Short-Training Damage Detection Method for Axially Loaded Beams Subject to Seasonal Thermal Variations.

Area of Science:

• Structural Health Monitoring (SHM)
• Vibration Analysis
• Damage Detection

Background:

• Vibration-based damage features are sensitive to structural changes but susceptible to environmental variations.
• Existing methods often require extensive training data, limiting their use in real-world scenarios with limited datasets.
• Monitoring existing or rapidly deteriorating structures poses challenges for traditional SHM approaches.

Purpose of the Study:

• To develop a novel damage index for structural health monitoring (SHM) that is robust to environmental conditions.
• To enable effective damage detection even with limited training data.
• To overcome the limitations of existing SHM strategies that rely on large datasets.

Main Methods:

• Utilized Principal Component Analysis (PCA) on vibration-based damage features.
• Developed a damage index based on PCA to filter environmental and operating effects.
• Validated the proposed index using simulated and experimental data from an axially loaded beam structure.

Main Results:

• The PCA-based damage index effectively filters temperature effects on vibration features (eigenfrequencies).
• The proposed index demonstrates capability in both short and long training set scenarios.
• The PCA-based strategy outperformed a Mahalanobis square distance-based index in temperature independence and damage sensitivity.

Conclusions:

• The novel PCA-based damage index provides a reliable method for structural health monitoring (SHM) under varying environmental conditions.
• This approach significantly enhances damage detection capabilities, especially when limited training data is available.
• The study demonstrates the superiority of the PCA-based method over traditional approaches for robust structural monitoring.