Laboratory tests on local damage detection for jacket-type offshore structures using optical FBG sensors based on statistical approaches

In this study, a local damage detection based on statistical approach for jacket-type offshore structures by principal component analysis (PCA) and linear adaptive filter (LAF) techniques using strain response data measured by FBG sensors was proposed while dynamic responses are being popularly utilized for damage detection of civil infrastructures including jacket-type offshore structures. In addition, environmental effects due to variations in temperature and external loading were intensively investigated and an efficient remedy was proposed using the nonparametric PCA and LAF models. Unlike many existing statistical damage detection methods, the mean of residual values eliminating the environmental effects was utilized as damage index for rational for enhancing the normality based on the central limit theorem and the normality was first checked before damage estimation using the mean of residual values. Laboratory tests for a scaled tidal current power plant structure, one of many jacket-type offshore structures, were conducted to investigate the technical feasibility of the proposed method for damage detection and localization. It was found that the PCA technique could more efficiently eliminate undesired environmental effects contained in the measurement data from FBG sensors without any additional information on the environmental changes, resulting in more damage-sensitive features under various environmental changes.