W SW Analysis Tools
W SW Analysis Tools
W SW Analysis Tools
User Guide
Limited Liability
The liability of DHI is limited as specified in Section III of your 'DHI
Software Licence Agreement':
'IN NO EVENT SHALL DHI OR ITS REPRESENTATIVES (AGENTS
AND SUPPLIERS) BE LIABLE FOR ANY DAMAGES WHATSO-
EVER INCLUDING, WITHOUT LIMITATION, SPECIAL, INDIRECT,
INCIDENTAL OR CONSEQUENTIAL DAMAGES OR DAMAGES
FOR LOSS OF BUSINESS PROFITS OR SAVINGS, BUSINESS
INTERRUPTION, LOSS OF BUSINESS INFORMATION OR OTHER
PECUNIARY LOSS ARISING OUT OF THE USE OF OR THE INA-
BILITY TO USE THIS DHI SOFTWARE PRODUCT, EVEN IF DHI
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
THIS LIMITATION SHALL APPLY TO CLAIMS OF PERSONAL
INJURY TO THE EXTENT PERMITTED BY LAW. SOME COUN-
TRIES OR STATES DO NOT ALLOW THE EXCLUSION OR LIMITA-
TION OF LIABILITY FOR CONSEQUENTIAL, SPECIAL, INDIRECT,
INCIDENTAL DAMAGES AND, ACCORDINGLY, SOME PORTIONS
OF THESE LIMITATIONS MAY NOT APPLY TO YOU. BY YOUR
OPENING OF THIS SEALED PACKAGE OR INSTALLING OR
USING THE SOFTWARE, YOU HAVE ACCEPTED THAT THE
ABOVE LIMITATIONS OR THE MAXIMUM LEGALLY APPLICA-
BLE SUBSET OF THESE LIMITATIONS APPLY TO YOUR PUR-
CHASE OF THIS SOFTWARE.'
Printing History
June 2004
August 2005
November 2006
October 2007
January 2009
3
4 WS Wave Analysis Tools
CONTENTS
5
1 ABOUT THIS GUIDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.1 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2 Assumed User Background . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1 Short Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.1 WS Linear Spectral Analysis module . . . . . . . . . . . . . . . . 12
2.1.2 WS Digital Filtering Analysis module . . . . . . . . . . . . . . . . 12
2.1.3 WS Directional Wave Analysis module . . . . . . . . . . . . . . . 12
2.1.4 WS Crossing Analysis module . . . . . . . . . . . . . . . . . . . 13
3 WS LINEAR SPECTRAL ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . 15
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2 Input File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3 Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3.1 Remarks & hints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.4 FFT Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.4.1 Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.4.2 Data window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4.3 Cut-off frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4.4 Remarks & hints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.5 Basic Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.5.1 Parameter definition . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.5.2 Remarks & hints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.6 Analysis Selections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.6.1 Remarks & hints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.7 Output File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.8 Scientific Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4 WS DIGITAL FILTERING ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . 29
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2 Input File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.3 Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.3.1 Remarks & hints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.4 Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.4.1 Remarks & hints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.5 Transfer Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.6 Basic Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.6.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.6.2 Remarks & hints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.7 Analysis Selections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
7
8 WS Wave Analysis Tools
Purpose
1.1 Purpose
The main purpose of this User Guide is to enable you to use the modules
included in the WS Wave Analysis Tools package for analysis of data
originating from physical model tests, field measurements or numerical
simulations.
It is assumed that you are familiar with the basic elements of MIKE Zero
file types and editors, the Plot Composer and the MIKE Zero Toolbox.
9
About This Guide
2 INTRODUCTION
11
Introduction
Figure 2.2 The WS Wave Analysis Tools can be used for analysis of time
series data from physical model tests, numerical model simulations
or field measurements
The module uses a class of filters called Finite Impulse Response (FIR)
filters also known as non-recursive filters. These filters are characterised
by filter coefficients (in the time domain), which are derived from speci-
fied transfer functions in the frequency domain. The actual filter operation
takes place in time domain by convoluting the input time series and the fil-
ter coefficients.
The module supports time series data of dfs0 and dfs2 type. This allows
you to perform digital filtering analysis of grid time series from numerical
model results, e.g. MIKE 21 BW Boussinesq Waves.
The module supports time series data of dfs0 and dfs2 type. This allows
you to perform directional wave analysis of grid time series numerical
model results, e.g. MIKE 21 BW Boussinesq Waves.
13
Introduction
3.1 Introduction
With the WS Linear Spectral Analysis module it is possible to perform lin-
ear frequency domain analyses based on the Fast Fourier Transform (FFT)
of time series data. Hence, it is presumed that the time series is a superpo-
sition of an infinite number of infinitesimal amplitude waves with uni-
formly and randomly distributed phases (Goda, 1985).
1 Auto spectrum
2 Cross spectrum (amplitude and phase)
3 Cross spectrum (real and imaginary parts)
4 Frequency response spectrum (gain and phase)
5 Coherent power spectrum
6 Coherence spectrum
The auto and cross-spectra, which correspond to the second moment of the
time series, are used for generation of the remaining spectra. Spectra 2
through 6 are dual analyses and require two sets of time series data. For
spectra 4 and 6 it is furthermore necessary to distinguish between input
and output time series as the order makes a difference.
15
WS Linear Spectral Analysis
The WS Linear Spectral Analysis module supports time series data with
equidistant time step. At present the following two binary file formats are
supported:
Whenever the editor recognises the used file format an info box contain-
ing the files header information appears in the lower half of the dialog.
You can view the selected data file using the MIKE Zero Time Series edi-
tor by clicking on the “View” button.
17
WS Linear Spectral Analysis
3.3 Time
The parameters given in the FFT Parameters dialog determines the fre-
quency resolution and the degree of smoothing of the spectrum.
3.4.1 Resolution
FFT size:
Number of data points in each FFT block or subseries. The length must be
a power of two and cannot exceed the total number of data points.
Number of frequencies:
Is half of the FFT size
FFT duration:
Gives the length of the FFT analysis in seconds
Frequency interval:
The difference between two discrete adjacent frequencies. If manually
changed, the value is automatically adjusted to the nearest possible value
(function of time step and FFT size)
19
WS Linear Spectral Analysis
Type
Three different data windows are available:
Overlap
The spectra (and response functions) are calculated using overlapping
subseries. The overlap controls the smoothing of spectra, but it also con-
trols the weighting of the different parts of the input time series.
Number of subseries
Based on the specified FFT parameters the number of subseries can be
calculated. It is almost always desirable to have a large number of sub-
series because it will reduce the variance of the spectral density and pro-
vide a smooth spectral estimate.
For non-overlapping time series the number of degrees of freedom for the
spectral estimate is two times the number of subseries. Spectral estimates
having 30-40 degrees of freedom is often used in practice.
Utilisation
Indicates the utilisation of the input data.
Minimum frequency
The minimum cut-off frequency must be greater than or equal to zero.
Maximum frequency
The maximum cut-off frequency must be less than or equal to the Nyquist
frequency, which is determined entirely by the sampling frequency f s of
the input data:
1
f max = 0.5 f s = -------- (3.1)
2 ∆t
Here ∆t is the time step of the input data. If the value is out of range a
warning is issued and the invalid maximum cut-off frequency is automati-
cally adjusted to the Nyquist frequency.
It is important to realise that the various dialogs are all synonyms for the
FFT size, and that changing one parameter also will affect the others.
21
WS Linear Spectral Analysis
In the Basic Parameters dialog you specify the name of the various
parameter-setting definitions.
For each specified Definition you select the spectra and the output to be
calculated.
Available Spectra
Auto spectrum
The auto spectrum gives the spectral density of the variance of the time
series as a function of the frequency. The auto spectrum is also referred to
as the frequency spectrum, power spectrum or variance spectrum.
Cross-spectrum
The cross-spectrum gives the correlation between the input and output
time series, i.e. for frequencies with spectral values of zero the input and
output is completely uncorrelated. For positive phase angles the output
time series leads to the input time series and vice versa. Output is available
as amplitude and phase values or real and imaginary parts of the generally
complex spectra.
Coherence spectrum
The coherence spectrum may be interpreted as the fraction of the output
time series that is linearly due to the input time series. A spectral value of
unity at a given frequency indicates a fully linear coupling, whereas for
zero value no linear relationship can be expected. Values between zero and
unity indicates that the time series are noise-affected, the relation between
the time series are not linear or the output time series is a function of sev-
eral input time series.
Output
You specify which type of integral spectral parameters to be calculated.
The estimates are saved in the log file, which as default is automatically
opened after execution.
Spectral moments
The first four spectral moments are calculated. The spectral moment is
defined as
f max
mn = ∫fmin SX ( f ) f n df (3.2)
Peak parameters
The spectral peak frequency fp is the frequency corresponding to the max-
imum spectral density.
2 fmax
Q p = ------ ∫ f S X ( f ) 2 df (3.3)
m 0 fmin
23
WS Linear Spectral Analysis
Spectral width
The spectral width (or broadness) parameter is defined as
m 22
ε4 = 1 – ------------- (3.4)
m0 m4
Peak-to-peak estimates
The significant wave height is defined as
H m0 = 4 m 0 (3.5)
1
Hp = 8m 0 1n --- (3.6)
p
Period estimates
The peak wave period Tp defined as
1
T p = --- (3.7)
fp
m0
T 01 = ------ (3.8)
m1
m0
T 02 = ------ (3.9)
m2
m2
T 24 = ------ (3.10)
m4
The selection of items for analysis is chosen from a dynamic table con-
taining information about the selected data file. The two left columns con-
tain item names for input and output items (dual item analysis only). The
rightmost column contains a combo box for each item, by which it is pos-
sible to state if the item should be included in the analysis.
25
WS Linear Spectral Analysis
The “Parameter set” is set to “None” for disregarding the current setting.
In the Output File dialog you specify the output file name and folder
where to place the data file.
Log file
As default, the log file is automatically opened after model execution. The
file is opened in Notepad.
3.9 References
Bendat, J.L. and Piersol, A.G., 1986. Random Data. Analysis and Meas-
urement Procedures. John Wiley & Sons.
27
WS Linear Spectral Analysis
4.1 Introduction
The Digital Filtering Analysis module provides different digital filter
options. Basically filtering is a process of selecting, attenuating, or sup-
pressing certain frequency components of a time series signal by a desired
amount, allowing a digital filter to shape the frequency spectrum of the
signal.
The Digital Filtering Analysis module uses a class of filters called Finite
Impulse Response (FIR) filters also known as non-recursive filters. These
filters are characterised by filter coefficients (in the time domain), which
are derived from specified transfer functions in the frequency domain. The
actual filter operation takes place in time domain by convoluting the input
time series and the filter coefficients.
Figure 4.1 The figure shows a time series of pressure converted to a time
series of surface elevation by digital filtering
29
WS Digital Filtering Analysis
Whenever the editor recognises the used file format an info box contain-
ing the files header information appears in the lower half of the dialog.
You can view the selected data file using the MIKE Zero Time or Grid
Series editor by clicking on the “View” button.
31
WS Digital Filtering Analysis
4.3 Area
The Area dialog is only active when a dfs2-type data file has been speci-
fied as input. You can choose to make an analysis on the entire domain
(i.e. by selecting every grid point) or just a number of points by specifying
the model coordinates (j,k).
4.4 Time
In the Transfer Functions dialog you specify the name and define one or
more filter components. All filter components (transfer functions) are
defined in the frequency domain. The available basic transfer functions
and parameters are listed below.
Low-pass
The transfer function for a low-pass filter suppresses all frequency compo-
nents of the time series above the cut-off frequency.
High-pass
The transfer function for a high-pass filter suppresses all frequency com-
ponents of the time series below the cut-off frequency.
Notch
The transfer function for a notch filter suppresses a specific frequency
component of the time series.
33
WS Digital Filtering Analysis
Band-pass
The transfer function for a band-pass filter lets all frequency components
of the time series between the lower and higher cut-off frequencies pass.
Band-stop
The transfer function of a band-stop filter suppresses all frequency com-
ponents of the time series between the lower and higher cut-off frequen-
cies.
Integration
The transfer function for an integration filter integrates the input time
series to a given order.
Differentiation
The transfer function for a differentiation filter differentiates the input
time series to a given order.
cosh ( kh )
H pη ( f ) = ------------------------------------ (4.1)
cosh ( k ( z + h ) )
in which z is the distance from still water level to the pressure sensor
(must be negative). As the transfer function increases rapidly with fre-
quency and potentially causes a dramatic amplification of measurement
noise a maximum amplification factor may be given. The wave number k
is given by the dispersion relation:
Moving average
The transfer function for a moving average filter defines an equivalent
weighting of the time series data within the period.
Biésel (piston)
The transfer function for a Biesel (piston) filter attenuates all frequency
components of a piston position time series for a piston type wave maker
(with zero deviation) to an equivalent time series of surface elevation
(Biésel, 1951).
4sinh 2 ( kh )
A Xη ( f ) = ---------------------------------------- (4.3)
2kh + sinh ( 2kh )
Evanescent (piston)
The transfer function for an evanescent mode (piston) filter produces the
frequency dependent local disturbance due to a varying piston position of
a piston type wave maker (with zero deviation) (Schäffer et. al., 1994).
∞
4 sinh 2 ( kh )
D Xη ( f ) = ∑ -----------------------------------------
2kh + sinh ( 2k j h )
- (4.4)
1
Bartlett (low-pass)
The transfer function for a Bartlett or triangular filter yields a low-pass fil-
ter for which the attenuation above the cut-off frequency smoothly
decreases towards zero.
cosh ( k ( z + h ) )
H ηp ( f ) = ------------------------------------ (4.6)
cosh ( kh )
35
WS Digital Filtering Analysis
In which z is the distance from still water level to the level at which the
pressure should be calculated (must be negative and not exceed the water
depth). The wave number k is given by the dispersion relation.
2
( 2πf ) = gk tanh ( kh ) (4.7)
Elevation displacement
The transfer function of a filter for displacement of a surface elevation a
certain distance in the propagation direction produces a phase shift of all
frequency components (no attenuation occurs).
B x ( f ) = – kx (4.8)
2
( 2πf ) = gk tanh ( kh ) (4.9)
In the Basic Parameters dialog you define one or more parameter sets
(definitions) for the analyses. Already defined parameter sets can be tem-
porarily excluded from the analysis by setting the number of definitions to
“Include all definitions”.
37
WS Digital Filtering Analysis
4.6.1 Definitions
For each Definition dialog you select which of the defined transfer func-
tions that composes the filter. The FIR filter coefficients for a given filter
definition must be generated to match the time step of the input time
series. For each definition a filter width, and possibly a truncated filter
width, must be specified.
Filter width
The filter width is as minimum given by the required frequency domain
resolution as the frequency interval must be less than the minimum cut-off
frequency:
1
∆f = f min = ----------
-
2 N ∆t (4.10)
w = 2 N ∆t
Truncate filter
When filter coefficients are generated it is important to notice the possibil-
ity of truncating the filter coefficients. For filters that do not have very low
cut-off frequencies, the transfer function will only be marginally affected,
but the computational efficiency may be improved considerably. If the fil-
ter coefficients are truncated, the filter coefficients corresponding to the
largest time lag are omitted.
Truncating the filter coefficients will shorten the sequence of delete values
that always will appear at the start and end of the filtered time series.
39
WS Digital Filtering Analysis
The selection of items for filtering is chosen from a dynamic table con-
taining information about the selected data file. The left column contains
item names for input items. The rightmost column contains a combobox
for each item, by which it is possible to state, which of the defined param-
eter sets (filters) that should be used for filtering of the selected item. To
exclude the item from the analysis the “Parameter set” must be set to
“None”.
In the Output File dialog you specify the output file name and folder
where to place the data files containing the filtered time series and finite
impulse response functions (FIR). Further for dfs2-type input data the
energy of the filtered time series is calculated at all grid points. If one of
the output files already exist you can view the selected data files using the
MIKE Zero Time Series or Grid Editor by clicking on the “View” button.
Log file
As default, the log file is automatically opened after model execution. The
file is opened in Notepad.
4.9 References
Biésel F. (1951) 'Les appareils générateurs de houle en laboratoire' La
Houille Blance, Vol 6, Nos 2, 4 et 5.
41
WS Digital Filtering Analysis
5.1 Introduction
The WS Directional Wave Analysis module is an efficient tool for direc-
tional analysis of concurrent data of surface elevation and orthogonal
velocities (or fluxes). Hence the directional wave analysis can be per-
formed for data from
Figure 5.1 The calculated mean wave direction on top of a bathymetry. The
input data is based on a MIKE 21 BW Boussinesq Waves simulation
with directional wave input
43
WS Directional Wave Analysis
Figure 5.2 Calculated directional spectrum at the two read dots indicated on
the map above
The module provides the following different spectra, functions and prop-
erties:
The Directional Wave Analysis Module supports two types of input files:
Most often the latter type is only relevant for analysis of data originating
from a numerical time domain wave model like MIKE 21 BW.
You can view the selected data file using the MIKE Zero Time or Grid
Series editor by clicking on the “View” button.
45
WS Directional Wave Analysis
Figure 5.4 shows the two ways you save the required data in the Boussin-
esq Waves Editor. It is recommended to use the first procedure.
Figure 5.4 How to save data in the MIKE 21 BW editor for subsequent direc-
tional wave analysis in a single point or for an area
47
WS Directional Wave Analysis
of the internal memory in your PC, i.e. less than 0.5-2 GB. The most effi-
cient way to limit the file size is to save data on a coarser grid (e.g. very
5th grid point).
5.3 Area
The Area dialog is only active when a dfs2-type data file has been speci-
fied as input. You can choose to make an analysis on the entire domain
(i.e. by selecting every grid point) or just a number of points by specifying
the model co-ordinates (j,k).
Most often you will make an analysis for all data in the domain for assess-
ment of the spatial variation of e.g. the mean wave direction.
5.4 Time
49
WS Directional Wave Analysis
The parameters given in the FFT Parameters dialog determines the fre-
quency resolution and the degree of smoothing of the spectrum.
5.5.1 Resolution
FFT size:
Number of data point in each FFT block or subseries. The length must be
a power of two and cannot exceed the total number of data points
Number of frequencies:
Is half of the FFT size
FFT duration:
Gives the length of the FFT analysis in seconds
Frequency interval:
The difference between two discrete adjacent frequencies. If manually
changed, the value is automatically adjusted to the nearest possible value
(function of time step and FFT size)
Type
Three different data windows are available:
Overlap
The spectra (and response functions) are calculated using overlapping
subseries. The overlap controls the smoothing of spectra, but it also con-
trols the weighting of the different parts of the input time series.
Number of subseries
Based on the specified FFT parameters the number of subseries can be
calculated. It is almost always desirable to have a large number of sub-
series because it will reduce the variance of the spectral density and pro-
vide a smooth spectral estimate.
For non-overlapping time series the number of degrees of freedom for the
spectral estimate is two times the number of subseries. Spectral estimates
having 30-40 degrees of freedom is often used in practice.
Utilisation
Indicates the utilisation of the input data.
Minimum frequency
The minimum cut-off frequency must be greater than or equal to zero.
51
WS Directional Wave Analysis
Maximum frequency
The maximum cut-off frequency must be less than or equal to the Nyquist
frequency, which is determined entirely by the sampling frequency f s of
the input data:
1
f max = 0.5 f s = -------- (5.1)
2 ∆t
Here ∆t is the time step of the input data. If the value is out of range a
warning is issued and the invalid maximum cut-off frequency is automati-
cally adjusted to the Nyquist frequency.
It is important to realise that the various dialogs are all synonyms for the
FFT size, and that changing one parameter also will affect the others.
In the Basic Parameters dialog you specify general parameters for the
directional analysis and output types.
53
WS Directional Wave Analysis
Using spectrum:
z Directional spectrum
z Surface elevation spectrum (directional integrated)
z Mean wave direction function and spreading function
z Directional energy distribution (frequency integrated)
In the Analysis Selections dialog you specify which items (in the input
data files) the directional analysis should be based on. You specify:
The “Parameter set” is set to “Basic” for analysis or “None” for disre-
garding the current setting.
In the Output File dialog you specify the output file name and folder
where to place the data files.
Parameters
For dfs2-type input data following parameters are calculated at each grid
point:
55
WS Directional Wave Analysis
The calculation of mean wave direction and spreading complies with the
suggestions and definitions given by Frigaard et al. (1997).
f max S ( f )
MWD = θ X = arg ∫ -----------
X
- exp ( iθ m, X ) df (5.2)
fmin m 0, X
X ( f )σ θ, X ( f )
f max S
DSD = σ θ, X = ∫fmin -----------------------------
m 0, xX
- df
(5.3)
σ θ2, X = 2 ( 1 – c 1 )
2π 2 2π 2
c 1 2 = ∫ D X ( f, θ ) sin θ dθ + ∫ D X ( f, θ ) cos θ dθ
0 0
H m0 = 4 m 0
(5.4)
f max
m0 = ∫fmin S X ( f ) df
m0
T 02 = ------
m2 (5.5)
f max
m2 = ∫fmin S X ( f )f 2 df
1
T p = --- (5.6)
fp
Log file
As default the log file is automatically opened after model execution. The
file is opened in Notepad.
5.10 References
Nwogu, O., 1989. Maximum entropy estimation of directional wave spec-
tra from an array of wave probes. App. Ocean Res., 11, 176-182.
Frigaard, P., Helm-Petersen, J., Klopman, G., Stansberg, C.T., Benoit, M.,
Briggs, M.J., Miles, M., Santas, J., Schäffer, H.A. & Hawkes, P.J. /1997/
IAHR List of Sea state parameters - an update for multidirectional waves.
Proceedings of 27th IAHR Seminar, San Francisco, USA
57
WS Directional Wave Analysis
Hashimoto, N., 1997. Analysis of the directional wave spectra from field
data. Advances in Coastal and Ocean Engineering Vol. 3, ed. Liu, P.L.F.
World Scientific, Singapore, 103-143.
Bendat, J.L. and Piersol, A.G., 1986. Random Data. Analysis and Meas-
urement Procedures. John Wiley & Sons.
6 WS CROSSING ANALYSIS
6.1 Introduction
The WS Crossing Analysis module is a generalisation of the classical
zero-crossing analysis. For a predefined threshold, the input time series
item is divided into events, each of which is defined by the time series
value crossing the reference level in upwards (or downwards) direction.
59
WS Crossing Analysis
Figure 6.1 Scatter diagram for wave heights and periods detected within a time
series acquired during a physical model test
The WS Crossing Analysis module supports time series data with equidis-
tant time step. At present the following two binary file formats are sup-
ported:
Whenever the editor recognises the used file format an info box contain-
ing the files header information appears in the lower half of the dialog.
You can view the selected data file using the MIKE Zero Time Series edi-
tor by clicking on the “View” button.
61
WS Crossing Analysis
6.3 Time
In the Basic Parameters dialog you define one or more parameter sets
(definitions) for the analyses. Already defined parameters sets can be tem-
porarily excluded from the analysis by setting the number of definitions to
“Include all definitions”.
63
WS Crossing Analysis
6.4.1 Definitions
For each Definition dialog you select method, parameters and output from
the crossing analysis.
Method
Defines whether the crossing analysis must be performed in upward
(default) or downward direction.
Crossing level
Is the level, at which the up- or down crossing of which detects a new
event. You can specify a value or use the time series mean value (default).
Threshold level
The threshold level is a part of a low-level crossing filter for elimination
of noise. The threshold level defines a small amount by which the refer-
ence level must be exceeded before the beginning of an event is accepted.
You can specify an absolute value or define a level relative to the standard
deviation of the time series (default).
Threshold period
The threshold period is a part of a low-level crossing filter for elimination
of noise. The threshold period defines a duration that must be exceeded
before the can event can be accepted. An absolute value must be given
(default is zero).
Output
For each event the following output is available:
65
WS Crossing Analysis
The selection of items for analysis is chosen from a dynamic table con-
taining information about the selected data file. The left column contains
item names for input items. The rightmost column contains a combo box
for each item, by which it is possible to state which of the defined parame-
ter sets that should be used for analysis of the selected item. To exclude
the item from the analysis the “Parameter set” must be set to “None”.
In the Output File dialog you specify the output file name and folder
where to place the data file. If the output file already exists, you can view
the selected data file using the MIKE Zero Time Series editor by clicking
on the “View” button.
Log file
As default, the log file is automatically opened after model execution. The
file is opened in Notepad.
67
WS Crossing Analysis
69
Index
A I
Auto spectrum . . . . . . . . . . . . . 22 Integration . . . . . . . . . . . . . . . . 34
B L
Band-pass . . . . . . . . . . . . . . . . 34 Linear Spectral Analysis . . . . . . . . 12
Band-stop . . . . . . . . . . . . . . . . 34 Log file . . . . . . . . . . . . . . . . . . 27
Bartlett (low-pass) . . . . . . . . . . . 35 Low-pass . . . . . . . . . . . . . . . . 33
Biesel (piston) . . . . . . . . . . . . . 35
M
C Maximum Entropy Method (MEM) . . 12
Coherence spectrum . . . . . . . . . . 23 Maximum entropy method (MEM) . . 43
Coherent power spectrum . . . . . . . 23 Mean wave period . . . . . . . . . . . 24
Cosine squared . . . . . . . . . . . 20, 51 Method . . . . . . . . . . . . . . . . . . 64
Crossing analysis . . . . . . . . . . . 64 MIKE 21 BW Boussinesq Waves . 12, 30
Crossing Analysis Module . . . . . . . 13 MIKE Zero Time or Grid Series editor 31
Crossing level . . . . . . . . . . . . . . 64 MIKE Zero Time Series editor .41, 61, 67
Cross-spectrum . . . . . . . . . . . . . 22 Moving average . . . . . . . . . . . . . 34
Cut-off frequencies . . . . . . . . . 20, 51
N
D Non-recursive filters . . . . . . . . 12, 29
Data window . . . . . . . . . . . . 20, 51 Notch . . . . . . . . . . . . . . . . . . . 33
Definition . . . . . . . . . . . . . . . . 22 Nyquist frequency . . . . . . . . . 21, 52
Differentiation . . . . . . . . . . . . . . 34
Digital Filtering Analysis . . . . . . . . 12 O
Directional parameters . . . . . . . . 53 Output . . . . . . . . . . . . . . . . . . 65
Directional Wave Analysis . . . . . . 12 Output file . . . . . . . . . . . . . . . . 27
Directional wave analysis . . . . . . . 43
Dual item analysis . . . . . . . . . . . 26 P
Parameter set . . . . . . . . . . . . 26, 66
E Parameter sets . . . . . . . . . . . 63, 66
Elevation displacement . . . . . . . . 36 Peak parameters . . . . . . . . . . . . 23
Evanescent (piston) . . . . . . . . . . 35 Peak wave period . . . . . . . . . . . . 24
Peak-to-peak estimates . . . . . . . . 24
F Period estimates . . . . . . . . . . . . 24
Fast Fourier Transform (FFT) . . . 12, 15 Power spectrum . . . . . . . . . . . . . 22
Filter width . . . . . . . . . . . . . . . 38 Pressure to surface elevation . . . . . 34
Finite Impulse Response (FIR) . . 12, 29
FIR filter coefficients . . . . . . . . . . 38 R
Frequency response spectrum . . . . 23 Rectangular data window . . . . . 20, 51
Frequency spectrum . . . . . . . . . . 22
S
H Single item analysis . . . . . . . . . . 26
Hanning data window . . . . . . . 20, 51 Spectral moments . . . . . . . . . . . 23
High-pass . . . . . . . . . . . . . . . . 33 Spectral width . . . . . . . . . . . . . . 24
Surface elevation to pressure . . . . . 35
T
Threshold level . . . . . . . . . . . . . 64
Threshold period . . . . . . . . . . . . 65
Truncate filter . . . . . . . . . . . . . . 39
U
User background . . . . . . . . . . . . 9
V
Variance spectrum . . . . . . . . . . . 22
71
Index