Bus network decomposition for fault detection and isolation through power line communication.

Area of Science:

• Electrical Engineering
• Computer Science
• Network Engineering

Background:

• Complex embedded wired communication networks are susceptible to soft faults that can escalate into hard faults (short or open circuits).
• Existing fault detection methods may not adequately address the complexities of these networks, particularly concerning evolving fault types.

Purpose of the Study:

• To propose and validate a novel fault detection and isolation (FDI) method for complex embedded wired communication networks.
• To address the challenge of detecting and isolating soft faults that can degrade into hard faults.

Main Methods:

• A new FDI method is developed utilizing power line communication (PLC) transmission systems.
• Transmission coefficients are estimated using orthogonal frequency division multiplexing (OFDM) and compared to reference values to compute health indicators and residuals.
• A network decomposition methodology is employed, breaking down complex networks into smaller, manageable Y-shaped sub-networks for monitoring.

Main Results:

• The proposed FDI method was successfully validated on a Y-shaped network test bench using real-world data.
• The approach demonstrated effectiveness in a more complex network scenario through realistic simulations.
• The method accurately detected and isolated faults within the tested network configurations.

Conclusions:

• The developed FDI method offers a robust solution for monitoring complex embedded wired communication networks.
• The PLC-based approach, combined with network decomposition, provides reliable fault detection and isolation capabilities.
• This research contributes to enhancing the reliability and safety of communication systems in critical applications.