The ability to estimate the dynamic response parameters of the offshore structures such as the Tension Leg Platforms (TLP), ‘Spars’, and Semi- Submersibles, is of vital importance to the offshore industry. In this paper, system identification of offshore structures using empirical mode decomposition (EMD) and Hilbert Transform (HT) is presented. The method developed is tested on model test results from a TLP installed in Gulf of Mexico at a depth of around 3000 ft. The whitenoise wave test results are processed using EMD/HT and compared with the results obtained from conventional analysis. Verification of the model test results with data from site for another TLP installed in South Asia is also presented. It is shown that the EMD is highly effective in identifying the higher modes and damping ratios.
With increasing interest in deep water oil exploration, innovative floating platforms like TLP's, Spar and Semi-Submersible have gained in popularity. Since these structures exhibits complex dynamic behavior, traditional methods such as Fourier analysis may not be able to describe the system completely.
When a structure is subjected to an excitation, the energy in the excitation signal near the main modes of the structure causes the maximum response. The basic objective of time-frequency analysis is to devise a function that will describe the energy density of a signal simultaneously in time and frequency domains (Cohen 1995). Recently, the time frequency analysis method of Empirical Mode Decomposition and Hilbert Transform (EMD/HT) (Huang et.al. 1998) has been effectively used to identify the parameters of nonstationary and nonlinear signals.
Yang et.al (2000) used the EMD/HT algorithm along with random decrement technique, on the ambient response of structures to effectively estimate the modal parameters of tall structures.
