Investigating The Behaviour Of Air-silicone Oil Flows In Vertical And Horizontal Pipes For Effective Gas-liquid Transport

1.1  Introduction 

The simultaneous flow of fluids is difficult in a safe and controlled way, with the exception that various behaviors of the flow can be predicted with adequate reliability. It deals with the concurrent flow of fluids within different phases (i.e. gas, liquid and solid) or the different chemical properties but in the same phase, for example gas-liquid, gassolid, liquid-solid, liquid-liquid and gas-liquid-solid (Abdulkadir, 2015). Multiphase flows are encountered in industries like; the petroleum, chemical, and nuclear industries. The transportation of gas- liquid two-phase flow in the petroleum industry over long distance is quite common. This simultaneous flow is encountered in instances like the flow of oil from the reservoir to the separator, and to the process facilities. As pressure decreases, gas starts to evolve, thereby creating a two-phase flow in the pipeline. Various difficulties are encountered in the flow of these fluids, some of which are phase velocity differences and the existence of several flow regimes. These flow regimes include; bubbly, slug, churn, plug, and stratified, among many others. The existence of these flow regimes in transportation lines poses certain challenges to the industry because they increase the pressure drop, heat transfer, mass and corrosion rate in the pipeline. Since the accurate prediction of these flow patterns is essential to the success of designing multiphase flow systems in vertical and horizontal flows, there is therefore, need to investigate the behavior of the fluids in pipes of various inclinations, for effective transportation in the industry. 

1.2 Critics of Churn Flow  

One may wish to know the reason for this particular subheading, the critics of churn flow. The topic is discussed because of the various schools of thoughts and ideologies from different research about the existence of churn flow regime. The question is, does Churn flow exist as a distinctive flow pattern or, it is  just an extension of slug flow? To address this, this work is yet to find out as Mao and Dukler, and Hewitt and Jayanti presented different ideas about the existence of this particular flow. 

According to Mao and Dukler (1993) in their paper, “The Myth of Churn Flow?” they presented evidence that proved that churn flow pattern is a simple and continuous extension of the condition of slug flow and that no transition actually existed. Therefore it is not a distinctive and separate flow pattern on its own. In view of this, they presented two different pieces of evidence to buttress their point. These are visual evidence and instrumental evidence where experiments were performed to support their findings. 

1.2.1 Visual Evidence 

This evidence is no different from what other researchers observed in transparent pipes. Based on this, one can conclude that their observations proved the existence of churn flow as a unique and separate pattern that exists as a transition from, slug flow to annular flow. 

According to Mao and Dukler, stable slug flow is an upward motion of a quasi-periodic arrangement of alternating Taylor bubbles and liquid slugs at a constant speed of propagation, and that the length of the Taylor bubbles and the liquid slugs remain the same as they rise. The velocity associated with the bubbles and liquid slugs is uniformly upward and the same as that which exists in front of both the slug and bubble. As the gas rate increases, the flow becomes chaotic, and the size of the liquid slug and Taylor bubbles increase forming lumps of bubbles as they move up and down the pipe. The flow then becomes oscillatory and displays irregular periods. Again, the main characteristic of slug flow is the falling of the liquid film around the Taylor bubbles and it disappears. All these observations show characteristic features of churn flow, thereby proving that it exists as a unique flow pattern. 

1.2.2 Instrumental Evidence 

In instrumental evidence, three different experiments were carried out to provide data for the work by using a 50.6 mm diameter vertical pipe. These data included the axial profile of the void fraction in the liquid slug, the thickness profile of the liquid film around the Taylor bubble, the axial profile of the magnitude and the direction of the shear stress along the bubble and slug, and the velocity propagation of the front and back of the bubbles and slugs. 

From the first experiment conducted using an RF probe at three gas flow rates, three gas velocities and a constant liquid superficial velocity, it concluded that; the lowest gas velocity, 0.76 m/s showed traces of full slug flow. The second experiment of 1.41 m/s showed slug flow with some up and down motion of the slugs, while the highest velocity 3.42 m/s, showed a churn flow with all the chaotic behaviors. From the second experiment, the velocity of 200 bubbles and slugs was determined by cross-correlation of two axial conductance probe positioned parallel to each other. From this, the percentage of slugs moving downward was determined. It was concluded that, very few slugs moved downward to be able to characterize it as churn flow, and that the percentage was not too different from what was observed to be fully slug flow. The last experiment was based on using a model to calculate the characteristic of slug flow. The experiment was done by averaging the void fraction over a unit cell containing liquid slugs and Taylor bubbles. The experiment was conducted using different flow rate pairs, 24 for slug flow and 18 for churn flow proved that for the 24 flow rate pairs, the calculated and experimental void fraction was  7% in agreement; the error estimated being 6%, and the standard deviation was 3.1%. The standard deviation for the 18 runs was 5.3%. With these values, they concluded that since the agreement between the churn flow model and slug flow model for void prediction was not too different, then churn flow is not a distinctive flow pattern. From these three experiments, they finally concluded that churn flow does not exist but it is just an extension of slug flow, and their initial evidence of the existence of churn as a distinctive model was a mistake. 

Hewitt and Jayanti (1993), in their paper, “To Churn or Not to Churn”, raised an argument against Mao and Dukler concerning their findings. Hewitt and Jayanti disproved the findings based on the fact that, experiments were carried out at pressures close to atmospheric pressure. They came up with the fact that, the term churn has three uses and that the third type which is the distinctive flow pattern is characterized by the presence of flooding type waves throughout the flow regime. From this, they concluded that perhaps the gas velocity that was considered by Mao and Dukler, which is the 3.42 m/s was not high enough to get to the transition state between slug and annular since the flooding type waves would not have taken place. In this perspective, the 3.42 m/s gas velocity was taken as the second type of churn flow, which is the slug extension that was refferred to by Mao and Dukler. Hewitt and Jayanti stated that the main purpose of designating a flow regime is to be able to develop a model that can accurately predict a particular flow regime with a peculiar character, thereby reducing the number of flow patterns. In line with this, certain characteristics were listed which were only identifiable with churn flow and not annular or slug flow. These characteristics include; the flooding wave which is a predominant phenomenon associated with churn flow pattern and the periodic reversals of the liquid film; the increase in the entrainment fraction by the flooding waves, and droplet deposition dominated by  radial velocity  imparted at   the point   of  the droplet  creation.  The characteristics, as presented above, are significantly different from that of annular flow or slug flow. Since the characteristics of these two flow patterns are different, churn can be treated as a distinctive flow pattern. Again this flow pattern is noted to cover a wide range of flow velocity. Also, the transition of slug to churn exists at a gas velocity of 3.5 m/s and that this flow persists to velocities from 1015 m/s. Based on these findings, they concluded that Churn flow exists as a distinctive pattern and not an extension of slug flow. It is  also important to consider higher velocities of 10-15 m/s since most of the industrial equipment operates within these velocities. 

1.3 Problem Statement 

The success and safe design of multiphase flow systems depend on the accurate prediction of the various flow regimes that exist within the flow system. This is because, these regimes increase the rate of heat transfer, pressure drop and the corrosion rate of the system. These challenges intrigued a research in this area. To help solve this problem, there is a need for accurate determination of the void fraction which helps in the characterization and prediction of the flow regimes existing at a particular point in the flow system. Although numerous works have been done in this area, the results obtained still require further scrutiny since they reported poorly especially for churn flow. As the name implies, churn flow is a highly disturbed flow of liquid and gas in which the liquid motion is oscillatory, going up and down alternating, although not in a periodic and regular manner (. The net flow of liquid is generally in the direction of the flow of gas, although it may be zero or even negative. This flow regime is normally identified between the slug and annular flow regimes in vertical or near vertical upward flow. However, it occurs at low liquid flow rates, where an uninterrupted change to an annular flow pattern may occur, either from slug flow or bubbly flow regimes. 

Churn flow is rare amid the flow regimes in that, several schools of thought exist regarding the mechanism of transition to this flow regime. It is also rather challenging to predict its existence. For example, McQuillan and Whalley (1985) reported only a 36% success rate in predicting it, whereas other flow regimes were predicted with an accuracy of up to 80%. Others, for example, Mishima and Ishii (1984), and Brauner and Barnea (1986) did not adopt such an approach to measure the effectiveness of their models, but even in their results, several instances of inappropriate prediction of churn flow can be found, and  the success rate was also rather poor. It is possible that this poor argument is as a result of one or both of the following: 

• The predictive methods do not model the underlying mechanisms to a satisfactory degree of complexity;  

• There may be more than one mechanism, over a range of flow conditions, governing this transition.  

Due to the inappropriate prediction of the flow regimes, there is the need for further investigation into the possibilities leading to the occurrence of the various flow regimes in both, horizontal and vertical pipes using unpublished experimental results obtained at the University of Nottingham, United Kingdom. 

1.4 Objectives of Research 

The objectives of this research are threefold: 

? To obtain void fraction data from wire mesh sensor (WMS) (67 mm internal diameter and 6 m long pipe) at gas superficial velocities up to 4.7 m/s using air– silicone oil as the model fluid from Nottingham University for the following; 

? To determine the effect of gas and liquid superficial velocities on void fraction. 

? To determine the effect of inclination angle on void fraction. 

? To identify the radial phase distribution of the flow. 

? To develop a drift-flux correlation that is capable of correctly predicting the void fraction for churn flows in vertical pipes. 

? Finally, flow regimes identification will be performed using the newly collected data to confirm the range of flow conditions which represent churn flow. 

1.5 Organization of the Research 

The report is structured in this format; 

? Chapter one; introduction of the work, defining the problem, state the aims and objectives, the methodology employed to carry out the work and showing how the entire work will be structured. 

? Chapter two; is the literature review. Previous work or research done on gas/liquid flow patterns in vertical pipes, transition of flow pattern, flow pattern Maps, identification in vertical and horizontal pipes for two-phase flow, and factors affecting the formation of churn flow, 

? Data acquisition and the methodology employed to do this work is captured in Chapter three. 

? Chapter four; the results and discussions. This chapter details the various results obtained and the analysis performed on the results in achieving the objectives set for the work. 

? Chapter five will conclude the work and give Recommendations of further study if any.