Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | Proceedings of the 8th International Conference on Structural Dynamics, EURODYN 2011 |
| Herausgeber/-innen | G. Lombaert, G. Muller, G. De Roeck, G. Degrande |
| Erscheinungsort | Leuven |
| Seiten | 2907-2912 |
| Seitenumfang | 6 |
| Publikationsstatus | Veröffentlicht - 2011 |
| Extern publiziert | Ja |
| Veranstaltung | 8th International Conference on Structural Dynamics, EURODYN 2011 - Leuven, Belgien Dauer: 4 Juli 2011 → 6 Juli 2011 |
Abstract
In many engineering disciplines and industrial applications dynamical processes appear and must be analyzed in order to interpret the physical and mechanical characteristics. In a general case, these processes are affected by noise due to the uncertainity of input quantities, environmental and boundary conditions, system properties, and numerical operations (or experiments). It is assumed that a bundle of process records is available from numerical simulations or from experiments. Within these bundles, bunches of processes exist, which are more or less heterogeneous to each other. Each of these bunches is associated with a specific physical path which the mechanical system can follow, and different bunches characterize different physical behavior. The identification of these paths and of the associated physical and mechanical behavior is essential to understand a mechanical system and to design it for a specific purpose. In this paper a method for the identification of bunches and their representative paths is proposed. Cluster analysis methods are utilized to identify process bunches. This includes the formulation of specific distance measures to consider aspects of adjacent positions, correlation, and affinity between individual process records. The revealed information about the bunches and their representative paths is eventually used to assess the sensitivity and robustness of a dynamical system.
ASJC Scopus Sachgebiete
- Informatik (insg.)
- Hardware und Architektur
- Informatik (insg.)
- Computernetzwerke und -kommunikation
- Ingenieurwesen (insg.)
- Steuerungs- und Systemtechnik
- Ingenieurwesen (insg.)
- Elektrotechnik und Elektronik
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
Proceedings of the 8th International Conference on Structural Dynamics, EURODYN 2011. Hrsg. / G. Lombaert; G. Muller; G. De Roeck; G. Degrande. Leuven, 2011. S. 2907-2912.
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - Identification of representative paths in noisy processes
AU - Liebscher, Martin
AU - Reuter, U.
AU - Beer, Michael
AU - Mehmood, Zeeshan
PY - 2011
Y1 - 2011
N2 - In many engineering disciplines and industrial applications dynamical processes appear and must be analyzed in order to interpret the physical and mechanical characteristics. In a general case, these processes are affected by noise due to the uncertainity of input quantities, environmental and boundary conditions, system properties, and numerical operations (or experiments). It is assumed that a bundle of process records is available from numerical simulations or from experiments. Within these bundles, bunches of processes exist, which are more or less heterogeneous to each other. Each of these bunches is associated with a specific physical path which the mechanical system can follow, and different bunches characterize different physical behavior. The identification of these paths and of the associated physical and mechanical behavior is essential to understand a mechanical system and to design it for a specific purpose. In this paper a method for the identification of bunches and their representative paths is proposed. Cluster analysis methods are utilized to identify process bunches. This includes the formulation of specific distance measures to consider aspects of adjacent positions, correlation, and affinity between individual process records. The revealed information about the bunches and their representative paths is eventually used to assess the sensitivity and robustness of a dynamical system.
AB - In many engineering disciplines and industrial applications dynamical processes appear and must be analyzed in order to interpret the physical and mechanical characteristics. In a general case, these processes are affected by noise due to the uncertainity of input quantities, environmental and boundary conditions, system properties, and numerical operations (or experiments). It is assumed that a bundle of process records is available from numerical simulations or from experiments. Within these bundles, bunches of processes exist, which are more or less heterogeneous to each other. Each of these bunches is associated with a specific physical path which the mechanical system can follow, and different bunches characterize different physical behavior. The identification of these paths and of the associated physical and mechanical behavior is essential to understand a mechanical system and to design it for a specific purpose. In this paper a method for the identification of bunches and their representative paths is proposed. Cluster analysis methods are utilized to identify process bunches. This includes the formulation of specific distance measures to consider aspects of adjacent positions, correlation, and affinity between individual process records. The revealed information about the bunches and their representative paths is eventually used to assess the sensitivity and robustness of a dynamical system.
KW - Cluster analysis
KW - Distance measures
KW - Noisy processes
KW - Representative paths
UR - http://www.scopus.com/inward/record.url?scp=84959918382&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84959918382
SN - 9789076019314
SP - 2907
EP - 2912
BT - Proceedings of the 8th International Conference on Structural Dynamics, EURODYN 2011
A2 - Lombaert, G.
A2 - Muller, G.
A2 - De Roeck, G.
A2 - Degrande, G.
CY - Leuven
T2 - 8th International Conference on Structural Dynamics, EURODYN 2011
Y2 - 4 July 2011 through 6 July 2011
ER -