396
CHAPTER 7. LABORATORY WAVE GENERATION
(1987). They pointed out that (in most cases) we do not know what sequence of wave heights and periods will be the worst condition for each
specific project being tested in a physical model. For example, is a high
wave crest followed by a low wave trough more damaging to a rubble-mound
structure than a low trough followed by high crest? Therefore, rather than
attempting to select the “worst condition” a priori, lengthy test runs will
increase the chances that the relevant worst condition for each model test
will occur in the wave tank. Of course, there remains the finite probability
that the test run will end before the worst condition has occurred.
Several synthesis methods fitting the nondeterministic definition for irregular waves have been reported in the literature. These methods use a
random ; meaning the signal can never be repeated) or a pseudo-random
noise source that has a long cycle time. The only constraints placed on
the simu'ation are the wave energy spectrum as represented by the spectral
shape and the total energy contained in the spectrum zeroth moment).
One commonly used method for synthesizing irregular waves according
to the nondeterministic definition is referred to as the Nondeterministic Spectral Amplitude NSA model *
Tuah and Hudspeth 1982, Hudspeth. Nath, and Sollitt 1985) . The NSA model assumes the sea surface
ûucxaatâons at a point can be represented by Gaussian white noise given as
*Fiaüx-. Maasaid aod Dai (19SS) referred to üws method *s the Axsdcci Complex
ACS aaethed
x/2
^(f) = y +
4-it sm-U *
(7.181)
■Imjc the discrète radian frequencies,
= 2w/a. are densely spaced along
the frequency interval 0 < f < oo. The first term is re-rog^Ted as the mean,
u: tne coefficients
and Eu are independent
inm variables that are
u-urmam» distributed with zero mean and with a standard deviation given
fey
\
)rA/
(7.182)
wrere •_
are discrete values of the ron:one-sided power spectral
dessky function.
Details for numerical simulation, of an irregular tire series realization
œaforming to a specific spectral shape are riven in Tuas and Hudspeth
v 1^-82 and in Hudspeth Nath, and Solhtt (1985 Basically the procedure
fotkws these steps:
1 Generate a sequence of paired random numbers. a and
-X- that are normally distributed with zero mean and unit
*
CHAPTER 7. LABORATORY WAVE GENERATION
(1987). They pointed out that (in most cases) we do not know what sequence of wave heights and periods will be the worst condition for each
specific project being tested in a physical model. For example, is a high
wave crest followed by a low wave trough more damaging to a rubble-mound
structure than a low trough followed by high crest? Therefore, rather than
attempting to select the “worst condition” a priori, lengthy test runs will
increase the chances that the relevant worst condition for each model test
will occur in the wave tank. Of course, there remains the finite probability
that the test run will end before the worst condition has occurred.
Several synthesis methods fitting the nondeterministic definition for irregular waves have been reported in the literature. These methods use a
random ; meaning the signal can never be repeated) or a pseudo-random
noise source that has a long cycle time. The only constraints placed on
the simu'ation are the wave energy spectrum as represented by the spectral
shape and the total energy contained in the spectrum zeroth moment).
One commonly used method for synthesizing irregular waves according
to the nondeterministic definition is referred to as the Nondeterministic Spectral Amplitude NSA model *
Tuah and Hudspeth 1982, Hudspeth. Nath, and Sollitt 1985) . The NSA model assumes the sea surface
ûucxaatâons at a point can be represented by Gaussian white noise given as
*Fiaüx-. Maasaid aod Dai (19SS) referred to üws method *s the Axsdcci Complex
ACS aaethed
x/2
^(f) = y +
4-it sm-U *
(7.181)
■Imjc the discrète radian frequencies,
= 2w/a. are densely spaced along
the frequency interval 0 < f < oo. The first term is re-rog^Ted as the mean,
u: tne coefficients
and Eu are independent
inm variables that are
u-urmam» distributed with zero mean and with a standard deviation given
fey
\
)rA/
(7.182)
wrere •_
are discrete values of the ron:one-sided power spectral
dessky function.
Details for numerical simulation, of an irregular tire series realization
œaforming to a specific spectral shape are riven in Tuas and Hudspeth
v 1^-82 and in Hudspeth Nath, and Solhtt (1985 Basically the procedure
fotkws these steps:
1 Generate a sequence of paired random numbers. a and
-X- that are normally distributed with zero mean and unit
*