Vibration-Based Damage Identification of a Steel Frame Using an Output-Only Algorithm

There are various approaches used by different research groups to identify structures and structural changes, and the success of a certain methodology may depend on the context in which it is applied. Therefore, it is crucial to verify promising methodologies by testing them on different structures and damage cases. The objective of this study is to investigate a statistical pattern recognition-based method of Structural Health Monitoring (SHM) using a laboratory structure. Sophisticated finite element models and traditional modal parameters are not used in the implementation of the statistical pattern recognition techniques, as they require significant user interaction. Instead, the statistical approaches presented in this paper are solely based on the signal analysis of the measured vibration data. This makes this approach attractive for the development of an automated health monitoring system. A large-scale laboratory structure was constructed at the Qatar University structures laboratory, and a large dataset of vibration signals was obtained under several structural damage scenarios. This paper suggests a statistical moments-based technique to identify damage using the vibration signals. The method does not require labor-intensive supervised learning, and only acceleration sensor data is required to detect damage. Overall, the proposed approach has the potential to be a cost-effective and efficient solution for SHM of various infrastructures.