1. TUSTP 2003 By Ciro A. Pérez May, 2003 DOE Project: HORIZONTAL PIPE SEPARATOR (HPS © )

2.  Objectives  Physical phenomena in HPS  Modeling approach  Experimental program  Conclusions - Future work Topics

3.  Study the behavior of oil-water mixtures in horizontal pipes  Develop a mechanistic model that predicts separation efficiency for given fluids, geometry and flow rates  Compare/refine model with data obtained in this study and from literature  Study effects of using manifolds to install multiple separators in parallel Objectives

4.  Objectives  Physical phenomena in HPS  Modeling approach  Experimental program  Conclusions - Future Work Topics

5. Zone 1Zone 2 Zone 3 Zone 4 Oil Oil with Water droplets Packed water droplets in oil Packed oil droplets in water Water with Oil droplets Water Physical phenomena in HPS Inlet Direction of flow Outlets

6.  Oil-Water mixture enters HPS, with droplet distribution function of processes upstream. Some mixing can occur at inlet (Zone 1)  Inside HPS the velocity decreases, turbulence decreases (laminar flow might be reached), settling and coalescence are promoted (Zone 2), layers begin to develop  Up to 6 layers can develop (Zone 3): -Pure Oil -Oil with water droplets -Packed water droplets in oil -Packed oil droplets in water -Water with oil droplets -Pure water  Eventually steady state is reached (Zone 4) Physical phenomena in HPS

7.  Regimes of operation in HPS  Laminar flow is desirable as it promotes segregation  Oil is more likely to flow to be in laminar flow conditions due to higher viscosity  So, desirable flow regimes are: -Laminar Oil Flow - Laminar Water Flow -Laminar Oil Flow - Turbulent Water Flow  Study flow in HPS requires: -Steady state conditions: max segregation -Transient conditions: how long it will take Physical phenomena in HPS

8.  Objectives  Physical phenomena in HPS  Modeling approach  Experimental program  Conclusions - Future work Topics

9.  Previous studies  Proposed model Modeling approach

10. Previous studies a.1D Mechanistic approach: Barnea-Brauner (1991) b.2D Analytical approach (for laminar flows): Brauner (1998) c.Numerical approach -Shoham-Taitel (1984, gas-liquid) -Elseth et al. (2000, VOF method) -Gao et al. (2003, VOF method) 1D mechanistic approach leads to simple solutions, so it will be used as an initial approach Modeling approach

11.  Proposed model:  1- 1D stratified flow pattern model is applied for given fluids and flow rates. If flow is stable, flow characteristics are given by the model  2- If flow is unstable, following procedure applies: -An amount of more viscous phase is assumed to flow to the less viscous phase -For this new flow rate, properties are calculated for mixture, segregated flow is assumed, and stability is checked. Migration stops when stability is reached -No convergence means non segregated flow Modeling approach

12.  Preliminary results Model tested against experimental data (Shi et. Al (2000))  Test conditions: -Oil properties: 3 cp, 800 kg/m 3 -Water properties: 1 cp, 1100 kg/m 3 -Pipe: 0.1 m ID, 18m long -Mixture velocity: 0.4 to 3 m/s -Water Cut: 0.2, 0.4, 0.6, 0.8  Trallero (1995) model used, Sheltering Factor assumed 0  Increased interfacial friction factor as mixing and waves form at the interface Modeling approach

13.  Results: Pure oil and water layer thickness Modeling approach

14.  Objectives  Physical phenomena in HPS  Modeling approach  Experimental program  Conclusions - Future work Topics

15. Test Section Experimental program

16.  Calibration:  Level: Pipe centerline was leveled in +-3/32” range from the horizontal  Level sensors: For operating conditions, level meters are able to detect continuous interface with error of 3/32” Experimental program

17.  Typical level meter signal at interface Experimental program Sensor stem gap: 7/32”

18.  Pitot / Isokinetic sampling probe Previous works: -Khor, Mendes-Tatsis and Hewitt (1996) -Vedapuri, Bessette and Jepson (1997) -Shi, Cai and Jepson (1999) -Cai, Gopal and Jepson (2000) Experimental program

19.  Pitot / Isokinetic sampling probe  Characteristics -ID= 3/16” -OD= 11/32” -Operating dP: 0 to 1” H 2 O, accuracy dP 0.15% -Range of operation:. Min. velocity: 0.06 m/s (error 10% ). Max. velocity : 0.7 m/s (error 0.073% ) Experimental program

20.  Photo of assembled probe: Base Pitot Pressure outlets Sampling outlet Experimental program

21.  Pitot / Isokinetic sampling probe in place Experimental program

22.  Pitot / Isokinetic sampling probe - Calibration results for single phase Experimental program

23.  Pitot / Isokinetic sampling probe -Problems when measuring oil-water flow. After flushing with oil, water floods pitot, capillarity causes oscillations in dP while flooding -Improved with wider pressure taps. dP values to be taken at initial plateau, before flooding occurs. Experimental program Plateau Flooding

24.  Calibration results: Effects of oil-water flow  Pitot filled with oil, mixture flowing Vsl=0.6 m/s, WC 60% Experimental program

25.  Objectives  Physical Phenomena in HPS  Modeling approach  Experimental program  Conclusions/ Future Work Topics

26.  Initial model for all flow conditions is proposed. Actual model underpredicts thickness of pure fluid zones  Model requires higher interfacial shear stress when mixing layers are present  Pitot measurements for low velocities are affected by capillarity in pitot pressure taps Measurement criterion was adapted for this condition Conclusions/Future Work

27.  Measurement of velocity profiles for experimental matrix  Measurement of hold up for experimental matrix  Hold up/Interfacial friction factor adjustment with experimental data and literature data Future work

28.  Questions? HORIZONTAL PIPE SEPARATOR (HPS©)