivity.
• Dilated linear matrix inequalities and slack matrix variables are used for decoupling and general Lyapunov functions.

Main Results:

• The proposed scheme effectively generates residuals robust to disturbances and sensitive to faults.
• The event-triggering mechanism significantly reduces information communication.
• The FD decision scheme, based on residual intervals, proves effective.
• Simulation results validate the proposed fault detection method's effectiveness.

Conclusions:

• The developed event-triggered FD scheme successfully balances communication efficiency and detection performance.
• The method provides a robust and sensitive approach to fault detection in discrete-time linear systems.
• This work contributes to more efficient and reliable control system monitoring.