The Gas-Liquid Cylindrical Cyclone (GLCC) separator is commonly used for the separation of oil and gas mixtures flowing from the well head. Similar to the design used by other separators, it has an inlet and two outlets for gas and liquid respectively. However, the inlet to the separator can either be single or dual type. The pipeline connection from the upstream preconditioning equipment (CFC) is inclined downwards and has a tangential inlet slot. The essence of having a downward inclination is to promote pre-separation of the fluid phases. On the other hand, a tangential inlet promotes circular fluid motion thereby inducing separation of the fluid phases by centrifugal forces. Due to the complex behaviour of the flow within the GLCC, liquid carry over (LCO) as drops into the gas phase pipeline and gas carry under (GCU) as bubbles into the liquid phase pipeline are inevitable. Both phenomena greatly reduce the purity of the fluid phases at the outlets. To overcome this challenge, it has been proposed from field experiments carried out by Chevron Energy Technology Company, to precondition the influent flow in an upstream vertical pipe before entrance to the GLCC. In order words, a suggested solution to overcoming liquid carry over (LCO) and gas carry under (GCU) is to precondition the oil/gas mixture by forcing small bubbles/drops of 3 - 5mm in diameter to coalesce in an upstream vertical pipe. The upstream vertical preconditioner is known as a Churn Flow Coalescer (CFC). This is because the churn flow regime is the most suitable for the coalescence of both liquid and gas phases. Therefore, it is in the scope of this research work to carry out detailed preconditioning experiments within an upstream vertical pipe that serves the purpose of a Churn Flow Coalescer (CFC). All experiments in this research work have been carried out in the Chemical and Environmental Engineering L3 laboratory at the University of Nottingham. Although, the churn flow regime is specifically the most suitable for the GLCC, the operational envelope for the initial set of experiments spans the bubble to churn regimes. This is because the experiments were performed with the aim of delineating the conditions for the inception of typical churn flow in a large diameter pipe. These set of experiments were conducted in a 121mm internal diameter, 5.3m in length vertical pipe using air and water as the operating fluids. In these experiments, slug flow characterised by a Taylor bubble and a liquid slug was not observed. The churn flow regime is made up of two sub-regimes namely: liquid bridging of the gas core and formation of huge waves. The former is a phenomenon that occurs when the liquid phase forms a bridge as a result of radial coalescence of the wave crests flowing about the pipe centreline and momentarily blocks the entire pipe cross-section. The huge waves occur when the liquid phase flows as waves on the inner walls of the pipe and about the pipe centreline having large amplitudes. Between bubble to churn flow regimes in these experiments, four regimes were observed namely, discrete bubbly flow and spherical cap bubbly flow which make up the bubbly flow regime, churn turbulent regime (transition region) and typical churn flow regime. These experiments paved way for detailed experiments to be carried within the churn flow regime. Detailed churn flow experiments were then carried out in a large scale closed loop facility having an internal diameter of 127mm and a longer vertical pipe of 11m. The rationale for performing the experiments in this facility is because the facility offered a wider range of conditions within the churn flow compared to the first experimental rig facility. Data was acquired at L/D = 2.4, 7.1, 30.7, 35.4 and 82.7 which represent different axial distances from the gas-liquid injection at the base of the test section. Air and water were also used as the operating fluids. The void fraction data acquired at different axial distance from the injection varies logarithmically with increase in axial distance. The flow can be considered to be developed at L/D = 82.7 based on the void fraction. In addition, the frequency of liquid bridging of the gas core decays with increasing distance from the injection (downstream) while the frequency of the huge waves and liquid structures entrained in the gas core increases downstream from the injection due to coalescence. Finally, the effect of viscosity in the churn flow regime was investigated using air-glycerol/ water as the fluid pairs in the same large scale loop facility. Two glycerol/water mixtures were used having viscosities of 12.2cP and 16.2cP respectively. The data was acquired at a suitable axial distance from the injection at L/D = 65.5. In this experimental campaign, the size and frequency of the liquid structures entrained in the gas core are larger compared to the liquid structures present when experiments were carried out using air-water as the operating fluids. As a result, this gives a bi-modal probability distribution for air-glycerol/ water compared to air-water. Similar to the air-water experiments, the liquid bridging operating condition gives a high degree of coalescence of both phases. The mechanism of entrained liquid structure formation has been proposed based on the comparative study to the air-water experiments. A model has also been developed that predicts the effective length of pipe for the Churn Flow Coalescer (CFC). Overall, the liquid bridging sub-regime of churn flow should be the prevailing condition in the CFC to enhance proper separation of gas and liquid phases in the downstream GLCC separator.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:605808 |
| Date | January 2013 |
| Creators | Kanu, Aimé Uzochukwu |
| Publisher | University of Nottingham |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://eprints.nottingham.ac.uk/27654/ |
Page generated in 0.0028 seconds