An Experimental Study on Damage Detection of a Building Structure Model

Abstract

The collapse of a large infrastructure such as bridges, dams and harbors cause serious losses. Therefore, there has been an increasing demand in using Structural Health Monitoring techniques for the building structures so that their maintenance cost and time can be reduced. In this study, a structural health monitoring methodology using dynamic responses is proposed for damage detection of a prototype building structure during shaking table testing in different positions of structural damage. Damage characters are calculated from the measured acceleration data and applied to outlier analysis, one of unsupervised learning based pattern recognition technique for a damage index and a decision making. And a threshold value for the outlier analysis is determined based on confidence level of the probabilistic distribution of the acceleration data. For more systematic damage detection, several control parameters to determine and optimal decision boundary for the unsupervised learning based pattern recognition are optimized. Finally, further research issues will be discussed for realworld implementation of the proposed approach.