This research investigation focused upon the development of a new approach to interpret complex hydrodynamic phenomena using data from experimental studies or field measurements. A promising system identification technique was further developed to address nonlinear and frequency dependent aspects common to marine riser and ocean platform dynamics. The sequential development of analytical formulations and the use of computer simulations and model basin experiments are documented herein and in archival refereed publications.
Initially, a distributed parameter formulation incorporating the reverse system identification technique for multiple input/ single output nonlinear problems was investigated. This combined time/frequency domain method was used to illustrate the propagation of various types of hydrodynamic nonlinearities along the length of a marine riser. The marine riser response model introduced the combined method of normal modes and system identification procedures. Numerical simulations using this new approach demonstrated that the parameters of interest were convergent for each of the modes that were included. Further, the sensitivity of this methodology to predict the selected parameters over a range of frequencies and the degree of variation that could be expected under ideal simulation conditions, and the propagation of nonlinearities along a riser were illustrated.
The further development of this time/frequency domain system identification technique addressed the precise evaluation of the frequency dependence of parameters such as added mass, stiffness and damping, as well as the use of fully correlated signals in the process of parameter estimation. Although some multiple riser data was available, it was not suitable for this aspect of the study and a recent series of rigid and compliant mini-TLP model tests was used. Practical issues regarding the application of this approach, the utilization of force and moment measurements, the over specification of nonlinearities in a predictive model and the ordering of non-linear contributions by their importance were addressed. In general the results demonstrate that the methodology is quite robust and yields predictions that are most accurate for the parameters associated with the largest motions of the ocean system being investigated. At this point the methodology is sufficiently developed to where it can be applied to investigate a wide range of marine riser, ocean structures and floating platform systems.