IMPROVEMENT ON ESTIMATION ACCURACY OF DIRECTIONAL SPECTRUM BY A BAYESIAN METHOD FOR SWELL OBSERVATION

Abstract

 Although swell can be dominant for damage of coastal structures, there is not enough research on directional wave spectrum of swell for design use. We improved Bayesian Directional Method (BDM) to calculate more precise estimator for directional spectrum using Non Negative Least Square method (NNLS) for optimization. We examined the estimation accuracy of directional wave spectrum of swell using the data obtained submerged Doppler-type directional wave meter (SDDWM) for the extended maximum liklihood method (EMLM), Bayesian Directional Method (BDM), Bayesian Directional Method with NNLS (BDM-NNLS). Provided directional spectrum from Bretschneider-MITSUYASU spectrum and MITSUYASU directional function, we simulated the data obtained from SDDWM. As a result, BDM-NNLS showed the best agreement for the estimation accuracy of directional wave spectrum of swell.