Download or read book Investigation of Time Series Analysis Based Damage Detection Methodologies for Structural Health Monitoring written by Qipei Mei and published by . This book was released on 2014 with total page 107 pages. Available in PDF, EPUB and Kindle. Book excerpt: Structural Health Monitoring (SHM) is widely accepted as a valuable tool for monitoring, inspecting and maintaining infrastructure systems. Damage detection is one of the most critical components of SHM to identify the existence, location and severity of damage so that effective preventive actions can be taken to improve the condition of the structure. In this study, time series analysis based damage detection methods using output-only vibration data are investigated for global condition assessment of structures. The main body of the thesis falls into two key parts. In the first part, a novel damage detection method based on Auto-Regressive models with eXogenous (ARX models) and sensor clustering is proposed. Two different Damage Features (DFs) based on the difference of fit ratios and ARX model coefficients are considered in this part. Applying this method to experimental data from a steel grid type structure and a 4-span bridge type structure, it is demonstrated that the existence, location and severity could be successfully assessed by both two DFs for most of the cases. In the second part, an improved version of the previous method is developed to separate the changes in stiffness and mass of the structure using output only data. In order to verify this approach, it is first applied to a 4-DOF mass spring system and then to the shear type IASC-ASCE numerical benchmark problem. It is demonstrated that the approach can not only accurately determine the location and severity of the damage, but also distinguish between changes in stiffness and mass. This study constitutes the first approach of its kind which can distinguish between change of stiffness and mass by using output only vibration data. At the end of the thesis, the limitations of current methods, recommendations, and future work are also addressed.