Browsing by Author "Saeed Far, Soheil"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Determination of threshold value for estimating significant wave height and depth of similarity
    (Universiti Teknologi Malaysia, 2016) Saeed Far, Soheil
    Probabilistic and deterministic evaluations of surface waves have been undertaken to estimate the significant wave height, and evaluate the kinematic and dynamic properties of wave heights. In the probabilistic description, two extreme wave analysis models, namely Peaks-Over-Threshold (POT) and Generalized Pareto Distribution (GPD), were developed to highlight their merits and limitations, and improve the POT model. Studies have shown that the POT model is not equipped with a suitable approach to determine true threshold value. This study proposed a Coupled method to specify true threshold value for the model. The employed data in the modeling were the wave height data observed during a 41-year period from 1949 to 1989, in Labuan offshore area, Sabah, Malaysia. Presently, the minimum ratio of residual correlation coefficient (MIR) criterion is used as a goodness-of-fit method in the POT model. However, examinations indicated that the MIR method has not always been a reliable approach in determining the best fitting distribution function. This study proposed an alternative approach, known as Norm of Residuals to replace it with MIR. Validity of Norm of Residuals was examined by the Chi- Square test, and the modeling of the two extreme wave analyses was conducted in MATLAB programing. In the deterministic evaluation, the significant wave height obtained by the probabilistic evaluations was used to develop two simulation models. Linear and non-linear wave theories were applied to specify identical wave properties (Depth of Similarity) in conjunction with a fixed jacket platform. However, Depth of Similarity has not yet been determined. This study determined the depth of similarity for Labuan offshore area by simulating the two wave theories. The results of this study demonstrated that the Coupled method for the POT model is reliable and feasible, and the threshold value for the Labuan dataset is u = 2 m. Moreover, the proposed Norm of Residual method for selecting the best fitting distribution is very stable, and matched the POT model. The best fitting distribution function for the Labuan data is Weibull distribution with shape parameter k = 1:4. In addition, the depth of similarity obtained by the simulation models initiates from 10m below the surface toward the seabed with 99% similarity. In conclusion, probabilistic evaluation has shown that the modified POT provides reliable results compared with modern models, and the deterministic evaluation indicated that the depth of similarity could be different from one region to another