| Summary: | 博士 === 國立成功大學 === 水利及海洋工程學系碩博士班 === 96 === A sea surface wave field is a useful way to present wave features in both time and
space domains. Several algorithms have been proposed for analyzing the wave field so as to
obtain significant wave parameters. Due to the non-homogeneity of the wave field in space
domain, the algorithms which assume the wave field non-homogeneous may not be suitable
for discussing the spectral characters of the non-homogeneous wave field.
The wavelet transform is now recognized as a useful, flexible and efficient technique to
analyze intermittent, non-stationary and non-homogeneous signals as well as remote sensing
images. It should be workable to discuss a non-homogeneous wave field using wavelet
transform although, previously, this issue has received little attention. The aim of this
research was to study wave field analysis by applying spatio-temporal wavelet transform, in
order to understand the features of non-homogeneity when discussing wave propagation on
a varying topography.
For the sake of applying wavelet theory to extract wave information from the
non-homogeneous wave field, the algorithm of spatio-temporal wavelet transform was
developed. The relationship between wave parameters and the wavelet analyzed results
were also derived from this study.
The numerical algorithm of spatio-temporal wavelet transform was tested in a
simulated wave field of regular and irregular waves in order to understand the practicability
and accuracy of wave calculation from a wave field by wavelet transform. This is the first
time wavelet transform has been used to analyze a non-homogeneous wave field, the
influences of the wave image features and the wavelet function upon wave field analysis.
In order to understand the non-homogeneity of sea areas, the non-homogeneity of
different wave field cases were quantized using the wavelet transform. It was revealed that
the higher the sea bed slopes and the wave periods, the higher the non-homogeneity indexes.
The influence of the wave period on the non-homogeneity indexes was less than the
influence of bathymetry. In addition, the non-homogeneity index was also influenced by the
wave nonlinearity which is obvious in the shallow water area.
Finally, wave field images, acquired from X-band radar, were applied to the analysis.
For the areas near shore, it was revealed that the observed spectral energy deviated from the
linear dispersion relation. The primary spectral energy distribution agreed with the
relationship between the harmonic frequency and wave number which presents the
nonlinear dispersion relation It could, therefore, be concluded that the use of
spatio-temporal wavelet transform in analyzing non-homogeneous wave fields of random
waves was feasible, even in coastal areas.
|
|---|
