Research

Abstract: 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 is 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.

Keywords: Structural health monitoring, statistical pattern recognition, vibration signals, laboratory structures, damage identification, automated system, statistical moment analysis

Abstract: Structural health monitoring (SHM) is crucial to detect damage in structures at an early stage, allowing timely maintenance and repair to ensure their safety and longevity. This paper presents a study that investigates the feasibility of using a statistical pattern recognition-based method for SHM using a laboratory structure. The proposed approach relies solely on signal analysis of the measured vibration data, making it cost-effective and attractive for the development of an automated health monitoring system. Unlike traditional SHM methods, the proposed approach does not require labor-intensive tuning, expert knowledge, or extensive training, reducing the time and cost required for SHM. The large-scale laboratory structure at Qatar University provides a unique platform to obtain a large dataset of vibration signals under several structural damage scenarios. The study presents a technique to identify damage using Mahalanobis distance between vibration signals of damaged and undamaged conditions. The proposed approach has the potential to be a practical and efficient solution for SHM in civil, mechanical, and aerospace engineering applications, contributing to the development of reliable and accurate health monitoring systems for structures.

Keywords: Structural health monitoring; Statistical pattern recognition; Vibration analysis; Mahalanobis

distance; Automated monitoring systems.