Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Hybrid physics-data driven spectral forecasts of semisubmersible response

Ian Milne, Lachlan Astfalck, Matthew Zed, Jack Lee-Kopij, Edward Cripps

Abstract

A framework for probabilistic forecasting of vessel motion is developed and validated for a semisubmersible operating in long period swell. Bayesian statistical methods are applied to predictions of the heave response from a physics model using numerical wave spectra and measured motion data. Model diagnoses motivate an additional level of complexity required for the error structure in the Bayesian model, specifically to account for heteroskedasticity and time-correlated errors. The hybrid model forecasts were evaluated during periods where the heave resonance and cancellation frequencies were excited. The method is demonstrated to be effective for providing reliable quantification of uncertainty and correcting bias in the raw physics model predictions. This justifies its value for improving the efficiency and safety of offshore operations.

Hybrid physics-data driven spectral forecasts of semisubmersible response

Abstract

A framework for probabilistic forecasting of vessel motion is developed and validated for a semisubmersible operating in long period swell. Bayesian statistical methods are applied to predictions of the heave response from a physics model using numerical wave spectra and measured motion data. Model diagnoses motivate an additional level of complexity required for the error structure in the Bayesian model, specifically to account for heteroskedasticity and time-correlated errors. The hybrid model forecasts were evaluated during periods where the heave resonance and cancellation frequencies were excited. The method is demonstrated to be effective for providing reliable quantification of uncertainty and correcting bias in the raw physics model predictions. This justifies its value for improving the efficiency and safety of offshore operations.

Paper Structure

This paper contains 18 sections, 10 equations, 10 figures, 1 table.

Table of Contents

  1. Introduction
  2. Physics-based approach
  3. Semisubmersible hydrodynamics
  4. Surface wave forecasts
  5. Vessel response forecasts
  6. Efficacy of a pure physics model forecast
  7. Bayesian forecasting, probabilistic scoring metrics and the hybrid model
  8. Bayesian statistics and probabilistic forecasts
  9. Proper scoring rules and validation metrics
  10. The hybrid model
  11. Data
  12. Measurements of vessel heave
  13. Compilation of predictions for forecast horizons and training-test data split
  14. Diagnostics and results
  15. Diagnostics
  16. ...and 3 more sections

Figures (10)

  • Figure 1: Example of the heave RAO a semisubmersible MODU with and without vortex drag, with wave energy from ocean swell coinciding with the heave resonance and cancellation frequencies.
  • Figure 2: From top: time histories of the significant wave height (Hs), peak spectral wave period (Tp) and significant heave displacement ($2\sqrt{m_0}$). Comparisons are shown between the observed and predicted values (from the numerical forecast or, for the heave motion only, the wave buoy). Results are at an 1h time step over a 1 month period.
  • Figure 3: Graphical interpretation of the CRPS. The distance in the integral in Equation (9) is represented by the shaded area. Forecast $F_t(z)$ is a cummulative distribution function -valued quantity.
  • Figure 4: Autocorrelation plots (upper panels) and absolute residuals versus numerical model heave (lower panels) for the linear adjustment model of equation \ref{['eqn:linear_adjust']}. From left to right: 0, 12 and 96 hour horizons.
  • Figure 5: Autocorrelation plots (upper panels) and absolute residuals versus numerical model heave (lower panels) for the linear adjustment model of equation \ref{['eqn:linear_adjust_errors']}. From left to right: 0, 12 and 96 hour horizons.
  • ...and 5 more figures