Local flow regime identification for boiling two-phase flow by BP neural networks approach

The pressure fluctuation signals of the boiling two-phase flow in a upward tube were analyzed by statistical and fractal theory. Five parameters, i. e. the heat flux, standard deviation, Hurst index, correlation dimension and Kolmogorov entropy were obtained and used as the characteristic vector of BP neural network. Results show that the flow regime characteristic vector which was obtained by statistical and fractal parameters could reflect the difference between various flow regimes. The method has the merits such as easy computation and easily quantifying the characteristics of the measured signals.