Improvement study of modal analysis for offshore structures based on reconstructed displacements

Compared with traditional acceleration-based modal analysis, displacement can serve as a more reliable and robust identification parameter for operational modal analysis (OMA). However, for offshore structures operating in the marine environment, it is difficult to find a fixed reference point for measuring structural displacements. Additionally, when displacements are reconstructed by integrating measured accelerations, unrealistic drifts are inevitably generated due to the unknown initial velocity and displacement as well as the influence of noise. Aiming at providing more accurate and stable results for identifying the modal parameters of offshore structures, this paper developed a displacement-based OMA strategy based on displacement reconstruction and stochastic subspace identification (SSI) techniques. The strategy starts from integration of the measured accelerations, and removes the drift terms by employing Chebyshev polynomials to fit the integrated displacements. Then, the reconstructed displacement-based SSI is applied to offshore structures for modal parameter identification. To demonstrate the differences between results obtained by using the reconstructed displacements and accelerations for OMA, two numerical examples are used, i.e., a synthesized ideal acceleration signal and a two-degree-of-freedom (2-DOF) system. Numerical results show that the low-frequency components in the reconstructed displacements can be more easily identified than those in the accelerations, which implies that the developed strategy may be more suited to offshore structures with low-frequency characteristics. To further investigate the performance of the developed displacement-based strategy, experimental data of a cantilever beam with similar structural properties to a monopile offshore wind turbine (OWT) is then analysed. The results show that the displacement-based strategy can produce more accurate modal parameters than using accelerations. Finally, two sets of accelerations from field tests of an OWT at difference installation stages are used: one set of data is measured from a foundation before a turbine is installed, and the other is collected from a completed OWT. Reference results are calculated by performing the eigensystem realization algorithm (ERA) on measured accelerations under ship collision excitations and the OMA results show that for low-frequency dominated structures like OWTs, the developed displacement-based OMA strategy is more stable and accurate than traditional acceleration-based modal analysis.