Research On Improving The Accuracy Of Modal Identification For Offshore Platforms

It is valuable in theoretical research and engineering application to develop global damage detection technology for offshore platforms. The global damage detection technique is based on modal parameters identification. Due to the difficulty in implementing free vibration or forced vibration tests for such a huge structure as an offshore platform, it is recognized that employing the ambient vibration measurement to detect structural damage should be the only practical method. The key point of global damage detection technique is to solve the problem of the disturbance ,so as to improve the accuracy of modal identification for offshore platforms. The research of modal parameters identification is still on stage of theory and modal experimentation, and the critical issues is to reduce the influence of environment load,environmental factors and measurement noise in order to improve the accuracy of modal identification for in-situ offshore platforms. In this dissertation, the major contents aiming at the two impacts are shown as following:Firstly, the disturbance due to the colored noise of wave resultant force is almost avoided in this research. The response of offshore platforms is decided by wave, and the colored noise of wave resultant force can disturb modal identification which may appears around the frequency of low order mode of offshore platforms. In order to improve the accuracy of modal identification for offshore platforms under irregular waves, a method to find the peak and trough frequencies of wave resultant force aiming at 4-leg offshore platforms is proposed by analyzing the mechanism that the distance between legs along the wave propagation direction affects the frequency property of wave resultant force. Furthermore, on the basis of it, a method to choose the appropriate range of wave striking angles for avoiding the disturbance of colored noise induced by wave resultant force on modal identification is proposed corresponding to each lower-frequency mode. Appropriate range of wave striking angles is found out corresponding to three lower-frequency modes respectively. Secondly, the disturbance of measurement noise on the accuracy of modalidentification is investigated by numerical simulation. Modal identification is carried out employing dynamic response signals with measurement noise under white noise earthquake loads. Also modal identification is carried out employing dynamic response signals with measurement noise under irregular wave loads. The two impacts are studied respectively. And according to the comparison, the conclusion is that the ARMA method is robust against measurement noise and the impact of measurement noise on modal identification can be reduced.The modal identification is disturbed mainly by the colored noise of wave resultant force, and the accuracy of modal identification can be improved by choosing the appropriate wave propagation directions. The influence due to measurement noise can avoided by mean treatment for identified modal parameters and using the ARMA identification method.