APPLICATION OF MEP METHOD TO THE STUDY OF WAVE CLIMATE STATISTICAL CHARACTERISTICS

摘要

The Maximum Entropy Principle (MEP) method is elaborated, and the correspondingrnprobability density evaluation method for the random fluctuation system is introduced,rnthe goal of the article is to find the best fitting method for the wave climate statisticalrndistribution. For the first time, a kind of new maximum entropy probability distributionrn(MEP distribution) expression is deduced in accordance with the second order moment ofrna random process. Deferent with all the fitting methods in the past, the MEP distributionrncan describe the probability distribution of any random fluctuation system convenientlyrnand reasonably. If the moments of the random signal is limited to the second order, that is,rnthe ratio of the root-mean-square value to the mean value of the random variable isrnobtained from the random sample, the corresponding MEP distribution can be computedrnaccording to the deduced expression in this essay. The concept of the wave climate isrnintroduced here, and the MEP distribution is applied to fit the probability densityrndistributions of the significant wave height and spectral peak period. Take the MexicornGulf as an example, three stations at different locations, depths and wind wave strengthsrnare chosen in the half-closed gulf, the significant wave height and spectral peak periodrndistributions at each station are fitted with the MEP distribution, the Weibull distributionrnand the Log-normal distribution respectively, the fitted results are compared with the fieldrnobservations, the results show that the MEP distribution is the best fitting method, and thernWeibull distribution is the worst one when applied to the significant wave height andrnspectral peak period distributions at different locations, water depths and wind wavernstrengths in the Gulf. The conclusion shows the feasibility and reasonability of fittingrnwave climate statistical distributions with the deduced MEP distributions in this essay,rnand furthermore proves the great potential of MEP method to the study of wave statisticalrnproperties.