Extracting Stochastic Airwake Models from a Database for Engineering Analysis and Simulation

Abstract

A framework is presented for extracting interpretive models of airwake autocorrelation and autospectrum as well as crosscorrelation and cross-spectrum from a database. These models have a simple analytical structure that aids routine simulation and application as a predictive tool. Airwake refers to turbulence shed from the ship superstructure, and the database, to a set of spectral (autospectral and cross-spectral) points of flow velocity data from experimental and computational fluid dynamics–based investigations. The framework is developed from first principles: It is based on perturbation theory; it addresses all three velocity components, and it is tested against a comprehensive database under different superstructure and wind-over-deck conditions. For each velocity component, the autocorrelation and cross-correlation are represented by separate perturbation series in which the first terms have a form of the von Karman longitudinal or lateral correlation function. These series are then transformed into equivalent perturbation series of autospectra and cross-spectra. The perturbation coefficients are evaluated by satisfying the algorithmic constraints and fitting a curve on a set of selected spectral data points in the low-frequency bandwidth (0≤f(Hz)≤1.6); the emphasis is on extracting spectral models for this bandwidth. Generally, no more than a second-order perturbation correction (a three-term perturbation series) is necessary, and the extracted models lend themselves well to construction of shaping filters driven by white noise. The framework's strengths and weaknesses are discussed as well.