Friday Oct 19, 2018

Extreme Value Analysis

The design wave's estimation is of great importance for the structural computations. Many of the phenomeno which are analyzed in different engineering and scientific fields, are of random nature. Therefore, it is not possible to comment about occurrence and not occurrence or accurate behavior of them confidently. Nevertheless, by the assistance of statistical and probabilitistic methods, it is possible to extract some rules about the phenomeno’s occurrence probability and some relevant indexes (e.g. the average values, the maximum, etc.) from the past behaviors and regarding the other effective  parameters.
Extreme value analysis is a scientific method based on which we can relate a random phenomena value to the probability or return period of it’s occurrence. Up to now several relations in this field have been presented by the researches. Because of the difference of random phenomena with each other, each of the relations (which are known as probabilistic distribution functions) is valid for special phenomena (e.g. flood, rain, wave, wind, etc.)
In many of the engineering projects, due to the measurements, observations or previous simulations of a phenomenon, one of the recommended distribution functions which has the most compatibility to the present values is selected as the behavior pattern of the considered phenomenon. Usually in distribution functions, there are also parameters that we can increase the goodness of  fit to the data of the distribution function.
Finally, if standard statistical tests show that there exits efficient compatibility between the data and the selected distribution function, we can extract  the considering values form the distribution function. Off course we shouldn’t expect that by using the data corresponding to a limited period we can obtain accurate values for very large return periods.
In order to determine the design wave two different approaches can be used. First of all, the design wave determination can be accomplished by extrapolation of the wave data resulting from measurement or simulation in a long-term period using different distribution functions. The second approach is estimation of the design wind using appropriate probabilistic distributions, and converting it to the design wave in deep water region and transferring it to the shallow water. Therefore, in most of the coastal design projects, due to the lack or inadequacy of the wave measured information in the considered area, using the extreme value analysis on the wind data, the wind speed is calculated for different return periods and then by the assistance of wind to wave conversion relations, wave heights values are obtained for corresponding return periods.
In the present study, Naiband bay and Assalooyeh areas have been selected as the reference site and, the design wave has been determined in the site by different methods and is compared and analyzed with other studies. Thereafter, using the obtained information, the design wave has calculated close to important structures around the second phase studies area with the return period of 100 years.

 انتخاب داده‌های حدی از یک سری زمانی به روش AMS و PDS
Extreme data selection from a time series by the method of AMS and PDS

موقعیت نقاط مورد نظر برای تعیین موج طرح آب عمیق
Considered points locations for determining the deep water design wave

 

 


Last Update : Mar 14, 2012 08:36