Noise analysis using chaos theory

Chaos theory finds application in noise signal analysis in systems of information acquisition with diagnosis value. In the present article it is examined the theoretical foundation of noise discrimination problem from a fractal dynamics point of view. It is examined the analytical procedure that makes it possible to determine the embedding space dimension exported from a temporary series. When chaos is present, a strange attractor with fractal dimension is obtained. In this way it is possible to discriminate a random signal from a chaos determinist one. The analytical procedure for this technique is validated by mathematical modelling of the Hennon map and the Rossler system. Also, this technique is applied to a neutronic temporary series derived from vibration model- experiment carried out in a subcritical natural uranium-water assembly. (author). 13 refs, 7 figs