|
Crystal Clear SSI
A new feature resulting in extremely clear
stabilization diagrams is now introduced as an add-on to the well-known SSI
algorithms available in ARTeMIS Extractor Pro and ARTeMIS Analyzer Pro. This new
revolutionary feature is Crystal Clear Stochastic Subspace Identification
(CC-SSI).
A Conditional Estimation Technique
The SSI techniques rely on linear least
squares estimation of the model using the raw measured time series.
In the past this estimation was unconditionally. Now the
revolutionary Crystal Clear SSI estimation feature allow a
conditional estimation emphasizing a user defined number of physical
modes and suppressing the remaining parts of the information in the
data. This results in extremely clear stabilization of the number of
modes specified and nearly no noise modes inside the visible
frequency range.
The basic idea is to specify the maximum
number of significant poles (eigenvalues) present in the
measurements that should be estimated. The estimation algorithm will
then focus on the modes having these poles and any less significant
noise poles are returned with a natural frequency estimate much
higher than the Nyquist frequency, and a damping ratio of 100 %. The
maximum number of poles can be specified by the user, or is
estimated automatically by the software using a special
data-dependent algorithm.
Crystal Clear SSI option can either be
enabled in automatic mode where it by itself suggest the number of
poles (eigenvalues) to use in the conditional estimation, or in
manual mode where the user specify the number.
The algorithm has proven to be extremely
robust in many difficult cases such as: ·
Due to the highly consistent estimation of
the poles, the search for the optimal model order is less critical
when using this new feature. Especially the damping estimates are
much more robust now than ever before.
Crystal Clear SSI analysis of a fighter jet wing with many
modes. The modes indicated with a red A on the top of the diagram are
extracted using the new
Automatic Mode
Estimation features. See the close-up picture below.
|