Related Experiment Videos
A substructure approach for the midfrequency vibration of stochastic systems
Abhijit Sarkara1, Roger Ghanem
1Department of Mechanical Engineering, McGill University, 817 Sherbrooke Street, Montreal, Quebec, Canada. asarkar@mecheng.mcgill.ca
The Journal of the Acoustical Society of America
|April 22, 2003
Summary
A new substructure coupling technique uses proper orthogonal decomposition to efficiently model vibrations in uncertain systems. This method enables adaptive bases and reduced-order models for accurate probabilistic predictions.
Area of Science:
- Mechanical Engineering
- Computational Mechanics
- Vibration Analysis
Background:
- Modeling mid-frequency vibrations in linear dynamical systems with parameter uncertainty is computationally challenging.
- Existing methods may not efficiently handle the probabilistic characterization of complex structures.
Purpose of the Study:
- To present a novel substructure coupling technique for mid-frequency vibration analysis.
- To develop an efficient method for probabilistic characterization of model predictions in uncertain systems.
Main Methods:
- Utilizing proper orthogonal decomposition (POD) for adaptive basis derivation within subsystems.
- Constructing reduced-order models for global structures based on subsystem POD.
- Integrating the substructure approach with stochastic finite element methods (SFEM).
Main Results:
- The proposed methodology enables efficient derivation of adaptive bases and reduced-order models.
- The technique facilitates accurate probabilistic characterization of predictions for uncertain systems.
- Comparison with component mode synthesis highlights similarities and distinctions.
Conclusions:
- The novel substructure coupling technique offers an efficient approach to mid-frequency vibration analysis.
- The method effectively integrates with stochastic finite element analysis for uncertain systems.
- Proper orthogonal modes derived from frequency and time domains show varying suitability based on system behavior.