Flow regime identification of gas/liquid two-phase flow in vertical pipe using RBF neural networks

Abstract

The gamma ray scattering energy spectrum detected by one detector was presented to distinguish the gas liquid two-phase flow regime of vertical pipe. The simulation geometries of the gamma ray scattering measurement were built using Monte Carlo software Geant4. Computer simulations were carried out with homogeneous flow, annular flow and slug flow. The results show that the scattering energy characters of homogeneous flow and annular flow have significantly different. The scattering spectrum of slug flow is similar to annular flow for long gas slugs and similar to homogeneous flow for short gas slugs. The RBF neural networks were used to predict the flow regime. The results show that the homogeneous flow and annular flow can be completely distinguished and most of the slug flows were recognized by the neural network. It was demonstrated that the method of one detector scattering energy spectrum has the ability to identify the typical gas liquid flow regime of vertical pipe and fit the applications in engineering.