Self-healing FBG sensor network fault-detection based on a multi-class SVM algorithm.

Area of Science:

• Optical sensor networks
• Fiber optic sensing technology
• Network reliability and fault tolerance

Background:

• Fiber Bragg Grating (FBG) sensors are crucial for monitoring in various applications.
• Traditional FBG networks face challenges with fault detection and network resilience.
• Need for robust self-healing capabilities and accurate fault identification in FBG sensor systems.

Purpose of the Study:

• To propose and evaluate a novel three-layer ring architecture for FBG sensor networks.
• To enhance the network's reconfigurable capabilities and self-healing properties.
• To develop and validate methods for soft fault detection and reliability estimation.

Main Methods:

• Implementation of a three-layer ring architecture with enhanced reconfigurability.
• Development of a self-healing mechanism for resilience against multiple fiber link failures.
• Application of a multi-classification support vector machine (multi-class SVM) algorithm for soft fault detection.
• Utilization of an artificial neural network (ANN) model for reliability estimation.

Main Results:

• The proposed network demonstrates self-healing capabilities, successfully recovering from three fiber link failures.
• The multi-class SVM algorithm achieved a 99% accuracy in detecting soft faults within the FBG sensors.
• The ANN reliability estimation model accurately predicted the network's reliability at a reasonable computational cost.

Conclusions:

• The proposed three-layer ring architecture significantly enhances the resilience and fault tolerance of FBG sensor networks.
• Advanced algorithms like multi-class SVM and ANN provide effective solutions for fault detection and reliability assessment.
• This research contributes to the development of more dependable and robust optical sensing systems.