Droplet size and velocity in dual continuous horizontal oil���water flows Talal Al-Wahaibi a,*, Panagiota Angeli b a Department of Petroleum and Chemical Engineering, Sultan Qaboos University, P.O. Box 33, Al-Khoud, P.C. 123, Oman b Department of Chemical Engineering, University College London, Torrington Place, London WC1E 7JE, UK 1. Introduction Two-phase liquid���liquid systems are common in the oil and chemical industries and they often take the form of dispersion of one phase into the other. Knowledge of drop size and drop size distribution are important factors for determine the rheology and stability of a dispersion. They also have a very significant effect upon equipment performance as in liquid��� liquid contactors or phase separators for example. Several studies have been carried out to model the size and the distribution of the dispersed phase, mainly in stirred vessels (for a comprehensive review see Zhou and Kresta, 1998). In contrast, there is very limited amount of data on pipeline flow particularly of unstable liquid���liquid dispersions, because of the complex nature of the experiments involved. In such systems drop break-up and coalescence will determine the final drop size distribution. These two phenomena have been studied separately and models have been suggested in a number of studies (for example Hinze, 1955 Shinnar, 1961 Sleicher, 1962 Kubie and Gardner, 1977 Thomas, 1981), while in a few studies the phenomena were combined using drop population balance equations (Valentas et al., 1966 Tsouris and Tavlarides, 1994 Hu, 2006). Angeli (1996) and Hu (2006) presented extensive summary of the drop break-up and coalescence literature on dispersed liquid���liquid systems. Experimentally, a number of investigators have presented data on drop size distributions in turbulent pipe flow of unstable dispersions mainly in fully dispersed dilute flows (where one phase is complete dispersed into the other). A comprehensive summary of that can be found in Lovick and Angeli (2004b). Recently, Angeli and Hewitt (2000) used an endoscope located inside the pipe to measure drop size distribution in turbulent flow of two immiscible liquids (with dispersed phase volume fractions from 3.4 to 9%). Experimental results, obtained in fully dispersed flow revealed that dmax decreased c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 8 3 ��� 9 3 a r t i c l e i n f o Article history: Received 15 November 2006 Accepted 3 October 2007 Keywords: Liquid���liquid Horizontal flow Dual continuous Chord length Drop size Drop velocity a b s t r a c t The size and vertical distribution of drops during horizontal dual continuous oil���water flow were studied experimentally. The investigations were carried out in a 38 mm ID stainless steel test section with water and oil (density 828 kg/m3 and viscosity 5.5 mPa s) as test fluids. Drop velocities and sizes were obtained with a dual impedance probe which allowed measurements at different locations in a pipe cross-section. The measurements indicated chord lengths up to 20 mm in some cases. Drop concentration and chord length decreased with increasing distance from the oil���water interface. Also, oil drops were found to be larger than water drops since oil tends to lose its continuity at relatively low volume fractions compared to water. The number density of large drops was found to decrease as the water superficial velocities increased while there was no clear effect of oil superficial velocities on drop size. Water drops were in general faster than the velocity of the upper, oil continuous, layer while oil drops could be either faster or slower than the lower, water continuous, layer. There was no clear effect of the layer velocity on the size of drops dispersed in that layer. Finally, it was found that none of the available correlations on maximum drop size was able to predict the present experimental data. These correlations were developed for drop breakage in a turbulent flow field. # 2007 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. * Corresponding author. Tel.: +968 2414 1320 fax: +968 2414 2517. E-mail address: alwahaib@squ.edu.om (T. Al-Wahaibi). available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/cherd 0263-8762/$ ��� see front matter # 2007 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.cherd.2007.10.012

as the continuous phase velocity increased, while water drops in oil appeared to have smaller average sizes compared to oil drops in water at the same dispersed phase volume fractions. A strong effect of the pipe wall material on drop size distribution was found which was attributed to differences in wall roughness and turbulence generated. Simmons et al. (2000) compared two different optical techniques, based on light backscattering (Par-Tec 300C) and diffraction (Malvern 2600) to investigate drop size distribution. Their results reflect a decrease in the Sauter mean drop diameter, d32, with increasing mixture velocity. Simmons and Azzopardi (2001) used the same laser techniques to study drop size distributions in both vertical upward and horizontal flows. The authors argued that the Rosin���Rammler distribu- tion is not physically sound since it does not have a mathematical upper-cut off which means that it has a tail to infinite drop sizes. For this reason, the upper limit log normal function was used. The upper limit log normal function was found to fit their experimental results reasonably well. The maximum drop diameter was found to agree with the theory of Hinze (1955) for dilute dispersions (up to 3%) but not for concentrated ones. Lovick (2004) designed a dual impedance probe that can measure chord lengths during oil���water flow. Using this probe, Lovick and Angeli (2004b) obtained chord length distributions in horizontal oil���water flows along a vertical pipe diameter. Apart from the fully dispersed regime, they also studied the dual continuous regime, where both phases retain their continuity at the top and bottom of the pipe, respectively, but there is drop entrainment and dispersion of one phase into the other. The experimental results revealed that drop size decreased with increasing distance from the interface. Also, a slight effect of layer velocity on the maximum and median chord lengths was found. In all cases studied, the velocity of the water drops in the upper oil continuous layer was found to be higher than the average velocity of the layer while that of the oil drops in the lower water continuous layer could be either higher or lower than the average layer velocity. None of the available correlations on maximum drop size was able to predict their experimental data. Using the same probe Hu and Angeli (2006) studied chord length distributions in downward and upward vertical flows. Since phase inversion appeared at high dispersed oil volume fraction the water continuous mixtures were more concen- trated than the oil ones. Larger drops were found in the pipe central region for both downward and upward flows however the distribution of the water drops in the less concentrated oil continuous dispersions was more uniform than that of the oil drops in the highly concentrated water continuous phase. Also, the measurements showed smaller water drops in oil compared to oil drops in water. The probe, similar to other local probes, would favour sampling of the large drops, which was taken into account in the development of a model for the transformation of the experimental chord length distributions to drop size distributions (see Hu et al., 2006). Using this model, Hu (2006) showed that d32 of the Sauter mean diameter, d32, of the oil drops increases significantly as the dispersed fraction of oil in water increases for both downward and upward flows. In this paper, drop velocity, chord length and drop size distribution were studied in horizontal liquid���liquid dual continuous flow. The dual continuous pattern has been reported under different names in a number of works in liquid���liquid flows (Guzhov et al., 1973 Cox, 1985 Scott, 1985 Trallero, 1995 Nadler �� and Mewes, 1997 Vedapuri et al., 1997 Angeli and Hewitt, 2000 Lum et al., 2006 for a comprehensive review see Lovick and Angeli, 2004a). Despite its common occurrence in liquid���liquid flows, however, the only other drop size data available in that flow pattern is by Lovick and Angeli (2004b). Compared to that study, in the present work a modified test section inlet has been used which ensures that the two fluids are introduced with minimum mixing and that all drops are formed from the evolving interfacial waves (see Al-Wahaibi et al., 2007 for the mechanism of formation through interfacial waves). In addition, wider range of flow conditions (starting from mixture velocity of 0.95 to 2.5 m/s) have been studied focusing only on dual continuous flow while in the work of Lovick and Angeli (2004b) the focus was on both dispersed and dual continuous flows. 2. Experimental set up The experimental studies on dispersed drop velocities and chord length distributions were carried out in the flow facility shown in Fig. 1 that has a 38 mm ID stainless steel test pipe (for a detailed description of the facility used in this work see Lovick and Angeli, 2004a). The test pipe consists of two 8 m long sections joined by a U-bend. The inlet is a Y-junction through which the two fluids enter in a stratified manner. Experiments were carried out at the first section of the test pipe. Oil and water were used as test fluids with properties shown in Table 1. A dual impedance probe designed by Lovick (2004), located at 7 m from the inlet was used for drop velocity and size distribution measurements. The dual impedance probe con- sists of two impedance sensors working independently. Each impedance sensor is made of two coaxial conductors that are separated by an insulator the whole sensor is enclosed in an insulator which leaves only a sensitive tip exposed to the flow. The sensors measures the impedance of the phase that surrounds the sensitive tip. The output signal of each sensor is a time series of the impedance values of the fluid at the point of measurement. High signal values are obtained when the tip is surrounded by water and low values when it is surrounded by oil. High signal values in an oil continuous dispersion will therefore indicate the passage of water drops, while low signal values in a water continuous dispersion will indicate the passage of oil drops. The two sensors are 10 mm apart, placed one behind the other in the flow direction. A specially designed spool allows the two sensors to move together along the same diameter in a pipe cross-section. To each probe, an alternating current with frequency between 2 and 45 kHz can be applied via a signal controlling box. For each measuring location the output signals from the two sensors are logged in a computer over a period of 3.4 s by a FORTRAN program. The signals from the two sensors are then cross-correlated to obtain the average time the drops need to travel from one sensor to the other at this location. Combining this time with the known distance between the sensors, the average drop velocity can then be calculated. When the drop Table 1 ��� Properties of the test fluids Properties Oil Water Density 828 kg/m3 1000 kg/m3 Viscosity 5.5 mPa s at 25 8C 1.0 mPa s at 25 8C Interfacial tension 39.6 mN/m at 25 8C c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 8 3 ��� 9 3 84