Vibration-Based Loosening Detection of a Multi-Bolt Structure Using Machine Learning Algorithms

Area of Science:

• Engineering
• Artificial Intelligence
• Machine Learning

Background:

• Machine learning (ML) is increasingly utilized in structural health monitoring (SHM).
• Early detection of bolt loosening is critical for structural integrity.

Purpose of the Study:

• To propose a vibration-based technique for early-stage bolt loosening detection and identification using ML algorithms.
• To evaluate the performance of different ML classifiers for this task.

Main Methods:

• Established experimental cases with fastened and loosened bolts on a motor.
• Collected vibration data under three different operating conditions (800, 1000, 1200 rpm).
• Applied Short-Time Fourier Transform (STFT) for feature extraction and trained various ML classifiers.

Main Results:

• The developed ML technique demonstrated satisfactory accuracy in detecting bolt loosening.
• The system could identify specific bolts that began to lose preload across different working conditions.
• The most accurate classifier was identified for practical implementation.

Conclusions:

• The proposed vibration-based ML approach is effective for early bolt loosening detection and identification.
• The findings support the implementation of this technique for online monitoring of multi-bolt structures.