doctoralThesis
Análise experimental e numérica da evolução dos parâmetros do escoamento bifásico gás-líquido em golfadas em uma tubulação horizontal
Fecha
2020-08-14Registro en:
RODRIGUES, Rômulo Luis de Paiva. Análise experimental e numérica da evolução dos parâmetros do escoamento bifásico gás-líquido em golfadas em uma tubulação horizontal. 2020. Tese (Doutorado em Engenharia Mecânica e de Materiais) - Universidade Tecnológica Federal do Paraná, Curitiba, 2020.
Autor
Rodrigues, Rômulo Luis de Paiva
Resumen
The two-phase slug flow pattern is characterized by alternating the liquid and gas phases. This flow is mostly found in pipelines of oil and gas production and transportation. The slug flow hydrodynamics is important in the design of facilities dealing with this flow pattern. The goal is to ensure an experimental and numerical analysis considering the influence of the flow evolution effect of water and air in a 25.8-mm ID pipe and 37.9-m long, in the experimental apparatus available in NUEM/UTFPR labs. Five resistive sensors are used to detect and track the typical structures of the slug flow. The experiments were made for different flow combinations of water and air to obtain after signal processing, the elongated bubble translational velocity, the unit cell frequency, the liquid slug and the elongated bubble lengths, the void fraction in the unit cell and the pressure drop. Numerical simulations were made using the slug tracking model developed in the NUEM-UTFPR labs. An evaluation of the inclusion of the wake effect as well the model singularities were made. These numerical simulations enabled comparison with the experimental data, and efforts were focused in trend detections and the behavior of dimensionless parameters, superficial fluid velocities and pipe diameter. The numerical and experimental results were analyzed using the kernel type probability density functions (PDF) to discuss the shape behavior considering the statistical moments of third and fourth orders (skewness and kurtosis). Goodness of fit tests (maximum likelihood and Kolmogorov-Smirnov) enabled inferences on distribution types and comparisons on the numerical and experimental tests.