Investigation On Airgap Response For Semisubmersibles Based On Model Test Results And Statistical Analysis

Semisubmersibles are important facilities to explore and produceoffshore oil. With more severe sea climate for semisubmersibles, theconcept and basic design in defining airgap is vital. Various scholars havebeen delving into this problem utilizing CFD method and potential code,however, the offshore industry is still relying on model tests to finalizeplatform design involving airgap problems.This thesis conducts a comprehensive study on airgap using potential,time-domain and statistical analysis based on model tests’results.The application of potential methods on airgap is firstly discussed. Wehave calculated airgap response RAO for different locations around theplatform and compared calculated response RAO with frequency resultstransformed from white noise test. It’s been found that linear potentialmethod could preliminarily determine airgap response at differentlocations; what’s more, absolute wave elevation at different locations hasstark difference. The cross spectral analysis from incident irregular wavesand airgap response time series is compared with white noise tests. It’sbeen found that spectral analysis from white noise results is generallyreliable. Consequently, for defining absolute wave elevation, we couldreplace irregular wave tests with white noise results.Considering the influence from platform motions and mooring systemon airgap, we have also studied the feasibility of two time domainmethods. The first is concerned about potential time domain bytransforming the traditional airgap problem into double floating bodies,which could be solved by hydrodynamic code, thus directly obtaining therelative wave elevation compared with statistical values stemming fromtime series of irregular model tests. The comparison is good with closestandard deviation. The second method is combining Jonswap spectrumwith response RAO extracted from white noise. The result shows that thesecond method is relatively small compared to time series statistics frommodel tests. Last, we have made a comprehensive construction on statistical modelsfor airgap amplitudes and periods. Inspired by statistical models fromwave spectrums, we have constructed empirically-suited models forairgap amplitudes and periods using parameter estimating methods. Theapplication of Gumbel distribution for obtaining reliable extreme airgapamplitudes is advanced. Also, parameter estimating methods could beused to assess the influence of vertical motions on airgap. Thebi-distribution for joint airgap amplitudes and periods is estimated bynon-parametric KDE method, plotting a contour curve of probabilitylevels and discovering the modes of amplitudes and periods.