| dc.description.abstract | Water separation in an upward, inclined separation tube, with distributed tapping points has been studied, verified and modeled in this thesis. Phase segregation in an upward inclined pipe is widely occurring in oil and gas production flow systems, but using the phenomenon for a controlled phase separation is somewhat novel and requires the establishment of understanding of the governing parameters. The thesis research consists of two major parts;
1. Conducting laboratory experiments to validate the separation concept, to quantify its efficiency and to acquire data to develop the separation model.
2. Modeling the separation process analytically to establish the governing parameters and their significance.
The experimental separation observations have been conducted on an inclined test tube with distributed tapping points. The test tube is 5m long and is 0.154m in diameter, a size which minimizes diameter scaling-up to field applications scale. The tests were conducted at a low pressure with a mixture of two-phase, water and synthetic oil. In addition, certain experiments were conducted with low volumes of compressed air, simulating the small quantities of free gas flowing in and near the well. Variations of full scale separator length were studied by conducting an extensive matrix of test conditions on the short 5m-tube. These tests results can be combined in a ”pancake” fashion to model the performance of a full-length separator.
The core of the modeling is the establishment of an empirical drainage function per tapping point that relates the oil and water flow rates in the tube to the oil fraction obtained in the tapping stream. Once a single tapping performance is known and expressed in terms of separation efficiency, the total separator efficiency with multiple tapping points is calculated. The drainage function is based on a physical realization that there is a region, or a ”drainage volume” in the flow stream affected by the fluid withdrawal in the tapping point. The oil fraction in the tapping stream depends on this drainage volume and this dependency can be expressed by two factors; (a) the height above the draining point and, (b) a relationship that expresses how much oil is drained from each position vertically above the tapping point. These two relationships, which were established from the experimental results, were then used to assign a weighted oil contribution from each height from the bottom of the tube. When these weighted contributions are employed on the oil dispersion model, it calculates the compound effect of the dispersion and withdrawal to yield the tapping stream oil fraction.
The dispersion model calculates the distribution of the fraction of each oil droplet-size class in the pipe cross section, and the distribution of the total oil fraction. The dispersion model is based on the balance of forces acting on oil droplets perpendicular to the flow direction. The applied forces are gravity, buoyancy, lift and turbulent dispersion. The model is unified. That is, it captures the oil distribution at all the observed flow patterns. The model has been used to study the effect of variation of the parameters that were kept constant in the experiments, such as water and oil density and interfacial tension and pipe diameter. In addition, the developed and tuned model has been tested for predicting separation efficiency in single and multiple tapping arrangements.
The studied separation variables are; inlet rates of the fluids, inclination angle and diameter of the separation pipe, oil and water densities, interfacial tension, number of tapping points and distance between the tapping points.
The results concluded that the flow pattern is the most significant factor affecting the separation performance. High separation efficiency is obtained for flow patterns with a water-rich bottom layer. This layer can be either upwards flowing, or a back-flowing film. The pipe inclination angle, the pipe diameter, water and oil densities and the interfacial tension are strongly affecting the flow patterns, and thus the separator performance. On the other hand, separation performance was not greatly affected by the drainage rate or the number of drainage points per section. Furthermore, as expected intuitively, the presence of air increases the turbulence and thus decreases the performance of the separation. However, it was also observed that in the presence of free gas, oil-water flow patterns with a thick, clear (transparent) and back-flowing water layer yielded better separation than the ones with upward flow films.
In summary, the thesis provides a basic understanding of the process of separation in inclined flow using single and multiple tapping, gives information on the range of expected performance, and provides some fundamental knowledge for design and engineering of separators based on this principle. | nb_NO |
