EPSILON ENTROPY FOR A TIME-SERIES OF THERMAL TURBULENCE

We consider a dynamic entropy of a Rayleigh-Benard flow which varies with the characteristic scales epsilon of the process. Such an epsilon entropy is used to distinguish and classify the different dynamic regimes recently observed in this system: oscillations, chaos, and soft, hard, and developed turbulences. Moreover, data analysis of an experimental time series reveals a scaling behavior of the epsilon entropy in a hard turbulent regime, demonstrating that this notion of entropy is suitable to describe dynamic processes with a hierarchy of characteristic instabilities.