University of Thessaly (UTH)
Ανάλυση και ταυτοποίηση χωροχρονικών φαινομένων με χρήση προχωρημένων μεθόδων ανάλυσης χρονοσειρών
Abstract
dc:descriptionThe objective of the present thesis is the analysis of time series, using mainly the non-linear analysis methods, giving also emphasis to the most recently developed method of transforming time series to complex networks. The main research question that was the possibility to the identification regions presenting different dynamical behavior in the case of spatiotemporal phenomena, i.e. dynamical systems whose behavior evolves both in space and time. Our aim was to identify the dynamical state of the system, through time series analysis (system identification), to understand the dynamics of the underlying process and extract characteristic temporal/spatial characteristics of the system. The analysis was applied both to experimental and field time series. The experimental time series concern the study of various turbulent jet flow, given that in a turbulent flow there are small and large scale structures (vortices) in different temporal and spatial scales a fact that offers an excellent case study of spatiotemporal phenomenon. The field time series originated from seawatch buoys that have been developed by the Hellenic Center for Marine Research (HCMR) in the framework of the POSEIDON project. Moreover, in this thesis the case of reduction of the data of the time series through a selection of successive pairs of maximum-minimum values and the effect on the identification of different regions.. Temporal correlations were examined by using the cross correlation function and the Granger causality analysis. This analysis concluded that the study of experimental or field time series can significantly contribute to the identification of changes in the system dynamics. The classical methods of statistical analysis combined with other methods, such as Hjorth analysis, the average mutual information, and the Hurst exponent analysis, allow the detection of regions which present different dynamical behavior, while at the same time allows us to draw conclusions about their temporal characteristics. In addition, this thesis reveals the importance and usefulness of time-series study through converting them into complex networks. It seems that, in this way, the system dynamics are significantly depicted in the resulting network while there is a large number of topological measures of the resulting network which allow us to distinguish the different characteristics of the dynamics of the systems. Additionally, the study of time series through maxima-minima allows the identification of various regions with different dynamical behavior in the case of spatiotemporal phenomena, which is especially important in cases where the length of data is large and in cases of analysis that are conducted in real time as it reduces the processing and the decision making time. In summary, the proposed methodology of this study can be applied to a wide range of applications. In particular, the results of this study could institute the basis for future research of more complex spatiotemporal phenomena in the engineering problems and methodologies could be integrated into automatic detection systems that could identify changes of state in a system such as in early warning systems.
Degree
thesis:*- Grantor dc:publisher
- University of Thessaly (UTH)
- Year dc:date
- 2015
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Charakopoulos, Avraam
- Χαρακόπουλος, Αβραάμ
Subjects
dc:subject × 26- Χρονοσειρές, μη γραμμική ανάλυση
- Δυναμικά συστήματα, μη γραμμικά
- Ταυτοποίηση συστήματος
- Χωροχρονικά φαινόμενα
- Μη γραμμική ανάλυση χρονοσειρών
- Πολύπλοκα δίκτυα χρονοσειρών
- Διασυσχέτιση
- Αιτιότητα χρονοσειρών
- Τυρβώδης ροή
- Non linear time series analysis
- Dynamical systems, nonlinear
- System identification
- Spatiotemporal phenomena
- Non-linear analysis
- Complex network time series
- Cross correlation
- Granger causality
- Turbulent flow
- Φυσικές Επιστήμες
- Φυσική
- Επιστήμες Μηχανικού και Τεχνολογία
- Άλλες Επιστήμες Μηχανικού και Τεχνολογίες
- Natural Sciences
- Physical Sciences
- Engineering and Technology
- Other Engineering and Technologies
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/39811