In Situ Damage Detection Method for Metallic Shear Plate Dampers Based on the Active Sensing Method and Machine...

Area of Science:

• Structural Engineering
• Materials Science
• Signal Processing

Background:

• Metallic Shear Plate Dampers (MSPDs) are crucial for passive vibration control.
• Current visual inspection methods for MSPDs are inefficient and cannot detect hidden damage.
• Rapid post-earthquake assessment of MSPDs is vital for structural integrity.

Purpose of the Study:

• To develop a novel, accurate, and efficient method for detecting damage in MSPDs.
• To utilize active sensing and machine learning for automated damage assessment.
• To overcome the limitations of traditional visual inspection techniques.

Main Methods:

• Simulated various damage states in MSPDs through quasi-static tests.
• Applied sweep-frequency excitation and used piezoelectric sensors to capture stress wave signals.
• Processed signals using Wavelet Packet Transform (WPT) and energy spectrum analysis for feature extraction.
• Trained and validated a k-nearest neighbor (KNN) algorithm with extracted time-frequency features.

Main Results:

• The KNN model achieved a high validation accuracy of 98.9% with all data and 98.1% with a single channel.
• The method demonstrated robustness, achieving 92.6% accuracy on a structurally varied MSPD.
• The active sensing approach effectively identified different damage conditions in MSPDs.

Conclusions:

• The proposed active sensing and KNN-based method provides a highly accurate and efficient solution for MSPD damage detection.
• The technique shows significant practical potential for real-world post-earthquake inspections.
• The method's robustness suggests applicability to MSPDs with minor structural variations.