Publication: Compressing pressure data by a method based on neural networks

All || By Area || By Year

Title	Compressing pressure data by a method based on neural networks
Authors/Editors*	X. Gavalda1a, J. Ferrer-Gener1b, G.A. Kopp2, and Francesc Giralt1a
Where published*	To be submitted
How published*	Other
Year*	2008
Volume
Number
Pages
Publisher
Keywords
Link
Abstract	The amount of data produced by the experiments carried out in aerodynamic wind tunnels by civil and wind engineers are mostly pressure data over building and structures at high Reynolds number flow. This data are stored in large databases such as the NIST aerodynamic database for wind loads on low buildings, ready to be used by the engineers in their future calculations. This data are occupying a lot of space on the storage media so the reductions of their volume without sensitive lost in their properties is of the major interest. In this paper, a new compression method is presented based in Fuzzy ART Neural Network that preserves the statistical characteristics of data, the four first moments, and also engineering parameters like Lift, Drag and Power Density Spectra. An application of the method to experimental pressure data is presented. The major drawback is the CPU time consumed to deal with the best compressed data. The reduction of this time is being investigated.