464244 A Simplified Dynamic Model of a Continuous Subsea Oil-Water Gravity Separator for Estimation of Unmeasurable Oil and Water Purities

Wednesday, November 16, 2016: 5:21 PM
Carmel II (Hotel Nikko San Francisco)
Tamal Das, Department of Chemical Engineering, Norwegian University of Science and Technology, Trondheim, Norway and Johannes Jäschke, Chemical Engineering, Norwegian University of Science and Technology, Trondheim, Norway

Gravity separation is a widely used separation process in industry for separation of liquids based on difference in densities. In oil and gas industry, gravity separators are often used for subsea oil-water separation. However, a strict control over the quality of the outlet streams does not exist because subsea sensors for sufficiently accurate measurement of oil in water and water in oil concentrations are not available. Hence, there is a need for dynamic models for gravity separator that can predict the concentration of outgoing streams sufficiently accurately.

A common gravity separator design is a horizontally oriented cylindrical vessel, in which mixed fluids enter from one end and the separated fluids are removed from the other. Inside the separator, the fluids separate into layers. The layers from bottom to top are water layer, emulsion layer and oil layer. The emulsion layer consists of two sections: sedimentation layer and dense-packed layer. At the end of the separator, a weir prevents the emulsion layer and the water layer from flowing over the weir. The liquid levels inside the separator are controlled in such a way that only oil flows over the weir to the oil outlet section, while the water is removed continuously from the bottom of the separator.

In previous years, the modeling of gravity separation has focused on studying mechanisms that govern batch separation of two liquids, such as oil and water. The models for gravity separation presented in the literature have been summarized by Frising [1]. The models can be classified as sedimentation-based models and coalescence-based models. Sedimentation based models [2-6] consider sedimentation of droplets and interfacial coalescence between droplets and interface. These models are typically tuned to match the results from the batch experiments that demonstrate four layers: clear oil, sedimentation zone, dense-packed zone and clean water. On the other hand, coalescence based models [7,8] focus on the dense-packed zone considering binary as well as interfacial coalescence of the droplets of dispersed phase. They predict mean droplet size based on lifetime of a single film and assume that the presence of surfactant only scales the coalescence time. A model by Henschke [8], which is a coalescence-based model, is, perhaps, the most relevant for modeling continuous gravity separation because it predicts drop size in the dense-packed zone using a single coalescence parameter, which is independent of experimental equipment, mixing intensity and volume fraction of dispersed phase.

Some of these mechanisms of oil-water separation have been used for modeling continuous gravity separation. Sayda et al.[9] developed a simplified dynamic model for a continuously operating gravity separator based on material balances, and phase equilibrium equations. Their model takes into consideration simplified hydrodynamics, thermodynamics and spatial evolution of separation profiles. They used a standard form of Stoke's law for vertical settling velocity of droplets based on a constant average droplet diameter of oil droplets in water. Hence, their model does not consider coalescence effects that are typically very important in oil-water separation. Another attempt to model a continuous oil-water separation was conducted in [10], in which they model a horizontal pipe separator. They combine some of the principles from sedimentation based models [11] and Henschke’s model [8]. Their approach is quite useful for modeling a gravity separator because horizontal pipe separator is quite similar to the conventional gravity separator. However, they model a pipe separator that is completely filled with fluids and does not have a weir. On the other hand, gravity separators often include a weir and are operated with the separator only partially filled with liquids.

In this work, we present a dynamic model for a continuous cylindrical gravity separator for oil-water separation to predict the purity of outlet streams based on operating conditions by assuming that there exists three layers of liquids: oil, emulsion and water inside the separator. It is based on lumped total mass balances and partial mass balances for each layer, in which the inter-layer mass transfer equations are derived using principles from the previous work in sedimentation based models and Henschke’s model.

The model is solved in MATLAB using solver ode15s. It captures the effect of total inlet flow rate, inlet oil cut and inlet droplet size on purities of outlet streams and thickness of each layer of liquid. We use the model in a Kalman-based non-linear estimator for estimating unmeasurable variables, such as oil and water purities using level and density measurements and other measurements of inlet conditions, such as inlet flow rate, oil cut, droplet size and water outflow rate. Further, we study the effects of the variations to water level set point on the performance of the separator.

[1] Frising, T., Noïk, C. and Dalmazzone, C., 2006. The liquid/liquid sedimentation process: from droplet coalescence to technologically enhanced water/oil emulsion gravity separators: a review.Journal of Dispersion Science and Technology,27(7), pp.1035-1057.

[2] Jeelani, S.A.K. and Hartland, S., 1986. Prediction of dispersion height in liquid-liquid gravity settlers from batch settling data. Chemical engineering research & design64(6), pp.450-460.

[3] Jeelani, S.A.K., Pandit, A. and Hartland, S., 1990. Factors affecting the decay of batch liquid‐liquid dispersions. The Canadian Journal of Chemical Engineering68(6), pp.924-931.

[4] Jeelani, S.A.K. and Hartland, S., 1993. The continuous separation of liquid/liquid dispersions.Chemical engineering science48(2), pp.239-254.

[5] Jeelani, S.A.K., Panoussopoulos, K. and Hartland, S., 1999. Effect of turbulence on the separation of liquid-liquid dispersions in batch settlers of different geometries. Industrial & engineering chemistry research38(2), pp.493-501.

[6] Jeelani, S.A.K., Hosig, R. and Windhab, E.J., 2005. Kinetics of low Reynolds number creaming and coalescence in droplet dispersions. AIChE journal51(1), pp.149-161.

[7] Lobo, L., Ivanov, I. and Wasan, D., 1993. Dispersion coalescence: Kinetic stability of creamed dispersions. AIChE journal39(2), pp.322-334.

[8] Henschke, M., Schlieper, L.H. and Pfennig, A., 2002. Determination of a coalescence parameter from batch-settling experiments. Chemical Engineering Journal85(2), pp.369-378.

[9] Sayda, A.F. and Taylor, J.H., 2007, July. Modeling and Control of Three-Phase Gravilty Separators in Oil Production Facilities. In American Control Conference, 2007. ACC'07 (pp. 4847-4853). IEEE.

[10] Pereyra, E., Mohan, R.S. and Shoham, O., 2013. A Simplified Mechanistic Model for an Oil/Water Horizontal Pipe Separator. Oil and Gas Facilities,2(03), pp.40-46.

[11] Hartland, S. and Jeelani, S.A.K., 1988. Prediction of sedimentation and coalescence profiles in a decaying batch dispersion. Chemical engineering science43(9), pp.2421-2429.


Extended Abstract: File Not Uploaded
See more of this Session: Estimation and Control of Uncertain Systems
See more of this Group/Topical: Computing and Systems Technology Division