Stagnant Liquid Column Research Articles

Numerical investigation on influences of injection flow rate on bubbling flow at submerged orifices

In this study, we numerically investigated the formation of bubbles by gas injection through an orifice in a stagnant liquid column and the flow field surrounding the bubbles after detachment under several flow conditions. To this end, we employed an enhanced coupling method that combines the advantages of mass conservation property of the volume-of-fluid (VOF) method with the boundedness and continuity of a conservative level-set (LS) function for multiphase computations to achieve a smoothened and sharpened interface. The coupling method was integrated into an incompressible Navier–Stokes solver to investigate the bubbling flow under various injection conditions. The simulated results were first validated against the experimental data of a benchmark test case to assess its reliability. The capability of the solver was also confirmed by examining the computed results and comparing them with those of the other studies. Subsequently, we simulated bubbling flow under different injected gas conditions and several orifice radii. The simulated results indicated that at low gas injection flow rates, the detachment time of bubbles, pinched off the orifice rim, could be predicted depending on the orifice radius and volume flow rate of the gas, whereas orifice radius had a notable influence on the detached volume of bubbles. In addition, the gas injection flow rate had a significant influence on the pressure field surrounding the bubbles after detachment, which is useful for studying of the bubbling flows and for the optimal design of small-scale systems.

Different nanofluids effect on bubble characteristics at the isothermal bubble column

AbstractIn this study, the effect of different nanofluids such as SiO2, Fe3O4, and Al2O3 on bubble characteristics is studied experimentally. The nanoparticles concentration for the SiO2 nanofluid is 0.05 wt% and for other nanofluids is 0.005 wt%. Bubbles are formed by injection of air at a constant gas flow rate (between 600‐1200 mL/h) into a stagnant isothermal liquid column. Experimental data of formation, growth, and detachment of the air bubbles were recorded by a high‐speed digital camera (1200 fps), and the image processing method was used to analyze the bubble characteristics. In the present study, bubble characteristics such as diameter, size, aspect ratio, and detachment frequency were studied in four different liquids. The results show that a bubble has the biggest size in the pure water and adding nanoparticles to the pure water decreases bubble size. Also, the variation of detachment frequency has an inverse relation with the bubble size behaviour. Between different nanoparticles, Fe3O4 has maximum and SiO2 has minimum effect on bubble features. By adding Fe3O4 to pure water, bubble diameter decreases nearly 7%‐8% and bubble detachment frequency increases nearly 15%‐20%. In order to analyze the variation of the bubble characteristics, a force modelling on bubbles during the growth was proposed. Moreover, a correlation has been proposed for the prediction of bubble diameter at detachment time in three different nanofluids and in the pure water. The mean absolute error of this correlation is 2.65% and has a good agreement with experimental results.

Experimental investigation of bubble growth and detachment in stagnant liquid column using image – based analysis

An experimental study has been carried out to characterize bubble formation, growth, and detachment mechanisms in a stagnant liquid column. Both bubble frequency and bubble detachment size were measured in different gas flow rates, injector diameters and orientations, submergence height, and liquid properties. Experiments were performed for air injection flow rate ranges between 200 mlph and 1200 mlph using needle diameters of 1.6, 1.19, 1.07, and 0.84 mm submerged in liquids with viscosities of 0.001, 0.1, 0.35, and 1 Pa.s. The data for bubble formation was obtained using a high-speed imaging technique. The results show that the bubble diameter at the departure increases as the needle diameter, liquid viscosity, and gas flow rate increase. In addition, the decrease in the submergence height results in a larger bubble at the departure. In order to analyze the changes in bubble detachment characteristics, a force modelling on a growing bubble was proposed. The experimental data were utilized for training a feed-forward back propagation neural network system to estimate the bubble detachment diameter. They were also used to propose a correlation to predict bubble diameter at the departure. The proposed correlation is found to be in the range of ± 8% of the obtained experimental data.

Mass Transfer Enhancement in GLS Fluidized Beds using Nanomaterials

Nanomaterial has unique physical property which made it important for many applications and that is why the use of nanomaterials rapidly increasing in the field of science and engineering.1. This work focuses on mass transfer of solids into liquid in three phase fluidized beds in presence of nanomaterial. This include the study of effect of gas velocity, time and different concentration of nanomaterials on mass transfer coefficient in stagnant liquid column in three phase fluidized bed system. To measure coefficient of the mass transfer, known quantity of solid pellets ie benzoic acid and known amount of nanomaterial fraction ie Arachitol nano were charged in the test column of three phase fluidized bed system. At the beginning of each run, test section was partially filled with water which prevent breakage of particles. The experiments were conducted by sequentially varying gas velocity for different volumes of nanomaterial and measuring the rate of mass transfer by collecting samples directly from the outlet ports at the top subsequently analysed by volumetric titration method. The results show enhancement in mass transfer coefficient by addition of nanomaterials. Arachitol nano has been taken in different volumes ie 3ml, 7ml, 10ml and 20ml in (GLS) gas ,liquid and solid fluidized bed with air, water and benzoic acid pellets as three phases respectively in the system. The presence of nanomaterial increases the solid liquid mass transfer coefficient value with increasing fraction of nanomaterial, increasing gas velocity and increasing time although experimental run has been taken only for one hour.

Using a Dynamic and Constant Mesh in Numerical Simulation of the Free-Rising Bubble

A two-phase bubbly flow is often found in the process industry. For the efficient operation of such devices, it is important to know the details of the flow. The paper presents a numerical simulation of the rising bubble in a stagnant liquid column. The interFOAM solver from the open source Computational Fluid Dynamics (CFD) toolbox OpenFOAM was used to obtain the necessary data. The constant and dynamic computational grids were used in the numerical simulation. The results of the calculation were compared with the measured values. As expected, by using the dynamic mesh, the bubble trajectory was closer to the experimental results due to the more detailed description of the gas–liquid interface.

An Empirical Correlation for Zero-Net Liquid Flow in Gas-Liquid Compact Separator

Compact separators have significant application for subsea separation and offshore application. However, their operating envelope is usually narrow due to physical phenomena such as liquid carryover and gas carry-under. Before the occurrence of liquid carryover, the separator operates in what is termed zero-net liquid flow (ZNLF). Though there is an efficient separation during ZNLF; there is also liquid holdup in the upper section of the separator, which is termed as ZNLF holdup. The ZNLF holdup in a cyclonic separator during an actual gas-liquid separation was studied experimentally. The ZNLF holdup was measured directly using electrical resistance tomography (ERT). The direct measurement approach is an improvement of the existing method, which depends on measuring the pressure drop across the stagnant liquid column. The results showed that increasing gas flow rate at a constant liquid flow rate increase zero-net liquid holdup in the upper part of the separator. An empirical correction was developed, and the correlation predicted the experimental results with a ±10% error margin. The correlation could be useful as part of the input into a pressure drop model for calculating pressure drop across the gas leg of the cylindrical cyclonic separator. This correlation will be useful to process engineers for optimum design and operation of a gas-liquid compact separator.

Detached eddy simulations of rising Taylor bubbles

A numerical study of turbulent flow induced by Taylor bubbles rising in a stagnant liquid column is presented in this paper. Predictions of the void fraction, mean velocity and turbulent fluctuations are compared with available experimental results. In addition, we analysed the wall pressure and differential wall pressure fields as well as the temperature field produced by the interaction of a Taylor bubble with a heated tube wall. Implications for pressure measurements and thermal fluctuations in two-phase systems are discussed. A second simulation was performed to shed light on Taylor bubble coalescence, specifically highlighting the distinct mechanisms that operate in different parts of the wake region.

A novel numerical approach for investigation of the gas bubble characteristics in stagnant liquid using Young-Laplace equation

In the present study, the Young-Laplace equation was applied to simulate the adiabatic gas bubble growth from a submerged needle in stagnant liquid column. In order to solve the Young-Laplace equation the axisymmetric bubble height was used as input from experimental data. To increase the accuracy of Young-Laplace equations’ prediction during the bubble growth, the bubble was divided into four sections with the same height, and Young-Laplace equation was solved for each section individually. By dividing the bubble into four sections, the effects of viscosity and inertia forces within each section were reduced as compared to that of buoyancy and liquid-gas surface tension. Unlike the conventional Young-Laplace approach (one Young-Laplace equation for the entire bubble), the new approach was able to predict bubble characteristics reliably during the growth cycle. The bubble growth was investigated in a column of liquid with a triple contact line that fixed to the needle perimeter. To validate the numerical results, the bubble profiles that predicted by numerical simulation were compared with the experimental results. Experiments were performed by injection of air at constant gas flow rate of 600ml/h in the quiescent deionized water and SiO2 nanofluid. The nanoparticle concentrations were 0.05, 0.1 and 0.2wt%, and air flow injected from G14 and G17 standard needles. Eventually, evaluation of bubble characteristics, such as the bubble volume, the center of gravity, the instantaneous contact angle, and the bubble aspect ratio were investigated, and the effects of variation of liquid properties on the bubble characteristics were discussed. The results show that the present method can predict the bubble shape during 97.5% of growth time with mean absolute error of 6%. Furthermore, the results revealed that the bubble size decreased with increment of Bond number. Also, bubble instantaneous contact angle and bubble aspect ratio were almost irrelative to Bond number during the growth cycle.

Dynamics of a Taylor bubble in steady and pulsatile co-current flow of Newtonian and shear-thinning liquids in a vertical tube

A computational analysis is carried out to ascertain the effects of steady and pulsatile co-current flow, on the dynamics of an air bubble rising in a vertical tube containing water or a solution of Carboxymethylcellulose (CMC) in water. The mass fraction (mf) of CMC in the solution is varied in the range 0.1%⩽mf⩽1% to accommodate zero-shear dynamic viscosities in the range 0.009–2.99Pa-s. It was found that the transient and time-averaged velocities of Taylor bubbles are independent of the bubble size under both steady as well as pulsatile co-current flows. The lengths of the Taylor bubbles under the Newtonian conditions are found to be consistently greater than the corresponding shear-thinning non-Newtonian conditions for any given zero-shear dynamic viscosity of the liquid. In contrast to observations in stagnant liquid columns, an increase in the dynamic viscosity of the liquid (under Newtonian conditions) results in a concomitant increase in the bubble velocity, for any given co-current liquid velocity. In shear-thinning liquids, the change in the bubble velocity with an increase in mf is found to be relatively greater at higher co-current liquid velocities. During pulsatile shear-thinning flows, distinct ripples are observed to occur on the bubble surface at higher values of mf, the locations of which remain stationary with reference to the tube for any given pulsatile flow frequency, while the bubble propagated upwards. In such a pulsatile shear-thinning flow, a localised increase in dynamic viscosity is accompanied near each ripple, which results in a localised re-circulation region inside the bubble, unlike a single re-circulation region that occurs in Newtonian liquids, or shear-thinning liquids with low values of mf. It is also seen that as compared to frequency, the amplitude of pulsatile flow has a greater influence on the oscillating characteristics of the rising Taylor bubble. The amplitude of oscillation in the bubble velocity increases with an increase in the CMC mass fraction, for any given value of pulsatile flow amplitude.

An Effective Method for Modeling Non-Moving Stagnant Liquid Columns in Gas Gathering Systems

Abstract Gas gathering system modeling is often complicated by the presence of localized pressure losses that are not easily explained by traditional pressure loss correlations. The tendency is to assume the pressure loss correlation must be tweaked to match measured operating conditions. This often leads to inappropriate manipulation of tuning factors (efficiency factor or roughness). A more appropriate approach is to recheck the validity of the input data and, most importantly, visit the field prepared to gather additional data to resolve the causes of the localized pressure losses. The objectives of this paper are to discuss one of the causes of localized pressure losses —Stagnant Liquid Columns —and to present several cases where localized pressure losses were interpreted to be caused by stagnant liquid accumulations. Introduction The maturation of the gas gathering systems throughout North America has resulted in the majority of systems being operated well below their original design conditions. Consequently, it is common to encounter pressure losses that exceed those predicted by steady-state single-phase and two-phase correlations. The reasons for these pressure losses are varied and most often relate to measurement issues, poor understanding of the pipeline and facility connections and sometimes non-moving liquid accumulations called stagnant liquid columns. Liquid accumulations are a concern because their continuous removal is difficult and they increase backpressure for all upstream wells, which reduces well deliverability and can result in localized pipeline corrosion. One would think that the traditionally used steady-state two-phase pressure loss equations would be capable of predicting the pressure loss that occurs in these liquid accumulations. However, the steady-state flow of fluid must be, by definition, continuous: inlet rate equal to outflow rate. If the liquid enters the conduit and accumulates while the gas continues through, we no longer have steady-state flow and the validity of the correlation no longer holds. It has been the authors' experience that localized pressure losses are often associated with liquid accumulations. Typically, field staff usually know there is a problem and may or may not have already realized that it is due to liquids, but it is usually a surprise for head-office staff because there is little or no liquid production reported. This leads to the question, how do we reliably identify stagnant liquid columns and how should they be modeled? To answer this question, a discussion of recommended modeling procedures is required. Discussion Pressure Loss Categorization The existing steady-state pressure loss correlations generally do a very good job of estimating pressure losses when used appropriately. Consequently, a comparison of simulated line pressures with field measured line pressures should result in a reasonable match within a preset tolerance. Measured line pressures that do not match modelled line pressures must be scrutinized closely to determine the cause of the mismatch, rather than relying primarily on tweaking tuning factors such as pipeline roughness or flow efficiency. For wells or groups of wells where a reasonable match is not initially obtained, it is good practice to begin by attempting to classify the unmatched pressure losses as either a systemic or localized step pressure loss problem before making changes to the model to force a match.

Prediction of two‐phase flow distribution in parallel pipes using stability analysis

AbstractTwo‐phase gas liquid flow in pipes is a complex process. One of the problems that is hardly understood is how the two phases are distributed among two or more parallel lines with a common inlet manifold. Steady‐state analysis yields multiple steady‐state solutions. Linear and nonlinear (simulation) stability analyses are performed in order to determine the actual distribution of the flow that will take place in a real system. The analysis shows that when there are four parallel pipes, for example, the two‐phase flow mixture from the common inlet manifold can choose to flow in one, two, three, or in all four pipes, depending on the flow rates of the liquid, and the gas and on the pipes inclination. For low‐flow rates of gas and liquid, the flow tends to take place only in one line, while stagnant liquid columns are present in the other three pipes. As the flow rate increases the flow will take place in 2, 3 and finally in 4 pipes. Experimental data confirm the analysis although matching is only approximate. © 2006 American Institute of Chemical Engineers AIChE J, 2006

Gas–liquid phase distribution and void fraction measurements using MRI

We used a permanent-magnet magnetic resonance imaging (MRI) system to estimate the integral, spatially- and temporally-resolved void fraction distributions and flow patterns in vertical gas–liquid two-phase flow. Air was introduced at the bottom of the stagnant liquid column using an accurate and programmable syringe pump. The air flow rate was in upward vertical direction against the gravity vector. Air flow rates were varied between 1 and 200 ml/min. Bubble formation and detachment at the inlet of the vertical column was non-uniform. The cylindrical non-conducting test tube in which two-phase flow is characterized was placed in a 0.26 T permanent-magnet MRI unit. A roughly linear relationship has been observed for the integral void fraction versus volume flow rate of air through vertical stagnant liquid column. Integral or sample-averaged void fraction was obtained by volume-averaging of the spatially-resolved signals. Spatial averaging was performed in a radial or longitudinal axis of the column. The time-averaged spatially-resolved void fraction has also been obtained for the quasi-steady flow of air in a stagnant liquid column. No great accuracy is claimed as this was an exploratory proof-of-concept type of experiment. Preliminary results show that MRI can indeed provide qualitative and quantitative data and is especially well suited for characterization of averaged transport processes in adiabatic and diabatic multi-phase and/or multi-component flows. Better and faster MRI sequences could improve spatial and temporal resolutions significantly.

Level Swell and Flooding of R113 Two-Phase Mixture with a Falling Film in Vertical Pipes.

Using R 113 vapor and liquid, behavior of the two-phase mixture level was examined in both cases of bubbling a stagnant liquid column in vertical pipes and bubbling a liquid column to which liquid was supplied as a falling film. The pipe diameters used were 10.0, 14.8, 20.2 and 26.0 mm. The initiation condition of flooding in the latter case was also examined. Observed phenomena were quite similar to those in previous air-water experiments. However, the level fluctuation amplitude and the mixture level swell height in the experiments with the falling film were smaller than those in the air-water experiments. Physical models devised to predict the mixture level height and the flooding velocity under a slug flow condition, which were developed in previous work, were found to reproduce the experimental results well.

Countercurrent bubble and slug flows in a vertical system

A model for estimating void fraction in countercurrent bubbly and slug flows in vertical systems, based on the drift-flux approach, is presented. In countercurrent flow, as in cocurrent flow, the in situ gas velocity is aided in the upward direction by the buoyancy of the gas phase as well as its tendency to flow through the channel center. These two factors allow us to express the in situ gas velocity and the void fraction in terms of the bubble rise velocity, the mixture velocity, and the flow parameter. C 0. Single bubble rise velocity data gathered for various falling liquid velocities agreed well with the Harmathy correlation. The parameter C 0 in countercurrent bubbly flow was found to be 2.0, a value much higher than the generally accepted one of 1.2 for cocurrent flow. Significant liquid recirculation at the pipe wall during countercurrent bubbly flow, causing changes in the velocity and bubble concentration profiles similar to those observed in bubbly flow through stagnant liquid columns in large diameter pipes, is postulated to be the cause for this high value of C 0. The Taylor bubble rise velocity data gathered at various falling liquid velocities agreed well with the Nicklin equation. Visual observation indicated little distortion in the shape of Taylor bubbles due to liquid flow in the opposite direction. Void fraction during slug flow is computed using two different approaches. Good agreement is found between theory and experiment.

Drift velocity of elongated bubbles in inclined pipes

The drift velocity, which is defined as the velocity of long bubbles propagating into a stagnant liquid column, has been treated successfully in the past using inviscid flow theory. However, different approaches were used for the horizontal and the vertical cases. In this work, Benjamin's simple approach, which was originally applied for the horizontal case only, is extended to the inclined and the vertical cases, taking into consideration surface tension effects. Good agreement with experimental data is obtained.

Two-phase mixture level swell in vertical pipes

The behavior of the two-phase mixture level was examined in the case of bubbling of a stagnant liquid column in vertical pipes and also in the case of bubbling of a liquid column to which liquid is supplied as a falling film. In a range of low air flow rates where the flow pattern is of bubbly type, the mixture level swell and its fluctuation amplitude were small. However, these values increased sharply as the air flow rate was increased and the flow pattern turned into a slug type. The mean height of the two-phase mixture level and the level fluctuation amplitude were analyzed physically and compared with the experimental results.

Upward Vertical Two-Phase Flow Through an Annulus—Part I: Single-Phase Friction Factor, Taylor Bubble Rise Velocity, and Flow Pattern Prediction

Upward gas-liquid flow through vertical concentric and fully eccentric annuli was studied both experimentally and theoretically. A flow system was designed and constructed for this study. The system consists of a 16-m long vertical annulus with 76.2-mm i.d. casing and 42.2-mm o.d. tubing. A comprehensive experimental investigation was conducted for both concentric and fully eccentric annuli configurations, using air-water and air-kerosene mixtures as the flowing fluids. Included were definition and classification of the existing flow patterns and development of flow pattern maps. Measurements of volumetric average liquid holdup and average total pressure gradient were made for each flow pattern for a wide range of flow conditions. Additional data include single-phase friction factor values and Taylor bubble rise velocities in a stagnant liquid column. Data analysis revealed that application of the hydraulic diameter concept for annuli configurations is not always adequate, especially at low Reynolds number flow conditions. A more rigorous approach was thus required for accurate prediction of the flow behavior, especially for two-phase flow. Part I of the study includes experimental data and analyses of single-phase friction factor, Taylor bubble rise velocity, and flow pattern transition boundaries.

Two-phase Mixture Level Swell in Vertical Pipes.

This study examines the behavior of a two-phase mixture level in the case of bubbling of a stagnant liquid column in vertical pipes and also in the case of bubbling of the liquid column to which liquid is supplied as a falling film. In a range of low air flow rates where the flow state is a bubbly type, the mixture level swell and its fluctuation amplitude are small; however, these values increase sharply as the air flow rate increases and the flow state turns to a slug type. The mean height of the two-phase mixture level and the level amplitude are physically analysed and compared with the experimental results.

On the performance limit of closed two-phase thermosyphons.

An experiment was conducted to examine the performance limit of closed two-phase thermosy-phons, together with visual observation of flow state in the adiabatic section. The working fluids used were R 113, methanol and water. The flow state at the performance limit conditions was a violently disturbed slug type, in which the vapor plugs hold up the liquid slugs periodically to a high level, causing a local liquid circulation in the adiabatic section. This phenomenon is somewhat different from the flooding observed in open systems. Then, an equation correlating the vapor velocity at the performance limit to the rising velocity of vapor plugs in stagnant liquid columns is proposed. This correlation is in good agreement with the performance limit data covering a wide range of parameters.