On the role of buoyancy-driven instabilities in horizontal liquid–liquid flow

Citation data:

International Journal of Multiphase Flow, ISSN: 0301-9322, Vol: 89, Page: 123-135

Publication Year:
Usage 289
Abstract Views 289
Captures 16
Readers 16
Social Media 5
Tweets 5
Citations 5
Citation Indexes 5
Rhys G. Morgan; Roberto Ibarra; Ivan Zadrazil; Omar K. Matar; Geoffrey F. Hewitt; Christos N. Markides
Elsevier BV
Engineering; Physics and Astronomy; Chemical Engineering
Most Recent Tweet View All Tweets
article description
Horizontal flows of two initially stratified immiscible liquids with matched refractive indices, namely an aliphatic hydrocarbon oil (Exxsol D80) and an aqueous-glycerol solution, are investigated by combining two laser-based optical-diagnostic measurement techniques. Specifically, high-speed Planar Laser-Induced Fluorescence (PLIF) is used to provide spatiotemporally resolved phase information, while high-speed Particle Image and Tracking Velocimetry (PIV/PVT) are used to provide information on the velocity field in both phases. The two techniques are applied simultaneously in a vertical plane through the centreline of the investigated pipe flow, illuminated by a single laser-sheet in a time-resolved manner (at a frequency of 1–2 kHz depending on the flow condition). Optical distortions due to the curvature of the (transparent) circular tube test-section are corrected with the use of a graticule (target). The test section where the optical-diagnostic methods are applied is located 244 pipe-diameters downstream of the inlet section, in order to ensure a significant development length. The experimental campaign is explicitly designed to study the long-length development of immiscible liquid – liquid flows by introducing the heavier (aqueous) phase at the top of the channel and above the lighter (oil) phase that is introduced at the bottom, which corresponds to an unstably-stratified “inverted” inlet orientation in the opposite orientation to that in which the phases would naturally separate. The main focus is to evaluate the role of the subsequent interfacial instabilities on the resulting long-length flow patterns and characteristics, also by direct comparison to an existing liquid – liquid flow dataset generated in previous work, downstream of a “normal” inlet orientation in which the oil phase was introduced over the aqueous phase in a conventional stably-stratified inlet orientation. To the best knowledge of the authors this is the first time that detailed spatiotemporally resolved phase and velocity data have been generated by advanced measurement techniques in such experiments, specifically devoted to the study of long-length liquid – liquid flow development. In particular, the change in the inlet orientation imposes a Rayleigh–Taylor instability at the inlet. The effects of this instability are shown to persist along the tube, increasing the propensity for oil droplets to appear below the interface. Generally, the characteristics of the flows generated with the two inlet orientations are found to be comparable, although only six flow regimes are identified here, as opposed to eight for the original “normal” inlet orientation. The unobserved regimes are: (1) three-layer flow, and (2) aqueous-solution dispersion with an aqueous solution film. Furthermore, similar mean axial-velocity profiles are observed in the current study to those reported for the corresponding “normal” inlet orientation liquid – liquid flows. These findings are important to consider when interpreting published data from experiments performed in laboratory environments and attempting to draw conclusions relating to applications in the field. The generated data promote not only a qualitative, but importantly and uniquely, a quantitate understanding of the role of multiple instabilities on the development of these complex interfacial flows, with detailed insight into how the deviations manifest at distances 244 pipe-diameters downstream of the inlet (from high level information such as regime maps, to detailed flow information such as phase and velocity profiles). The data can be used directly for the development and validation of advanced multiphase flow models that require such detailed information.