|
|
 |
|
Signal Processing |
At a Glance
-
Decimation,
1-1000
times,
digital
anti-aliasing
filter,
cut-off at
0.8 times
Nyquist
frequency of
decimated
signals.
-
Filtering:
low-pass,
high-pass,
band-pass,
bandstop
Butterworth,
filter order
1-50 poles,
arbitrary
cut-off
frequencies,
test for
filter
stability.
-
Projection
Channels.
Specify is
all cross
information
of the
measurement
channels
should be
used or only
between
fewer number
of channels.
-
Spectral
estimation
using FFT
and Welchs
averaged
periodogram
method. Data
segment
length:
radix-2 only
limited by
the amount
of data,
overlap:
66.7 %,
window:
Hamming.
Estimates
the full
spectral
matrix if
the
Projection
Channel
option is
disabled.
-
Operating
Deflection
Shapes.
Specify the
type of data
imported to
ARTeMIS
Extractor
(displament,
velocity,
acceleration,
other) and
specify the
unit of the
imported
data to
ensure
correct
units of the
operating
deflection
shapes.
-
Common SSI
input matrix
estimation
to be used
in all
stochastic
subspace
identification
algorithms.
Maximum
state space
dimension:
Only limited
by the
available
amount of
data. Noise
mode
suppresion
through the
use of the
Projection
Channel
option.
-
Option for
reloading
original
uploaded
data.
-
Data
Presentation
of Test
Setups:
Spectral
magnitude
and phase,
singular
value
decomposition
of spectral
matrix,
average of
full
spectral
matrix,
average of
diagonal
elements of
spectral
matrix,
coherence of
spectral
matrix,
filter
characteristic,
correlation
functions.
Cursor read
out on all
curves.
-
Data
Presentation
of Reference
Data:
Spectral
magnitude
and phase,
average of
full
spectral
matrix,
average of
diagonal
elements of
spectral
matrix,
coherence of
spectral
matrix.
Cursor read
out on all
curves.
Signal
Processing Tools
As the first
step in the
identification
process the
measured
vibration data
has to be
carried through
different kinds
of signal
processing to
enhance the
physical
information that
the user is
looking for and
thus, to ensure
reliable
identification
results. In
ARTeMIS
Extractor, it is
possible to
decimate the
signal to reduce
the sampling
frequency, apply
high-pass,
low-pass,
band-pass and
band-stop
filters, to
estimate
spectral
densities and
estimate the
common SSI input
matrix that is
used later in
the Stochastic
Subspace
Identification
(SSI)
techniques.
Data
Presentation
At any stage of
the signal
processing it is
possible to view
the processed
data, as shown
below. It
is possible to
view all auto
and cross
spectral
densities and
condensed plots
like singular
value
decomposition of
the spectral
matrix or the
average of the
spectral density
matrix or its
main diagonal.
View Processed
Data window with
Singular Value
Decomposition
diagram and
cross-spectral
densities
presented.
Cursor and zoom
facilities are
available in all
displays.
Back to
Technical
Details
|
