Novel methods for coupled prediction of extreme wind speeds and wave heights

Abstract

Two novel methods are being outlined that, when combined, can be used for spatiotemporal analysis of wind speeds and wave heights, thus contributing to global climate studies. First, the authors provide a unique reliability approach that is especially suited for multi-dimensional structural and environmental dynamic system responses that have been numerically simulated or observed over a substantial time range, yielding representative ergodic time series. Next, this work introduces a novel deconvolution extrapolation technique applicable to a wide range of environmental and engineering applications. Classical reliability approaches cannot cope with dynamic systems with high dimensionality and responses with complicated cross-correlation. The combined study of wind speed and wave height is notoriously difficult, since they comprise a very complex, multi-dimensional, non-linear environmental system. Additionally, global warming is a significant element influencing ocean waves throughout the years. Furthermore, the environmental system reliability method is crucial for structures working in any particular region of interest and facing actual and often harsh weather conditions. This research demonstrates the effectiveness of our approach by applying it to the concurrent prediction of wind speeds and wave heights from NOAA buoys in the North Pacific. This study aims to evaluate the state-of-the-art approach that extracts essential information about the extreme responses from observed time histories.

Offloading operation bivariate extreme response statistics for FPSO vessel

“…In this study, authors employed a general Monte Carlo (MC) based technique, able to tackle inherent nonlinear effects without significant simplifications except those inherent in the numerical model itself. The target bivariate extreme response distribution was obtained by combining MC simulations with the ACER2D extrapolation method [12][13][14][15][16] . Unlike univariate statistical techniques, the ACER2D method accounts for bivariate response correlation.…”

Improving extreme offshore wind speed prediction by using deconvolution

“…The countless strategies endeavoured included the method presented by Mohande [ 26 ], which predicts wind speeds using neural networks or the different approaches to analyse wind speed estimates by Yan et al, as in Refs. [ [1] , [2] , [3] , [4] , [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] , [19] , [20] , [21] , [22] , [23] , [24] , [25] , [26] ]. Similarly, the authors of this paper have also previously used different statistical approaches to estimate better extreme statistical values such as, i.e.…”

“…Similarly, the authors of this paper have also previously used different statistical approaches to estimate better extreme statistical values such as, i.e. wind speed, wave height, response and load [ [1] , [2] , [3] , [4] , [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] , [19] , [20] , [21] , [22] , [23] , [24] , [25] , [26] ]. Furthermore, the authors developed a novel reliability approach to forecast COVID-19 epidemic levels, indicating the approach's versatility and reliability in various fields [ 14 , 16 , 26 ].…”

“…Similarly, other authors, as in Refs. [ [1] , [2] , [3] , [4] , [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] , [19] , [20] , [21] , [22] , [23] , [24] , [25] , [26] ], have attempted to use novel statistical approaches to estimate wind speed or pattern more accurately. Meanwhile, two recent and pertinent scientific articles [ 8 , 26 ] illustrated the importance of wind forecasting and strategies adopted to maximise efficiency.…”

“…The convolution technique also does not assume any specific extrapolation functional class. This is in contrast to many traditional extrapolation methods widely used in engineering practice where an assumed probability function is used [ [1] , [2] , [3] , [4] , [5] , [6] , 30 ]. Some other examples also include the Pareto-based distribution peak over the threshold (POT) [ 1 ], The Naess-Gaidai (NG) method fitting procedure was used in averaged conditional exceedance rate (ACER) method [ [7] , [8] , [9] , [10] , [11] ], bivariate ACER fitting method [ [12] , [13] , [14] , [15] , [16] , 47 ], Weibull distribution-based fitting method [ 38 ] and Gumbel distribution based fitting method [ 17 ].…”