|
|
 |
|
Time Domain Modal Analysis |
At a Glance
-
Basic Method: Data driven Stochastic Subspace Identification.
User choices: Implementation: Unweighted Principal components,
Principal Components and Canonical Variate Analysis.
-
Model orders: from one mode to the size defined by the common
SSI input matrix.
-
Stabilization criteria: Natural frequency deviation, damping
ratio deviation, mode shape MAC deviation, initial modal
amplitude MAC deviation, all limits user defined.
-
Physical mode separation: Damping ratio limits are user defined.
-
Select and link: Modes from the models chosen from each test
setups are selected and linked using snap functions and editing
facilities.
-
Uncertainty estimation: In case of several test setups, the
empirical standard deviation is calculated for natural
frequencies and damping ratios.
-
Comparison: Estimated model vs. FFT based auto and cross-spectra
of measurements.
-
Spectra and correlation functions of prediction errors between
estimated model and measurements.
The editor of
one of the three
Stochastic
Subspace
Identification
techniques
available in the
ARTeMIS
Extractor Pro.
Accurate
identification
In the time
domain approach
the user can
perform a very
accurate modal
identification
using up to
three different
types of data
driven
Stochastic
Subspace
Identification
(SSI)
techniques. The
techniques use
all response
data available
estimating a
full model in
discrete time.
Flexible
stabilization
criteria
It is possible
to specify a
model order from
one mode to the
maximum number
of modes
determined by
the size of the
common SSI input
matrix specified
in the signal
processing
configuration.
There is an
option to
specify model
stabilization
criteria based
on deviations of
damping ratio,
natural
frequency, mode
shape and
initial modal
amplitude. The
latter describes
the quality of
the statistical
modeling.
Model validation
The user can
validate the
quality of each
model by
selecting the
appropriate tabs
in the window to
inspect either
magnitude or
phase of the
spectral fit or
the fit on the
correlation
function.
Validation plots
are displayed to
show the
(synthesized)
modal model
curves compared
with the curves
directly
obtained from
processed
measurements.
Back to
Technical
Details
|
