A combined method of cross-correlation and PCA-based outlier algorithm for detecting structural damages on a jacket oil platform under random wave excitations

During the last decades, the detection of damages in offshore platforms has been one of the most important issues in ocean engineering. To achieve the goal of the accurate identification of structural damages, the damage sensitive features need to be extracted from the structural dynamic responses. However, it is well known that the changes in the structural dynamic responses do not depend only on the damage but also on the environmental conditions. Many researchers have pointed out that the principal component analysis (PCA) method is effective in discovering the hidden information from the grossed and blurred data. Additionally, the cross-correlation function has been proved effective in building structural characteristics by proceeding it between the dynamic responses from different structural positions. Thus, this paper presents a method in developing a novel PCAbased method for clarifying the structural health state from the structural characteristics built by the cross correlation function. Such method was verified by using the numerical model of a jacket oil platform subjected to random wave excitations. Subsequently, the noise resistance of the proposed method was also investigated. The results show that the proposed method is available in detecting the structural damages under the influence of the random wave excitations.