Parametric modelling of nonstationary platform deck motions
Abstract
Univariate system identification models are, in general, developed based on teh assumption that offshore dynamic systems are stationary random processes. For nonstationary processes, a method is adopted to transform the time series into wide-sense-stationary processes. Then the time series are modelled by the maximum entropy method which is formulated here for spectral estimation of platform deck displacements. The lower order maximum entropy spectra of nonstationary platform deck displacements are compared with fast Fourier transform spectra and show good agreement. However, the higher order maximum entropy model can be used for better representation of nonstationary motions. This method also reduces long time series of nonstationary offshore data into a few values which are teh coefficients of the maximum entropy model.
Collections
- Scholarly literature [8843]