Scholarlink

Introduction <p>This research presents an advanced structural health monitoring (SHM) approach that combines fiber Bragg grating (FBG) sensors in self-referencing configuration for guided wave (GW) sensing and machine learning for damage detection in a plate structure. FBG sensors offer unique advantages such as small size, low weight, multiplexing capability, and resistance to harsh environmental conditions. When used with edge filtering and remote bonding configurations, these sensors become suitable for high-frequency, dynamic measurements, overcoming the limitations of traditional wavelength-based FBG sensing.</p> Objectives <p>In the current work, a reference-free damage detection and localization strategy is proposed and a proof-of-concept is provided. Two case studies of damage localization on an aluminum plate are presented. In the first case, the plate is instrumented with 1 FBG and two actuator and damage is introduced on the line between the self-referenced bonds. Data driven techniques are then utilized for the damage localization. In the second case, the same plate is instrumented with four PZT actuators and two FBG sensors in orthogonal self-referencing configuration. Several different damage scenarios (1 healthy and 36 different damage scenarios) are simulated, and data driven approaches are employed for achieving damage detection and localization in 2D space without the need for baseline subtraction.</p> Methods <p>The data-driven approach makes use of principal component analysis (PCA) to reduce dimensionality and extract dominant features. A decision tree classifier is then trained to classify different damage states.</p> Results <p>The results show that this data-driven technique offers high classification accuracy and effective damage localization. Importantly, the system operates without requiring baseline data subtraction, which is especially valuable for aging or previously unmonitored structures.</p> Novelty <p>This paper is an experimental proof-of-concept of a novel sensor deployment (self-referencing) strategy and to the best of the author’s knowledge is the first implementation of the setup. The study demonstrates the potential of integrating optical fiber sensing with machine learning for scalable, efficient, and reliable SHM applications in civil and aerospace structures.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Reference-free Guided Wave-based Damage Localization using Fiber Bragg Grating Sensors in Self-Referencing Configuration: A Data-Driven Approach

  • 期刊论文

  • 发表日期：

    2026年2月18日

查看全文
  • Farzam Omidi Moaf,
  • Revati M. Wahul,
  • Abhishek Patange,
  • Rohan Soman

摘要

Introduction

This research presents an advanced structural health monitoring (SHM) approach that combines fiber Bragg grating (FBG) sensors in self-referencing configuration for guided wave (GW) sensing and machine learning for damage detection in a plate structure. FBG sensors offer unique advantages such as small size, low weight, multiplexing capability, and resistance to harsh environmental conditions. When used with edge filtering and remote bonding configurations, these sensors become suitable for high-frequency, dynamic measurements, overcoming the limitations of traditional wavelength-based FBG sensing.

Objectives

In the current work, a reference-free damage detection and localization strategy is proposed and a proof-of-concept is provided. Two case studies of damage localization on an aluminum plate are presented. In the first case, the plate is instrumented with 1 FBG and two actuator and damage is introduced on the line between the self-referenced bonds. Data driven techniques are then utilized for the damage localization. In the second case, the same plate is instrumented with four PZT actuators and two FBG sensors in orthogonal self-referencing configuration. Several different damage scenarios (1 healthy and 36 different damage scenarios) are simulated, and data driven approaches are employed for achieving damage detection and localization in 2D space without the need for baseline subtraction.

Methods

The data-driven approach makes use of principal component analysis (PCA) to reduce dimensionality and extract dominant features. A decision tree classifier is then trained to classify different damage states.

Results

The results show that this data-driven technique offers high classification accuracy and effective damage localization. Importantly, the system operates without requiring baseline data subtraction, which is especially valuable for aging or previously unmonitored structures.

Novelty

This paper is an experimental proof-of-concept of a novel sensor deployment (self-referencing) strategy and to the best of the author’s knowledge is the first implementation of the setup. The study demonstrates the potential of integrating optical fiber sensing with machine learning for scalable, efficient, and reliable SHM applications in civil and aerospace structures.