Search

Slugging flow is a condition caused by a liquid obstruction at the riser base. It exhibits cyclic behaviour. The cycle consists of a protracted time of no gas production at the riser's top, followed by the arrival of a liquid slug with a length greater than the riser height, and ultimately the breakthrough of a significant gas surge. The cycle time might range from a few minutes to a few hours, depending on the system size and flow conditions. In offshore oil production, feedback control is a practical and cost-effective way to prevent slug flow. To control the flow rate or the pressure in the pipeline, adjusting the choke valve opening on the topside facility is generally utilised as the control input. From a practical standpoint, designing a control system based on topside data rather than seabed measurements is preferable. Controlling the topside pressure alone is difficult and ineffective in reality, but combining it with the flow rate results in a more reliable control solution. Measuring the flow rate of a multiphase flow, on the other hand, is difficult and expensive. All the topside measurements-based slug control techniques was critically reviewed and necessary recommendations for enhanced control performance provided. In conclusion, this review acknowledged that slugging is a well-defined flow pattern, yet despite having been studied for several decades, current slug control methods still have robustness issues. Slug flow problems are expected to become even more intense in the future as a result of longer vertical risers driven by deep-water Exploration and Production (E&P).

Experiments on aeration from a bubble diffuser pipe having five aeration holes are carried out to investigate effects of the azimuthal angle of the holes on flows inside and outside the pipe. The azimuthal angle is varied by rotating the pipe. When the hole angle becomes larger than a certain angle, the liquid height inside the pipe is fixed just below the holes, which results in the prevention of slugging inside the pipe. The effects of hole angle on bubble generation mode and bubble size are small except for the large hole angles, at which generated bubbles are to break up due to an interaction between the bubbles and the pipe wall. Hence uniform aeration is easily realized just by rotating the pipe to some extent. Uniform aeration is also realized with a longer diffuser pipe having ten aeration holes by rotating the pipe. The power in aeration is evaluated from the pressure difference and the total gas inflow. Although the decrease in the hole diameter is also effective to realize the uniform aeration, the increase in the azimuthal angle is superior to the former from the point of view of energy saving.

Operating forces and magnitude of loads from gas-liquid slug flows exerted on a horizontally orientated 90 o bend are investigated. The distributed forces are either Newtonian, associated with the fluids motion or Configurational, inherent to the internal distributions of the phases. The forces are derived through the conventional balances of mass and linear momentum arising from the volume of fluid (VOF) description of gas-liquid flows. The study uses the integral form of the momentum balance to estimate the operating forces budget. Invoking dynamical time scales separation discloses the connection of the Lamb vector (vortex-force) to the local time rate of momentum. An interesting outcome being an explicit expression for Favre-Reynolds stress that reveals the contribution of void fraction fluctuations in the redistribution of the stress across the interface. Numerical simulations are performed to determine the magnitude of Newtonian loads on bend using a segmented domain technique to represent the fully established slug flow regime. The time-dependent traces of the relevant flow variables such as liquid hold-up, flow rates and resultant forces on the bend are recorded and analysed. Compared to the isotropic component, the deviatoric stresses are shown to have a marginal contribution to the total forces. It is also shown that loading cycles on bends are much higher than slugging cycles; this is an important feature for the structural integrity assessment of pipelines with bends.

The dynamics of gas fluidization is investigated by means of multiresolution analysis. Empirical mode decomposition (EMD) and the Hilbert-Huang transform approach applied to the signals of pressure fluctuation of the bed have been used for this purpose, operating in bubbling and slugging regimes. To elucidate the different components of the different scales, recurrence plots (RP) and recurrence quantification analysis (RQA) have been used. These techniques can distinguish the three different scales of gas fluidization: micro-scale, meso-scale and macro-scale, and classify every mode on its scale. Three modes from the EMD have been related to each dynamic component: particle interaction, local bubble dynamics and bed oscillation, showing evidence of this relationship. To show that the complexity of the modes matches with their characteristics, two measures have been computed: the apparent entropy and Lemped–Ziv complexity.

Active control has been recognised as one of the most promising methods to mitigate slugging by several researchers. However, choke valves used in offshore oil and gas production systems are often in a large size with a very large stroke time (time for a valve traveling from fully open to fully close or vice versa). This leads to not only a significant but also variable input time delay because the actual time delay also depends on the input increment value at each sampling point. In many practical offshore slug control systems, the large valve stroke time has been a major factor impacting significantly on the slug control performance. In this work, a modified Smith predictor is proposed to resolve this problem. Firstly, the traditional Smith predictor, which was mainly designed for fixed measurement time delay, is specially modified for variable input time delay using measureable delayed signal (actual valve position). Then, a linear model is adopted to present the slugging system without a time delay. Finally, the modified Smith predictor is implemented and tested using OLGA simulation software coupled with Matlab through OPC. Results show that the modified Smith predictor can significantly improve slug control performance when the choke valve has a large stroke time. With the Smith predictor, the system can operate at a larger valve opening, hence increase oil production.

All pipelines should be operated under engineering protection to assure that the oil and gas will be safely and economically transported from wells all the way to the downstream processing plant. Wax, asphaltene, and scales can deposit on the wall of pipe and block the oil and gas flowing inside the pipeline. Corrosion may occur at inner wall and outer wall of pipe, reducing the strength of the pipelines, causing loss of pipeline integrity. Severe slugging may form inside the pipeline and induce operational problems to the downstream processing facilities. This chapter provides fundamental knowledge of composition and chemistry of fluids being transported through pipelines, which are essential for attacking flow assurance issues that attribute to operational problems in oil and gas transportation.

In the co-current flow of gas-liquid mixtures through pipeline-riser system, severe slugging is frequently encountered and manifests in significant fluctuation of flow and pressure. This can pose serious threat to production facilities. The most common method for its mitigation is by choking which unfortunately could negatively affect production. The objective of this study therefore is to develop a technique that can help stabilise the system and maximise production simultaneously. In this paper, a new general method for multiphase flow system stability analysis was proposed based on a new passive attenuation method – the intermittent absorber. A series of experiments were carried out in a 4″ pipeline-riser system which is 55 m long with horizontal pipe inclined at −2° connected to 10.5 m high catenary riser followed by 3 m horizontal topside section. Air and water were used as experimental fluids. Numerical studies were also conducted on a 4″ pipeline-riser system to proof the concept. The results showed that the intermittent absorber possess potential for stabilising severe slug flow at larger valve opening and lower pressure. For the case studied, up to 9% reduction in the riser base pressure was reported which practically implies increased oil production.

Square nosed slugging fluidization behavior in a circulating fluidized bed riser using a polyethylene powder with a very wide particle size distribution was studied. In square nosed slugging fluidization the extent of mixing of particles of different size depends on the riser diameter, gas velocity, hold up and solids flux in the riser. Depending on the operating conditions the particle residence time distribution of a riser in the slugging fluidization regime can vary from that of a plug flow reactor to that of a well-mixed system. Higher gas velocities cause shorter particle residence times because of a significant decrease in the hold-up of particles in the riser at higher gas velocities. A higher solids flux also shortens the average residence time. Both influences have been quantified for a given polyethylene–air system. Residence time and residence time distribution were determined for different particle size and the influence of gas velocity, solids flux, hold up and riser diameter was studied. When comparing data from segregation and residence time experiments it is clear that segregation data can predict the spread in residence time as a function of overall residence time, particle size and gas velocity. The differential velocity between small and large particles found in the segregation experiments can predict the spread in residence time as found in the residence time distribution experiments with a powder with a broad particle size distribution. Raining of particles through the slugs was studied as a function of plug length, gas velocity and pulse length. It was found that raining is not the determining mechanism for segregation of particles.

In this paper the concept of micro-fluidized beds is introduced. A cylindrical quartz reactor with an internal diameter of only 1 mm is used for process conditions up to Click to view the MathML source and 244 bar. In this way, fast, safe, and inherently cheap experimentation is provided. The process that prompted the present work on miniaturization is gasification of biomass and waste streams in hot compressed water (SCWG). Therefore, water is used as fluidizing agent. Properties of the micro-fluid bed such as the minimum fluidization velocity (Umf), the minimum bubbling velocity (Umb), bed expansion, and identification of the fluidization regime are investigated by visual inspection. It is shown that the micro-fluid bed requires a minimum of twelve particles per reactor diameter in order to mimic homogeneous fluidization at large scale. It is not possible to create bubbling fluidization in the cylindrical micro-fluid beds used. Instead, slugging fluidization is observed for aggregative conditions. Conical shaped micro-reactors are proposed for improved simulation of the bubbling regime. Measured values of Umf and Umb are compared with predictions of dedicated 2D and 3D discrete particle models (DPM) and (semi)-empirical relations. The agreement between the measurements and the model predictions is good and the model supports the concept and development of micro-fluid beds.

Computational fluid dynamic (CFD) models must be thoroughly validated before they can be used with confidence for designing fluidized bed reactors. In this study, validation data were collected from a fluidized bed of (Geldart's group B) alumina particles operated at different gas velocities involving two fluidization hydrodynamic regimes (bubbling and slugging). The bed expansion, height of bed fluctuations and frequency of fluctuations were measured from videos of the fluidized bed. The Eulerian–Eulerian two fluid model MFIX was used to simulate the experiments. Two different models for the particle stresses—Schaeffer [Syamlal, M., Rogers, W., O’Brien, T.J., 1993. MFIX documentation: theory guide. Technical Report DOE/METC-94/1004 (DE9400087), Morgantown Energy Technology Centre, Morgantown, West Virginia (can be downloaded from Multiphase Flow with Interphase eXchanges (MFIX) website 〈 http://www.mfix.org 〉 ); Schaeffer, D.G., 1987. Instability in the evolution equations describing incompressible granular flow. Journal of Differential Equations 66, 61–74.] and Princeton [Srivastava, A., Sundaresan, S., 2003. Analysis of a frictional–kinetic model for gas–particle flow. Powder Technology 129(1–3), 72–85.] models—and different values of the restitution coefficient and internal angle of friction were evaluated. 3-D simulations are required for getting quantitative and qualitative agreement with experimental data. The results from the Princeton model are in better agreement with data than that from the Schaeffer model. Both free slip and Johnson–Jackson boundary conditions give nearly identical results. An increase in coefficient of restitution ( e ) from 0.8 to 1 leads to larger bed expansions and lower heights of fluctuations in the bubbling regime, whereas it leads to unchanged bed expansion and to a massive reduction in the height of fluctuations in the slugging regime. The angle of internal friction ( φ ) in the range 10 – 40 ∘ does not affect the bed expansion, but its reduction significantly reduces the height of fluctuations.

Publisher Summary Wellhead chokes are used to limit production rates for regulations, protect surface equipment from slugging, avoid sand problems due to high drawdown, and control flow rate to avoid water or gas coning. Two types of wellhead chokes are used: positive (fixed) chokes and adjustable chokes. Placing a choke at the wellhead means fixing the wellhead pressure and thus the flowing bottom-hole pressure and production rate. For a given wellhead pressure, by calculating pressure loss in the tubing the flowing bottom-hole pressure can be determined. If the reservoir pressure and productivity index is known, the flow rate can be determined on the basis of inflow performance relationship. This chapter presents and illustrates different mathematical models for describing choke performance.

The fluidization behavior of yttrium oxide (Y2O3) powders of high density and micronic diameter belonging to the group C of Geldart's classification has been investigated. Large interparticle forces lead to bed cracking, slugging and channelling, and cause the powder not to fluidize consistently. Different fluidization technologies have been tested, such as mechanical agitated fluidization, vibrated fluidization and addition of easy-to-fluidize large particles to fine particles. The quality of fluidization has been studied through pressure drop diagrams for decreasing gas velocities and for various fixed bed heights to column diameter ratios. In the case of stirred fluidization, several stirrer geometries have been tested (helix, turbine, etc.). However, the fluidization has not been satisfactory. By adding larger particles to fine powders, convenient fluidization conditions have been obtained. An inertia effect proportional to the initial bed weight seems to contribute to fluidization. Some evaluation of interparticle forces governing the tested mixture of fine/large particles has been performed by studying the influence of mass percentage of fine particles on the Hausner ratio and the angle of repose. Fluidization under vibration allows to partly overcome the adhesion forces between powders. The fluidization behavior has been improved for the highest vibration strengths.

Anti-slug control applied to two-phase flow provides a very challenging and important application for feedback control. It is important because it allows for operation that would otherwise be impossible, and challenging because of the presence of both RHP-poles and RHP-zeros. To conduct experiments on pipeline-riser anti-slug control, a small-scale slug-loop has been build. Theloop has been modeled andanalyzed usingasimplifiedmodel byStorkaas. The results from this analysis and experimental results using a PI-controller is presented in this paper. Some of the problems in the offshore oil industry that have received increasingly interest in the last years are related to multiphase flow. In multiphase flow different flow regimes can develop, depending on parameters such as flow rates, fluid properties and pipeline geometry. Slug flow is a flow regime which can cause a lot of problems for the production facilities. The slug flow is characterized by alternating bulks of gas and oil, and can be further divided into hydrodynamic and terrain induced slugging. Hydrodynamic slugs are caused by velocity differences between the phases and occur in near horizontal pipelines. These slugs are usually short and appear frequently. Terrain induced flow however, can contain a lot of liquid and therefore induce large pressure variations in the system. This flow is induced by low points in the pipeline geometry. When the low-point is realized by a downsloping pipe terminating in a riser, we get what is known as riser slugging. Because of the large and abrupt fluctuations in pipe pressure and gas and liquid flow rates at the outlet, these slugs cause huge problems for the processing equipment. Unwanted variations in the separator level give rise to poor separation and possible overfilling. The pressure fluctuations wear and tear on the equipment and can sometimes result in unplanned process shutdowns.

Recent advances in sub-sea oil production technology have changed the way off shore oil fields are developed. Many new field developments use sub-sea production facilities and pipe a multiphase mixture of gas, oil and water back to an existing platform or onshore facilities for separation and processing. These pipelines can be very long (up to 100km long) and complex—with complicated branching and crossovers. Dynamic features of multiphase flow, such as slugging, can cause disturbances in the processing system. Operators need help in handling these. Model-based estimates of pressure, holdup and flow along the pipeline are thus needed. An integrated system for operator support and operator training is being delivered for a large sub-sea gas field in the Middle East. It builds upon a dynamic process simulator (D-SPICE) and a multiphase flow simulator, (OLGA 2000). The simulator incorporates methods for simulating multiphase flow in closed-loop networks and tracking of hydrate inhibitors. The system is tuned to measured process performance. The system also supports predictive simulation facilities. The paper provides an, industrial case study of computer-aided process engineering in a technically challenging environment. Novel methods for multiphase modelling in closed loops, real-time model system design and model tuning are presented.

In 1989 it was demonstrated conclusively that the pressure drop caused by a bend in a pneumatic conveyor occurs not in the bend itself, but in the straight pipe immediately downstream of the bend, with the length over which the pressure drop is developed being of the order of eight meters for lean phase (suspension) flow and rather less (three to five meters) for slugging or plug flow. The prediction of the effects of bends in pneumatic conveying systems has previously eluded analytical study based on physical models, owing to the difficulty in obtaining a physical model which yields behavior representative of that observed in practice. This chapter presents the results of a study based on a simple physical model which, owing to its novel approach, does yield realistic behavior. The model is based on momentum transfer between two distinct phases of flow in the pipeline, and predictions made using this model have been compared with measurements taken from an industrial-scale pneumatic conveying system. The comparison is promising with regard to the potential for development of this line of modelling for the prediction of bend effects in pipelines. At present this model has been tested for lean phase conveying only.

Publisher Summary Gas—liquid contacting under conditions of slug flow occurs in many practical situations in the chemical industry. This chapter presents and reviews the research studies carried out in the field of physical modeling in slugging gas—liquid columns. It highlights the importance of the wakes of slugs as the primary agent in promoting mixing. The wake of the slug is a region of recirculatory flow spanning (nearly) the whole cross section of the tube and extending over a length of up to about one or two tube diameters below the slug. This local movement of the liquid is bound to be of great relevance to the process of mixing. The chapter also addresses the problem of Taylor dispersion for laminar liquid flow along a tube and presents an analysis of dispersion in co-current flow of gas slugs and liquid in vertical tubes.

Two-phase saturated steam-water flow was studied experimentally under atmospheric conditions. Wave growth and slugging from stratified flow regime in a horizontal duct were analyzed by using a video camera. The result for slug initiation was compared with the instability criteria based on linear analysis and the empirical correlations of other researchers. The required instability criteria is in good agreement. The reason for the necessary lower saturated steam velocity in comparison with Kelnin-Helmholtz relation and growth of short wavelength waves which create slugs were discussed and presented by some pictures. The result of this kind of careful experimentation was aimed in verification of inherent problems which are contained in two fluid models.

This chapter discusses the fluidization regimes of the fluidized state. The phenomenon of fluidization can be best visualized in terms of a simple experiment. Particles such as sand are poured into a tube provided with a porous plate distributor. Gas or liquid is then forced upward through the particle bed. This flow causes a pressure drop across the bed. When this pressure drop is sufficient to support the weight of the particles, the bed is said to be at minimum fluidization. A further increase in the flow rate of gas often causes bubbling that can be seen through the walls of a transparent tube or at the top of the bed. For narrow tubes and a sufficiently deep bed, the bubbles coalesce and form a slug. These gas slugs then move up the bed in a fairly regular periodic motion. Large-size particles, such as corn, can be made to spout if the gas is brought in through a small central tube. A spout is a centrally located, upwardly moving particle region. Above such a spouted bed, the particles drop down in a fountain and move down the tube in an annular region. As the gas velocity is increased substantially above the terminal velocity, the fluidized bed, the slugging bed, and the spouted bed are all blown out of the tube. Continuous operation in the dense-phase regime can be continued by rapidly feeding solids into the bottom of the bed. The resulting flow regime is called recirculating fluidization.

Publisher Summary Slug flow is a highly complex type of flow with an unsteady nature. The prediction of pressure drop, heat, and mass transfer for such flow is often considered a difficult task. The chapter deals with steady slug flow. The chapter introduces several options of modeling the hydrodynamic parameters and pressure drop using a unified approach that is applicable for the vertical, horizontal, and inclined cases. The chapter also reviews transient phenomena in slug flow by illustrating an example of severe slugging in a pipeline-riser system. It is noted that heat transfer in slug flow is of major importance for practical applications. The treatment of the two-phase flow is usually considered as a two-phase mixture. The severe slugging that consists of one riser and one pipeline is one of the simplest example of slug flow under nonsteady conditions. The Boe criterion differentiates between steady and cyclic operations with two exceptions. At high liquid flow rates, a steady flow can also exist within the severe slugging region predicted by the Boe criterion. There is a region outside the Boe criterion that is in an unsteady state and leads to unsteady oscillations.

In gas-solid slug flows in circulating fluidized beds, two different slugs can be distinguished: round-nosed and square slugs. Separated mechanistic models are proposed to describe these different slugs. Based on the models round-nosed slugs are shown most likely to behave like fast bubbles and square slugs to behave like slow bubbles. New criteria are established to predict the type of slug flow under which a circulating system will operate. Experiments of slugging transport bed were conducted using 1050 μm sand in 6.35 and 10.16 cm columns. Predictions of the present criteria satisfactorily agree with the data obtained in this study and the observations of the dense bed in Battel le's Multisolid Circulating Fluidized Bed Combustor being operated commercially. The criteria are verified to be more general than Stewart's criterion (1965) for the demarcation of different types of slug flows. The present models are also more general than Chen and Fan's (1987) by considering a more general case of continuous solid flow in the circulating system. The rise velocity of any slug can be described by a single equation by the models, U S = k 1 + k 2 (gD) 1/2 , where k 1 is greater than 1 for the round-nosed slug flow and is less than 1 for the square slug flow. The slower rise velocity of square slugs is attributed to the significant wall friction of non-fluidized particles.

We have developed a fluidized bed containing a draft tube for the purpose of gasifying low grade coal. The bed is effectively divided into two zones by the draft tube, and by conducting the gasification and combustion reactions separately in the two zones we have been able to produce synthesis quality gas using air and steam as the gasifying agents. In the draft tube where coal and steam react, the solids move upward in slug flow. It seems particularly advantageous to operate the draft tube in the square nosed slugging regime, because of the improved gas solid contacting which results from the high dense phase gas flow rate and low rise velocity of the slugs. The rate of solid transport induced by this slug flow up the draft tube and around the bed is experimentally and theoretically described.

Slugging in standpipe is a main factor influencing the operation stability of circulating fluidized bed. In this paper the causes of slugging are analysed. It is indicated that for steady and continuous flow of gas-solid mixture in standpipe the gas-solid relative velocity, U, should be greater than the minimun fluidization velocity, Umf and less than the slugging formation velocity, Usl. Some qualitative guidelines for designing standpipes with a small diameter are presented. Based on this analysis, a laboratory scale circulating fluidized bed was set up and operated satisfactorily. Reaction of oxy-dehydrogenation of butene to butadiene was conducted in this bed and excellent results were obtained.

Based on the observed vertical profile of average pressure in the bed and FFT analysis of its fluctuation, new criteria are proposed to distinguish higher gas velocity fluidization into different fluidization regimes; such as, slugging, transient, turbulent and dilute phase regime. To determine these fluidized regimes approximately from the measurement of pressure fluctuation in the bed, a simple method is also proposed. Observed data of transition Rep between each regime measured under both atmospheric and high pressure conditions can be correlated well as function of Ar. Effects of operating conditions upon the voidage in the turbulent fluidized bed and in its freeboard are also briefly discussed.

The multisolid pneumatic transport bed is a fluidized bed in which a circulating bed of fine particles is superimposed on a slugging or turbulent fluidized bed of dense particles. Pessure fluctuations were measured in the dense particle bed of a 0.102 m I.D. multisolid pneumatic transport bed using two differential pressure transducers. The pressure fluctuation data was used to determine the slugging characteristics of the dense particle bed and the transition from the slugging regime to the turbulent regime. The addition of circulating fine particles to the slugging dense bed resulted in a significant increase in slug velocity and caused the transition to the turbulent regime to begin at a lower velocity.

This work suggests that a fast fluidized riser consists of two parts as follows, (i) The lower part of the riser is occupied by a violently bubbling or slugging fluidized bed (analogous to churn-turbulent gas-liquid flow). (ii) The upper part of the riser contains a core, moving rapidly upwards, in which the particles are entrained by the gas; adjacent to the wall is a thin dense film of particles falling under gravity. Measurements of the flow rate of particles in this fallings film in a 0.15 m diameter fast fluidized bed have been made by means of scoops which collect the particles in the film; the film thickness appears to be about 1 — 2 mm. The flow rate of particles in the film, w, was found to decrease with increasing height, x, above the air distributor: it was found that w = w∞ + δw exp(-Kx) where w∞, δw and K are constants; this implies similar variation of the upward flow rate of solids in the core. These observations are consistent with the hypothesis that there is turbulent mass transfer of particles from the core to the film, and transfer coefficients for transport of particles have been inferred: these transfer coefficients are similar to those obtained for deposition of droplets in annular two-phase flow. The internal recirculation of particles thus generated within the riser may be several times the solids feed rate: this explains the high apparent slip velocities and gas-solids contacting efficiencies reported by other workers.

Publisher Summary The full benefits of fluidization, although with coarse and uniform-sized particles, are not always realized because of the growth of large bubbles in the bed causing a tendency toward slugging. This limitation of fluidized beds can be overcome if the gas enters the bed through a small opening at the center of a conical base instead of a uniform distributor. In the process of the bed becoming a composite of a central spout, a systematic cycle pattern of solids movement is established with effective contact between the gas and the solids, giving rise to a unique hydrodynamic system, which is more suitable for certain applications than more conventional fluid–solids configurations. This chapter signifies the growing acceptance of spouted beds lies primarily in the specific fluid and particle dynamics associated with them. Spouting action can be achieved with either a liquid or a gas as the jet fluid. The minimum particle diameter for which spouting is practical appears to be 1–2 mm. Fine materials tend to fluidize rather than spout. For a given solid material, column diameter and fluid inlet diameter, a “maximum spoutable bed depth” exists beyond which the spouting action degenerates into poor-quality fluidization. A typical spouted bed has a substantial depth, which in the case of a cylindrical vessel is usually at least of the order of two column diameters, measured from the inlet orifice to the surface of the annulus. The total pressure drop across a spouted bed is always lower than that required to support the weight of the bed. Spouting can be achieved only within certain limits of solids properties while whether or not a material having properties within these limits will spout depends upon the geometry of the column, including to some extent the design of the gas inlet.