FINDING CHAOS IN EXPERIMENTAL TIME-SERIES - THE CASE OF 2-PHASE FLOW

This paper focuses on a method proposed for diagnosing the presence of chaotic behaviour in physical systems directly from experimental time series. The procedure relies on the concept of short-term predictabil ity of chaotic motions. From a statistical point of view, it is comple tely non-parametric and works whatever the distribution function of th e data points may be. The method is robust to observational noise, and can distinguish merely linear stochastic processes, e.g., fractional Brownian motion, from truly non-linear deterministic systems. The capa bility of the method has been evaluated by applying it to intermittent gas-liquid flows in a horizontal pipe. A weak sign of deterministic c haos has been diagnosed in plug and slug flow regimes. Our analysis fu lly confirms the Ruelle-Takens view about the complexity in hydrodynam ical systems. However, the weak sign of chaos is in contrast with the Lorenz type systems (strong chaos) and supports the idea of Kolmogorov about irregular motion in hydrodynamical systems.