Quantifying and discritizing the uncertainity in the power production estimates of a wave energy converter

INTRODUCTION The ability to predictively assess Wave Energy Converter (WEC) power production is vital to the decision making process of WEC developers, electricity utilities and project financiers. However, when simulating a specific scenario there are numerous variables that introduce uncertainty in the prediction, even within a single seastate significant wave height (Hs) and energy period, (Te) bin. This work seeks to determine the causes of variability in mean power predictions and quantify the magnitude of these causes. The precise shape of the wave spectra and the exact values of Hs and Te are often not known, the random phases of the spectral constituents can never be predetermined, hydrodynamic parameters (i.e. viscous drag coefficients) may be in error and other environmental factors such as winds, currents and tidal heights are often omitted. This work quantifies and discretizes the probability distributions (PD) of these variables and conducts a Monte Carlo experiment with over 12 000 high-fidelity simulation runs. All simulation runs and environmental factors are from within a single seastate bin. Each run models the dynamics of a 7 degree of freedom axisymmetric two body point absorber WEC. The PD of the 20-min mean WEC power is presented. The mean power is systematically detrended and the deterministic influence of each random variable is quantified through trend-based data fitting. The detrended mean power data reveals a baseline level of uncertainty in the mean power recovered; even with known model coefficients and environmental conditions. This baseline uncertainty is associated with the randomness in the time series profile of the wave field caused by the random nature of the wave phases. This Monte Carlo simulation has only been conducted on one WEC design, one seastate bin and for a single location. Therefore its relevance to other WEC designs, seastates and locations cannot currently be ascertained. This work presents a robust procedure to determine uncertainty for a single seastate bin and demonstrates the importance of establishing this uncertainty.