Churn Flow Regime Research Articles

Fluid structure behaviour in gas-oil two-phase flow in a moderately large diameter vertical pipe

Intermittent flows in vertical pipes occur in many industrial settings including power generation and downstream oil-and gas production. This type of flows include cap bubble, slug and churn flow regimes. These regimes are of interest as downstream processes and control may heavily depend on the intermittency of the inflow. There are a number of correlations that predicts the features in such flows in vertical pipes. Most of the correlations were developed for air and water fluid pair for slug flow regime in vertical pipes with 25 to 50 mm inner diameter. In this paper, an attempt has been made to assess the suitability of several of these correlations specific to slug flow regime for a fluid pair that is different to air-water system. In this work, air-silicone oil flow development was experimentally investigated in a vertical pipe with an inner diameter of 68 mm. A Wire Mesh Sensor (WMS) and an Electrical Capacitance Tomography (ECT) sensor were installed in series at four locations (15D, 30D, 45D and 65D) downstream of the mixing section. The flow was visually observed using a high speed camera. The void fraction time series obtained from the WMS and the ECT were used to establish the flow characteristics such as slug length, slug frequency, void fraction in liquid slugs and Taylor bubble velocity. A comparison showed that the void fraction measurements using ECT and WMS are in good agreement. Axial measurements shows that the flow development beyond 45D is minimal. Change in physical properties of the liquid phase is responsible for the deviation associated with the existing slug flow models, particularly those developed to predict the gas holdup in liquid slugs.

Experimental and numerical evaluation of slugs in a vertical air–water flow

Multiphase flows are often present in chemical processes and they include gas-liquid transport, which may become intermittent depending on the operating conditions. One type of flow structure is the presence of large gas bubbles that can occupy the entire duct diameter, separated by liquid portions. This can result in large pressure fluctuations, which are normally associated with damage to equipment over time. The intermittent nature and the chaotic distribution of the phases in a turbulent flow complicate the design and optimization of the equipment. In this context, computational fluid dynamics is a useful technique since it can provide important information on several flow conditions, but the use of suitable models and boundary conditions is required. In this study, an air and water flow with the presence of slugs is evaluated in a system consisting of horizontal and vertical ducts, with an inner diameter of 94 mm. Experimental data on the pressure and void fraction were analyzed and compared to the predictions obtained from numerical simulations, obtained considering the VOF approach. It was observed that the use of the geometrical reconstruction scheme to simulate a compressible flow resulted in a better agreement with the experimental results. Using this setup, the average volume fraction and pressure values showed the expected behavior. The frequency of slugs, established using three different methods, was around 1.5 Hz for the evaluated conditions evaluated. The pressure probability density function (PDF) indicated good agreement between numerical simulations and experimental data, while the volume fraction PDF indicated that the flows evaluated have the characteristics of a churn flow regime. Moreover, it was found that an increase in the water flow rate can lead an increase or a decrease in the average velocity.

Similarities and differences in churn and annular flow regimes in steady-state and oscillatory flows in a long vertical tube

This study investigates the effects of induced-low-frequency flow oscillations on gas–liquid two-phase flows using a 42-m long, 0.048-m ID vertical pipe system. This work uses the experimental data set from a previous study on the effect of forced flow oscillations on churn and annular flow. However, in this previous study, the similarities and differences between flows in steady-state and under induced-low-frequency oscillations were not discussed. The time variation of liquid holdup, frequency of large liquid structures, and pressure gradient were analyzed under two superficial liquid velocities (0.017 and 0.3 m/s), and superficial gas velocities ranging from 4 to 21 m/s. The impact of flow oscillations in two-phase flow regime transitions is also investigated. The observations of flow regimes under oscillatory flow showed significant differences compared to the ones expected in steady-state flow under similar conditions. For higher pressures and oscillatory flow, churn flow regime was observed, when annular flow regime is expected in steady-state flow. Annular flow regime was not observed for higher superficial liquid velocities and high inlet pressure for oscillatory conditions. Surprisingly, the pressure gradient for oscillatory flow is significantly lower than in steady-state flow. However, this lower pressure gradient effect is only observed for low superficial gas and liquid velocities at the pipe inlet and outlet. This experimental observation is attributed to the formation of large liquid waves at the pipe inlet, which is a consequence of the induced-low-frequency flow oscillations.

Churn flow in high viscosity oils and large diameter columns

Churn flow is an important intermediate flow regime occurring in between slug and annular flow patterns in two-phase flow, with profound implications in chemical and petroleum industry. The majority of studies to date in churn flow have been carried out mainly using water or liquids of low viscosities and limited information exists regarding the behaviour of high viscosity liquids which resemble realistic process conditions. In this paper, a study that investigated churn flow and its characteristics in high viscosity oils (360 and 330 Pa.s) and large diameter columns (240 and 290 mm) is presented for a first time. Transition to churn flow regime starts when the structure velocity, length and frequency of the liquid bridges, which appear at the end of slug flow, increase. In churn flow, gas flows at the core of the oil column with a wavy passage, leaving the top surface open to atmosphere with a possibility of creating a very long bubble. The average length of the bubbles seen to decrease with increasing the gas flow rate. While, no considerable change is observed in void fraction, structure velocity and film thickness at this flow pattern.

The entrainment mechanism and droplet size distribution in gas-liquid churn flow

Knowledge of entrained liquid is important in calculation of pressure drop and in the determination of the film flow rate and thence its dryout and consequent deterioration in heat transfer in heating systems. Note that the excessive liquid entrainment may cause the liquid film completely removed from the wall (dryout), which could lead to disastrous accident. As one of the least understood gas-liquid flow regimes, churn flow appears a highly-disturbed flow of gas and liquid. According to the observation, churn flow is characterized by huge waves with flow reversal between the waves and the highly oscillatory liquid film, which is accompanied a continuous gas core with considered amount of entrained liquid. Although there have been enduring efforts to show that the entrained fraction is high in churn flow and reaches the minimum around the churn/annular transition, the underlying mechanisms of droplet entrainment in churn flow are still not well explored. Additionally, there is a dearth of data about entrained droplets in churn flow due to its chaotic nature. The measurement techniques which are suitable to determine entrained droplets in annular flow have limitations when applied to churn flow because they might not be able to distinguish between liquid carried as droplets and that transported in huge waves. In this study, we focused on the entrainment mechanism, liquid distribution and droplet size distribution under churn flow condition. The high-speed camera and the high-resolution camera were employed to capture a more detailed description of the huge wave breakup and the liquid distribution in the cross-section of the pipe with the inner diameter of 19 mm. In order to obtain the high-resolution images of entrained droplets, a delicate shadow detecting method was developed in the present paper, which provides a reliable approach to capture the entrained droplets with various sizes. Subsequently, Matlab was employed to analyze the complete series of pictures and recognize each of the blobs. The blobs were interpreted as droplets after the sharpness was satisfied. Every accepted edge datum is considered to be a droplet and thus its diameter can be obtained by Image J. The results indicate that there are three mechanisms for droplet generation in churn flow: liquid bridge breakup dominants at slug/churn transition, bag breakup plays a dominant role at low gas superficial velocities, but ligament breakup comes to gain greater importance with the increase of gas flow rate. Accordingly, the amount of liquid entrained is high in the chaotic churn flow regime and gradually decreases during the transition from churn flow to annular flow and finally reaches a minimum around the churn/annular transition boundary. Detailed process of droplet entrainment was also provided. Based on observation, large droplets (chunks) are related to the breakdown of slugs and bag breakup mechanism, whereas smaller droplets can be ascribed to the breakup of chunks, ligament breakup. No doubt, the present paper helps understand the mechanism of the entrainment in churn flow, which is essential for the development of mechanistic models to predict the pressure drop and dryout condition.

The effect of mesh-type bubble breaker on two-phase vertical co-flow

The influence of a mesh-type bubble breaker in a two-phase vertical co-flow is experimentally investigated using high-speed imaging. A range of liquid and gas flow rates were considered that correspond to the gas-to-liquid flow rates ratios (GLRs) from 0.05 to 39.6 covering bubbly, slug and churn flow regimes. The mesh-type bubble breaker was found to be effective in reducing the size of nozzle-generated bubbles in two-phase co-current vertical flow. The bubble size reduction by 50%-70% is observed in the presence of the mesh-type bubble breaker compared to that without a bubble breaker. The results also show that at a given GLR, the mesh-type bubble breaker delayed the flow regime transition compared to that without a bubble breaker. A Froude number correlation was proposed to predict the mean bubble size downstream of the bubble breaker for the bubbly flow regime, which accurately predicted the bubble size over a range of bubble breaker parameters. It is observed that the geometry of the bubble breaker influence the two-phase flow regime transition. At low liquid flow rates, small pore size of the breaker was found to be effective. This effect diminishes at higher liquid flow rate due to the elongation of the bubbles by the liquid inertia. The results also show that a longer bubble breaker located closer to the bubble generating nozzle is the most effective in producing bubbly flow over a longer range of gas-liquid flowrates ratios.

Entrained droplets in two-phase churn flow

The presence of droplets exerts a strong influence on liquid film flow rate, pressure drop, heat transfer and many other flow characteristics. Due to the complexity of churn flow, profound knowledge on the droplets in this flow pattern is not well documented in the existing literature. In the present paper, we study a gas-liquid two-phase churn flow in a 19-mm-inner-diameter pipe under atmospheric pressure and carefully design a shadow detection technique to capture the high-resolution images of the entrained droplets in churn flow. The images are subsequently binarized using an adaptive threshold. The results indicate that the amount of the entrained droplets is high in the chaotic churn flow regime and gradually decreases during the transition from churn flow to annular flow and finally reaches a minimum around the churn-annular flow transition. The detailed process of the droplet entrainment is discussed. Based on our experimental observations, it is concluded that the large droplets (chunks) are related to the breakdown of slugs and bag breakup mechanism, whereas the smaller droplets can be ascribed to the breakup of chunks, ligament breakup and impingement. A normalized size probability density function is implemented to analyze the effect of the entrainment mechanism on the range of droplet size. The Sauter diameter of the droplets in churn flow is compared with the existing empirical correlations in annular flow and a formula for the entrained droplet size in churn flow is proposed.

A simplified model for churn and annular flow regimes in small- and large-diameter pipes

This paper presents an improved model for gas-liquid two-phase flow in the churn and annular flow regimes for small- and large-diameters in vertical and near-vertical pipes. Many investigators consider churn flow to be the least understood flow regime in upward gas-liquid flows in vertical pipes. Nevertheless, this flow regime occurs commonly in several applications in the oil and gas industry, such as: gas-lift operations, production of gas-condensate wells, and two-phase flows with high gas-liquid-ratio in general. Additionally, the accuracy of two-phase flow models for large pipe diameters, irrespective of flow regimes, is still questionable. Only a limited number of studies can be found in the open literature that validates two-phase flow models for pipe diameters larger than 0.203m (8in.).The model developed in this study proposes a modification to an approach originally proposed in previous studies for pipe diameters smaller than 0.0508m (2in.). These modifications have proven to improve accuracy in the prediction of pressure gradient and liquid holdup for small and large pipe diameters under churn and annular flow conditions. The absolute average error in the prediction of experimental pressure gradient decreases from 155% (for the original model) to 36% (for the modified version). This study validates the proposed model with field and laboratory experimental data from several different studies from the literature, in terms of pressure-gradient and liquid holdup results, for pipe diameters ranging from 0.0318 to 0.279m (1.25–11in.), pressures from 1 to 613bara (14.6–8900 psia), and fluids types such as air-water and oil-natural gas systems.This study also compares its results with other widely accepted models in the commercial packages. When compared to field and laboratory data, the simulation results show that this new model has an overall better performance compared to other models widely used in the oil and gas industry.

Analysis of vertical co-current air-water flow with the Neptune_CFD

A two-fluid modelling approach with interface sharpening is applied for numerical simulations of flooding waves in air-water co-current flow in vertical pipe. The simulation capability of Neptune_CFD code is analysed. The results are compared with experimental data and with the simulations performed with the computational code ANSYS CFX. Isothermal flow of air and water in the churn flow regime is modelled. Turbulence is modelled for each phase separately using URANS approach. A typical geometry of a test section in the vicinity of the liquid injector is used as the computational domain. In general, both computational codes predicted similar wave frequency values for the pipe with internal diameter of 32mm and show good agreement with the experimental data. Detailed analysis of local flow phenomena in the case of 2D Neptune_CFD simulation however reveals a significant accumulation of liquid in the centre of the pipe that may be attributed to the axisymmetric assumption, drag force formulation, and possible numerical errors. It was demonstrated that the liquid accumulation in the centre of the pipe can be avoided by the simulations in the full 3D domain due to asymmetric nature of the churn flow.

Effect of forced flow oscillations on churn and annular flow in a long vertical tube

An experimental study has been carried out to investigate the effects of periodic forced flow oscillations on gas-liquid two-phase flows using a 42m long, 0.05m ID vertical pipe system. From the knowledge of the authors, there is no experimental work available in the literature on the forced flow oscillations for churn flow regime. Therefore, the main objective of this paper is to experimentally characterize churn and annular flows in a long vertical tube. The time variation of liquid holdup, pressure drop and pressure gradient were analyzed under two superficial liquid velocities (0.017 and 0.3m/s), and superficial gas velocities oscillations ranging from 4m/s to 21m/s. The impact of flow oscillations in two-phase flow regime transitions is also investigated.The observations of flow regimes under oscillatory conditions showed significant differences compared to the ones expected in steady-state flow under the same conditions. Under oscillatory conditions, the local liquid holdup and the amplitude of its waves decreased with the distance from the inlet. For the experimental runs with oscillatory gas flow and low superficial liquid velocities (Uls=0.017m/s), the liquid holdup at the bottom of the tube decreased while the gas flow rate was lowered, whereas in steady-state conditions the liquid holdup should increase while reducing superficial gas velocities. This process is described in this study by the formation of large waves in the bottom of the pipe as a consequence of the flow oscillations.

The effect of the diameter on air-water annular and churn flow in vertical pipes with and without surfactants

In this work, we present results of flow visualisation, pressure gradient measurements, and liquid holdup measurements for air-water flow without and with surfactants in vertical pipes with diameters of 34 mm, 50 mm, and 80 mm. The surfactants cause the formation of foam. This foam has a larger volume and a smaller density than the liquid. The larger volume results in a larger pressure gradient at large gas flow rates. At small gas flow rates, the lower density of the foam causes the transition between the regular annular flow regime and the irregular churn flow regime to shift to lower gas flow rates. As a result foam reduces the pressure gradient and the liquid holdup at small gas flow rates. Surfactants reduce the pressure gradient more effectively for thinner liquid films at the wall; therefore, they are more effective for small pipe diameters and small liquid flow rates.

The Effect of Surfactants on Vertical Air/Water Flow for Prevention of Liquid Loading

Summary From field experience in the gas industry, it is known that injecting surfactants at the bottom of a gas well can prevent liquid loading. To better understand how the selection of the surfactant influences the deliquification performance, laboratory experiments of air/water flow at atmospheric conditions were performed, in which two different surfactants (a pure surfactant, sodium dodecyl sulfate, and a commercial surfactant blend) were added to the water. In the experiments, a high-speed camera was used to visualize the flow, and pressure-gradient measurements were performed. Both surfactants increase the pressure gradient at high gas-flow rates and decrease the pressure gradient at low gas-flow rates. The minimum in the pressure gradient moves to lower gas-flow rates with increasing surfactant concentration. This is related to the transition between annular flow and churn flow, which is shifted to lower gas-flow rates because of the formation of an almost stagnant foam substrate at the wall of the pipe. At high surfactant concentration, it appears that the churn flow regime is no longer present at all and that there is a direct transition from annular flow to slug flow. The results also show that the critical micelle concentration, the equilibrium surface tension, the dynamic surface tension, and the surface elasticity are poor predictors of the effect of the surfactant on the flow.

Two-phase flow in short horizontal rectangular microchannels with a height of 300 μm

The two-phase flow in a narrow short horizontal channel with a rectangular cross section is studied experimentally. The channel has a width of 10, 20, or 30 mm and a height of 300 μm. The specifics of formation of such two-phase flows are investigated. It is demonstrated that the regions of bubble and churn flow regimes grow and constrain the region of jet flow as the channel gets wider. The boundaries of the regions of annular and stratified flow regimes remain almost unaltered.

Simulation of flooding waves in vertical churn flow

A transient simulation of flooding waves in the churn flow of air and water in a vertical pipe is performed by the means of two-fluid modelling approach with interface sharpening. The gas and liquid phases are considered immiscible and incompressible with no mass transfer between them. Inter-phase coupling of momentum is realized via interface drag force which is based on the interface area density and the relative velocity between the phases. Surface tension effects are modelled with the Continuum Surface Model. The flow is assumed isothermal. Turbulence is modelled for each phase separately using the two-equation eddy viscosity approach. Results are compared with the reported experimental data for churn flow regime in a vertical pipe (Wang et al., 2011a). Reynolds numbers of the gas flow are in the range from 6000 to 10,000, while the liquid mass flow rate upwards ranges from 25 to 32g/s. Prediction of critical and maximum amplitudes of the flooding waves show good agreement with experimental values. Results for wave frequencies indicate significant deviations, which can be attributed to the choice of the liquid inlet model.

Interfacial area concentration in gas–liquid bubbly to churn flow regimes in large diameter pipes

This study performed a survey on existing correlations for interfacial area concentration (IAC) prediction and collected an IAC experimental database of two-phase flows taken under various flow conditions in large diameter pipes. Although some of these existing correlations were developed by partly using the IAC databases taken in the low-void-fraction two-phase flows in large diameter pipes, no correlation can satisfactorily predict the IAC in the two-phase flows changing from bubbly, cap bubbly to churn flow in the collected database of large diameter pipes. So this study presented a systematic way to predict the IAC for the bubbly-to-churn flows in large diameter pipes by categorizing bubbles into two groups (group 1: spherical or distorted bubble, group 2: cap bubble). A correlation was developed to predict the group 1 void fraction by using the void fraction for all bubble. The group 1 bubble IAC and bubble diameter were modeled by using the key parameters such as group 1 void fraction and bubble Reynolds number based on the analysis of Hibiki and Ishii (2001, 2002) using one-dimensional bubble number density and interfacial area transport equations. The correlations of IAC and bubble diameter for group 2 cap bubbles were developed by taking into account the characteristics of the representative bubbles among the group 2 bubbles and the comparison between a newly-derived drift velocity correlation for large diameter pipes and the existing drift velocity correlation of Kataoka and Ishii (1987) for large diameter pipes. The predictions from the newly-developed two-group IAC correlation were compared with the collected experimental data in gas–liquid bubbly to churn flow regimes in large diameter pipes and their mean absolute relative deviations were obtained to be 28.1%, 54.4% and 29.6% for group 1, group 2 and all bubbles respectively.

On the effect of liquid viscosity on interfacial structures within churn flow: Experimental study using wire mesh sensor

In the churn flow regime, periodical interfacial structures such as liquid slugs and huge waves can coexist and undoubtedly, a phase property such as liquid viscosity can dominate the behavior of these structures. Regrettably, neither are the characteristics of churn flow widely understood nor have the effects of liquid viscosity on gas–liquid flow received enough attention.A Wire Mesh Sensor (WMS) with a 16×16 spatial resolution was employed to discover the effects of liquid viscosity on the behavior of churn flow in a vertical 76.2mm pipe. Three liquid viscosities of 1, 10, and 40cP, and superficial liquid velocities of 0.46, 0.61, and 0.76m/s were employed; whereas, superficial gas velocity ranged from 10 to 27m/s. Different techniques such as Probability Density Function (PDF), and 2-D and 3-D image reconstruction methods were applied to study the flow. It was noticed that increasing liquid viscosity not only affected the flow pattern but also the appearance frequencies of interfacial structures.

Intermittent flow regime of water/gas emissions as sinkhole trigger in alluvial plains: a study case at S. Vittorino plain (Rieti, Italy)

A case of natural intermittent flow regime (slug/churn) of a CO2-enriched spring water is discussed as an example of trigger process for the collapse of sinkholes. The hydrodynamic regime of the spring, situated in the area of S. Vittorino alluvial plain (Rieti, Italy), was investigated by monitoring acoustic emission, carrying out geochemical analyses and installing a sedimentation trap during a long-lasting phase of slug flow regime followed by a churn flow phase. The slug and churn flows can erode and transport upward fine and medium size particles of sand, progressively removing them from some underground shallow sandy levels where the groundwater mixes with gas of deep crustal origin to form the intermittent flow that can be observed at the surface. The slug flow regime was characterized by a stable slugging frequency whose value is in agreement with the estimated superficial velocities of Taylor bubbles and liquid slugs. The transition from the slug to the churn flow regime, the latter characterized by an increase of gas flux and by more severe erosion, was also observed during our study. In a span of 15 years, the transport of sand upward could produce a collapse volume similar to those of the sinkholes that recently originated in the same area. The study of such multiphase flows prolonged over time could help to monitor and prevent the sinkhole formation process.

Time–frequency analysis of acoustic and unsteadiness evaluation in effervescent sprays

The acoustic signals from effervescent sprays were experimentally investigated across a range of injection pressures, flow rates, and gas to liquid ratios by mass (GLR). An unconventional description of spray features under various operational conditions was obtained through the time–frequency–amplitude distributions of spray acoustic by adaptive optimal kernel (AOK). The acoustic energy was calculated through the Hilbert Huang transfer (HHT) method and was confirmed to be affected by both the GLR and air flow rate. A new evaluation of spray unsteadiness was established based on this finding. The results show that the gas flowing out of the orifice leads to an increase in the amplitude of the high frequency component. Discrete phenomenon in effervescent sprays was exactly observed by AOK spectrogram as the amplitude distribution was discontinuous with time and the amplitude fluctuations showed variation in different conditions. The factors influencing the acoustic energy include both air and liquid flow rate. The unsteadiness levels under different conditions were calculated after the new spray unsteadiness was evaluated through a simple atomizer. The unsteadiness levels had a low value when the GLR was higher than 5% and varied across a wide range when the GLR was below 4%. The flow regime had a significant effect on the spray unsteadiness as the bubbly and churn flow regimes produce a relatively stable spray, whereas the slug flow regime induces unsteadiness.

The effect of surfactants on air–water annular and churn flow in vertical pipes. Part 1: Morphology of the air–water interface

In this work, the influence of surfactants on air–water flow was studied by performing experiments in a 12 metre long, 50mm inner diameter, vertical pipe at ambient conditions. High-speed visualisation of the flow shows that the morphology of the air–water interface determines the formation of foam. The foam subsequently alters the flow morphology significantly. In annular flow, the foam suppresses the roll waves, and a foamy crest is formed on the ripple waves. In the churn flow regime, the flooding waves and the downwards motion of the liquid film are suppressed by the foam. The foam is transported in foam waves moving upwards superposed on an almost stagnant foam substrate at the pipe wall. Foam thus effectively reduces the superficial gas velocity at which the transition from annular to churn flow occurs. These experiments make more clear how surfactants can postpone liquid loading in vertical pipes, such as in gas wells.

The effect of surfactants on air–water annular and churn flow in vertical pipes. Part 2: Liquid holdup and pressure gradient dynamics

In this paper, we study the effect of surfactants on both the liquid holdup and the dynamics of the pressure gradient in annular and churn flow in vertical pipes. This effect is linked to the influence of the surfactants on the morphology of the air–water interface, which is studied in a related paper (van Nimwegen et al., 2014). The experimental results, obtained using a vertical flow loop with a 5cm internal diameter at ambient pressure, show three different effects of the surfactants on the measured quantities, depending on the air and water flow rates. (i) At large air flow rates, in the annular flow regime for air–water flow, the surfactants increase the pressure gradient; this is solely due to the increase of the frictional pressure gradient, caused by the larger interfacial stress between the foamy waves overlaying the foam substrate along the wall and the gas core. (ii) At low air flow rates and low water flow rates, in the churn flow regime for air–water flow, the surfactants decrease significantly both the average pressure gradient and the pressure gradient fluctuations. While the frictional pressure gradient increases, the liquid holdup decreases by more than a factor of two; the foam suppresses the flooding waves, making the flow much more regular, leading to the small pressure gradient fluctuations. Furthermore, there exists an optimum surfactant concentration for decreasing the average pressure gradient and the pressure gradient fluctuations. (iii) At low air and high water flow rates, in the churn flow regime for air–water flow, the average pressure gradient and the pressure gradient fluctuations are both somewhat decreased by the surfactants.