A5103062850- Fluid Dynamics and Turbulent Flows
- Heat Transfer and Boiling Studies
- Wind and Air Flow Studies
- Heat Transfer and Optimization
- Molecular Biology Techniques and Applications
- Heat Transfer Mechanisms
- Fluid Dynamics and Thin Films
- Forensic and Genetic Research
- Bacillus and Francisella bacterial research
- Nanofluid Flow and Heat Transfer
- Image and Signal Denoising Methods
- Forensic Anthropology and Bioarchaeology Studies
- Complex Systems and Time Series Analysis
- Model Reduction and Neural Networks
- Spacecraft and Cryogenic Technologies
- Plant Water Relations and Carbon Dynamics
- Fluid Dynamics and Vibration Analysis
- Chaos control and synchronization
Universidade da Coruña
2002-2019
Polytechnic University of Puerto Rico
2004
This paper is concerned with the problem of analyzing turbulent flow signals that are irregularly sampled by a laser Doppler velocimeter. The temporal irregularity sampling main addressed due to difficulties it introduces in use traditional analysis techniques. contribution this assess adequateness introducing signal-modeling strategy means delay-based artificial neural networks (ANNs) adaptively and continuously trained online on real data using an evolutionary-based called multilevel...
The analysis of turbulent flow signals irregularly sampled by a laser Doppler velocimeter is assessed means ANNs. This technique has been proven to correctly predict the time evolution signals. We are taking advantage this ability obtain models unevenly and thus be able reconstruct resample them at regular pace in order allow for their conventional
A Neural Network based system for reconstructing turbulent signals obtained by a Hot Wire Anemometer in the shear layer surrounding free yet of air is presented. The flow-field used as work bench presents very complex turbulence dynamics making measured difficult to predict. results presented here show good agreement prediction and present high potential an analysis tool.
The signals obtained by a hot wire anemometer when measuring at various points within two different and well known turbulent flow fields are used as test cases to asses an analysis system including delay based neural networks. method thus developed proves be suitable for reconstructing the and, additionally, extracting from them some of their main dynamic features corresponding large structures embedded in turbulence.