|
|
 |
|
NIA Seminar by Mohammad Hajj |
|
Date: Friday, July 13, 2007
Time: 10:30am
Location: NIA, Room 137
Additional Information: Presentation (PDF)
Nonlinear System Identification with Higher-Order Spectral Moments
Prof. Muhammad R. Hajj, Virginia Tech
One of the most applied procedures for data analysis is the estimation of the power spectrum which yields the energy content of different frequency components in a signal. Yet, there is much more information in stochastic and deterministic signals than is obtained from the power spectrum. This information can be obtained from the higher-order spectral moments. Because they preserve phase information, these moments can be used to identify nonlinearly coupled frequency components in one or more signals and quantify the extent of this interaction. This aspect inspires the idea of proposing them as tools for the analysis, interpretation, and modeling of random time series data as measured or numerically generated from linear and nonlinear systems. The objective of this talk is to show how additional information, as obtained from higher-order spectral analysis, can be used to identify and characterize system nonlinearities. Examples presented will stress the identification of nonlinear dynamics associated with the flutter of the Flexible Semispan Model of a High Speed Civil Transport wing configuration tested in the TDT at NASA LaRC and with Limit Cycle Oscillations (LCO) of F16 aircraft in flight tests. The concept of exploiting identified nonlinearities in the development of reduced-order models will also be discussed.
|
|
|
Home