Computational and Experimental Investigation of Oil Separation from an Oil-Water Two-Phase Pipe Flow and Its Accumulation in a Subsea Dead-Leg
Type of DegreeMaster's Thesis
MetadataShow full item record
Both 2- and 3-dimensional transient computational studies of two-phase turbulent flow within a horizontal channel/pipe with a vertical capped dead-leg placed downstream of an oil/water supply source simulating parts of a model subsea petroleum production system were considered. Two-phase continuity and momentum relations governing two immiscible, incompressible, Newtonian fluids were solved using the open-source software OpenFOAM. Baseline 2-D simulation results for a 90%-10% slurry of 32.8 oAPI crude oil-water mixture at the inlet port (Reynolds number of 2.1x105) were studied to elucidate the extent of evolving recirculating vortices and penetration of the oil phase into the vertical dead-leg. For the 2-D channel simulations, parametric studies of the average inlet velocity, width/length of the dead-leg and oil density/viscosity on evolving recirculation zones and accumulation of oil within a vertical dead-leg were performed to determine their effects on development of the transient turbulent flow, pressure, concentrations fields and the extent of oil separation and its accumulation in the vertical dead-leg. Computationally-intensive 3-D unsteady turbulent simulations within a vertical dead-leg for a 90%-10% oil-water mixture were also performed. The general trends of oil separating from the mixture and lifting in the vertical dead-leg as time progressed was captured by both 2-D and 3-D models. However, the 3-D model exhibits greater refinement of the complex three-dimensional vortical effects at the T-junction linked to separation of the two phases leading to migration of lighter oil to the top of the vertical dead-leg. An experimental flow loop using deionized water/olive oil mixtures treated with table salt as demulsifier was designed to test similar phenomenon. Oil accumulations tests were performed with 5% and 10% oil volume mixtures for dead-leg inclination angles of 90 (vertical), 60, 45 and 30 degrees. Following a system homogenization procedure, minimum of four oil separation/accumulation tests were performed for a minimum time period of 10 min. Videos of oil accumulation process were obtained for various inclination angles of the dead-leg. The instantaneous accumulated heights of the oil column versus time were tabulated. Through applying the least-squares linear fit method, empirical accumulation rates were derived. Variations of the instantaneous height of the oil column within the dead-legs were observed rise linearly for the 5% oil mixture, regardless of the inclination angle. For the 10% oil mixture, the instantaneous height of the accumulated oil column was rising linearly faster than the 5% oil mixture at the early instants, however for some inclination angles the accumulation rate declined. For the 10% oil mixture, limited 3-D simulation results for a period of 1-2 min matching the operating conditions in the laboratory were compared to experimental findings for dead-leg inclination angles of 90 (vertical), 60, 45 and 30 degrees. The complex recirculating flow zone in the vicinity of the T-junction was identified to be the major source of oil separation and its rise toward the capped end of the dead-leg. Another route was also identified where oil droplets were separated at the bottom of the complex recirculating zone and while coalescing lifted toward the expanding accumulating oil column. Limited experiments were performed for the case of a horizontal dead-leg. Experimentally-obtained oil accumulation data suggest that oil accumulation was limited to the top of the horizontal dead-leg and the zone of accumulated oil reached an asymptotic value early on.