Identification of the flow pattern from the experimental pressure signal in horizontal pipes carrying two-phase flows

Two-phase flow is common in many industrial applications, especially in oil-gas production and transportation. Identifying the configuration of the phases along the flow is extremely important since the way they are distributed dictates the behaviour of the gradient of pressure and quantities such as the heat transfer coefficients of the mixture, while also being relevant in outflow guarantee issues. This work proposes an approach for identifying the pattern of two-phase flow in horizontal pipes through noninvasive pressure measurements. The flow patterns are identified based on dimensionless parameters from a nonlinear analysis of time series, namely the correlation dimension, Lyapunov and Hurst exponents. Dynamic pressure measurements from piezoelectric sensors are obtained from a horizontal line of two-phase liquid-gas flow. The experimental bench uses tap water and compressed air as working fluid and is capable of generating different flow patterns classified as stratified-smooth, stratified-wavy, elongated-bubble, slug and dispersed-bubbles with different surface velocities of the gas and liquid phases. The dynamics of the 23 experimental points are investigated and the results indicate that the system response is chaotic for all flow patterns due to the positive sign of the Lyapunov exponent. From the Hurst exponent, it was possible to separate the oscillatory flows patterns, as slug patterns, from non-oscillatory flows patterns, as stratified. The correlation dimension split the smooth and wavy stratified patterns from the intermittent and dispersed bubbles ones. Subsequently, it shows that these parameters form a multi-dimensional map that cluster the experimental points according to the dynamics of each flow pattern. The proposed approach gives flow pattern identification based on the flow nonlinear dynamics from a single nonintrusive pressure measurement. This result opens new and exciting possibilities for applications in complex industrial systems.