ABSTRACT In the oil and gas industry, sand can be entrained with production of gas and liquid and can cause erosion damage and failure of pipelines, especially in elbows that are widely used to change the flow orientation. As a result, erosion failure of pipelines can occur which causes significant costs of workover and significant environmental problems can occur especially in offshore environments. Additionally, liquid dominated flows can be found in deep water, offshore oil-wells, subsea flowlines during production of gas and oil. Despite the need, there is a gap in the literature regarding erosion in elbows with liquid dominated multiphase flows, both numerically and experimentally. In this work, investigators have performed experiments to understand erosion patterns in elbow under liquid dominated flow conditions in a 50.8 mm diameter experimental facility. Additionally, Computational Fluid Dynamics (CFD) simulations have been performed and compared with experimental pattern results. In this study, a flow loop facility allowing flow of liquid, gas and sand was used with an acrylic elbow for erosion pattern visualization. CFD simulations are compared to experimental erosion patterns with both test results for sand-liquid and sandgas-liquid flows. The qualitative comparison between experiments and CFD simulations showed satisfactory agreement and is providing a basis for improving erosion models for these flow conditions.

ABSTRACT As production takes place in the petroleum industry, impurities are produced along with oil and gas. Sand is a common impurity associated with production. If sand is present, it is desirable to have continuous flow of even small concentrations of sand through the pipeline to avoid multiple issues like accumulation of particles and under-deposit corrosion. Experiments are performed at extremely low sand concentrations for both liquid-sand (single phase) and gas-liquid-sand (multiphase) flows to determine the critical flow rates necessary to keep sand moving in horizontal lines. In the case of gas-liquid-sand (multiphase) flow, the flow regime is stratified. The critical velocity is defined as the fluid velocity required to keep the solid particles continuously moving. Generally, the critical velocity is considered to have a direct relationship with the solid concentration for relatively low concentrations. That is for low concentrations, as the sand concentration increases the critical velocity needed for transport increases and vice versa. However, it can be reasoned that as the concentration approaches zero, the critical velocity should not be a function of concentration, since the particle spacing becomes large. In this work, this behavior has been shown experimentally, and the concentration at which the critical velocity is no longer a function of concentration is referred to as the threshold concentration. In this paper, the threshold concentration and the minimum critical velocity are examined for liquid-sand and gas-liquid-sand flows.

ABSTRACT An experimental study was conducted to investigate the effect of particle size on the critical velocity and near wall turbulence characteristics of water flow over the sand bed deposited in a horizontal pipeline. Industrial sands of four mesh sizes (20/40, 30/50, 40/70, 100) were used in the study. The particle image velocimetry (PIV) technique was used to determine instantaneous local velocity distribution near the stationary sand bed-fluid interface under flow conditions. The results indicated that the critical velocity for the onset of particle removal from the sand bed as well as the near wall velocity profile, Reynolds shear stress, axial and radial turbulent intensities were all influenced by the variation in particle size range.

ABSTRACT An experimental two-phase flow campaign with a viscous oil (100 cP) was conducted in the Large Scale 8" loop at the SINTEF Multiphase Flow Laboratory. The loop consisted of three main test sections, with pipe inclinations 0, 0.5 and 90 degrees, and with approximate lengths of 380 m, 380 m and 50 m, respectively. The experiments were performed at system pressures 20, 45 and 85 bara. The primary focus of the campaign was on liquid dominated flows in horizontal/near-horizontal pipes, with particular emphasis on the laminar-to-turbulent transition in the liquid. The main instruments were DP-cells for pressure drop measurements, and narrow beam gamma densitometers to measure the liquid height. In addition, traversing gamma densitometers were mounted on each of the two near-horizontal sections to measure the time-averaged liquid distribution and volume fractions. In this campaign, it was found that at high pressure, the slug flow region was very narrow. In particular, at low gas-liquid ratios, the prevailing flow regime was determined to be a kind of " pseudo slug flow", which was characterized by large waves that did not quite extend to the top of the line. Simulations with the commercial multiphase simulator LedaFlow [1] did not reproduce this at the time, and the penalty for this discrepancy was that the predicted pressure drop was too high. Consequently, it was concluded that the physical models in LedaFlow needed improvement to predict these conditions better. Through detailed analysis of the experimental data, it was found that at these conditions, slugs were often not able to form because the waves did not have sufficient inertia to sustain a slug front. This limitation was not accounted for in the flow regime criteria in LedaFlow, so to improve the situation, the existing slug flow model in LedaFlow was generalized to cover pseudo slug flow in addition to regular slug flow. By introducing this new pseudo slug flow regime, the pressure drop predictions in viscous oil systems became significantly more accurate than before.

ABSTRACT A set of air and highly viscous oil experiments was conducted to analyse the inner flow field of slug flow for low mixture Reynolds number conditions. The unsteady behavior of the wall shear stress profile was measured, modelled, and compared to an existing model. The Constant Temperature Anemometer (CTA) was utilized to measure the wall shear stress around the pipe circumference. Particle Image Velocimetry (PIV) was applied to investigate the inner liquid velocity field, simultaneously. A total of 9 flow conditions has been studied in a horizontal 50.8-mm (2-in.) inner diameter (ID) pipe and highly viscous fluid properties (nominal viscosity value of 0.68 Pa·s). From the slug front to the slug tail, a continuous increase of the velocity was observed by PIV, resulting in the increase of the wall shear stress in the liquid slug body at all the circumferential locations of the pipe as reflected by CTA. The wall shear stress decreased again from the tail of the slug into the liquid film region. This unsteady behavior of the wall shear stress profile was modelled by unsteady viscous flow theory. The proposed point model showed a fair agreement not only with the measured wall shear stress profile at the pipe bottom, but also along the pipe perimeter. On the other hand, the constant values provided by the existing correlation gave a deviation from the measured profiles.

ABSTRACT This paper considers the huge changes in the oil industry's understanding of multiphase flow over more than 4 decades, and the application of that understanding into the engineering discipline of flow assurance that has developed in the same timeframe. Initial test facilities were simple horizontal pipes carrying air and water at near atmospheric pressure and temperature. Then came larger, pressurised facilities using model hydrocarbon fluids. Now the oil industry has complex subsea production systems in deep water with reservoir fluids flowing over long distances and experiencing large changes in pressure and temperature. Fundamental understanding of the behaviour of the fluids and flow is key to the flow assurance discipline, and to the efficient design and operation of such production systems. Techniques for visualising and modelling the flow, required to manage the various characteristics of the fluids and flow behaviour, are discussed. Significant challenges remain for flow assurance before it can predict all aspects of fluid and flow behaviour sufficiently accurately for risk management in all applications. Opportunities exist in oil and gas field development projects, and in subsequent operations, to instrument and monitor flow and fluid behaviour both for the benefit of the operation and for validation and improvement of models used in design.

ABSTRACT This study presents experimental data for pressure gradient, liquid holdup, and flow regimes. To the knowledge of the authors, there are no studies in the literature on vertical downward two-phase flow in annulus. The experimental results for downward two-phase flow in the annulus is compared with downward flow in pipes. The comparison results show some differences between downward two-phase flow in annulus and pipes for the liquid holdup and flow regimes. This comparison shows a significant disagreement between flow in annulus and pipes for a wide range of conditions. The experimental results show that the liquid holdup is consistently higher for two-phase downward flow in annulus than in pipes for the annular flow regime, and these differences are as high as 45%. When the flow regime map for downward flow in annulus is compared to the ones in the literature for flow in pipes, intermittent flow is observed for flow in pipes with liquid velocities lower than the velocities for flow in annulus.

Annular flow is one of the most observed flow patterns in oil and gas production systems, thus understanding annular flow characteristics is of great importance. When gas flow velocity is low, depending on the rate of liquid, annular flow transitions to slug flow and an annular/pseudo-slug flow pattern is developed. The aim of the present study is to examine the annular/pseudo-slug flow characteristics such as void fraction using a state-of-the-art experimental approach before and after an elbow. Wire Mesh Sensor (WMS) has been used to determine the void fraction distributions in an upward vertical orientation before a 3-inch elbow and in a horizontal orientation after the elbow for different flow conditions. In lower gas velocities and higher liquid velocities and viscosities, churn-annular transition and pseudo-slug flow patterns were observed for vertical (before bend) and horizontal sections (after bend), respectively. On the other hand, increasing the gas velocity and decreasing the liquid velocity and viscosity cause the flow structure to shift toward the annular and wavy stratified patterns in vertical and horizontal orientations, respectively.

ABSTRACT Gas liquid slug flow can induce vibrations in process piping and large scale slugging can cause some oscillations of subsea flexible pipes. Pipe movements should be avoided, as fatigue can lead to pipe failure. Flow transients can also cause pipe failure, water slugs have lifted geothermal pipelines off support, and runaway hydrate plugs have caused fatalities. The flow – structure interaction is therefore of interest to study and to incorporate in dynamic simulators. Several cases of dynamic slug flow-structure interactions have been studied in small scale laboratory. A two way coupling has been established between a pipe structure model and a slug tracking model. A pigging case of a submerged flexible pipe is presented here. An initially liquid filled hose is hanging freely in a 4m long water tank. A pig is inserted at the inlet, and subject to a backpressure of air which drives the pig and the liquid out of the pipe. The subsequent pipe dynamics is video recorded and compared with the coupled flow-structure simulations. At high flow rates the initially U-shaped pipe rises above the water level and falls back to the surface. The experiments and the coupled models are described as well as key aspects of the numerical solvers.

ABSTRACT Experiments were performed with propose of mapping the flow patterns in two-phase flow in an upward inclinable 60 mm ID and 6 m long Plexiglas pipe. Air and viscous oil were used as working fluids and degree of inclination changed from 10 ? to 78 ?. Measurements includes pressure drop. Flow regime were recorded with high speed video cameras. Experimental results were compared with existing models by Barnea (1987) and with a commercial dynamic multiphase flow simulator (OLGA7.3 ® ). Predictions with unified model had acceptable agreement with experimental data. 1. INTRODUCTION The need of understanding two-phase gas-liquid flow has been increasing in the recent years, as well as the need of a technical solution for handling and controlling the flow behaviour. A correct prediction of multiphase flow dynamics is critical for the design of pipelines for deep-water riser and production tubing in offshore oil and gas field development, where parameters as pressure drop and liquid holdup are strongly dependent on the flow pattern. Generally, flow pattern in gas-liquid system can exist in a wide variety of forms, depending on the flow rate, physical properties of the phases, the geometry and inclination of the tube (McQuillan and Whalley, 1985).

ABSTRACT Based on the success of foam in subsurface applications it is of interest to investigate whether foam can also help overcome liquid management problems in surface flowline-riser systems. Therefore, flow experiments were carried out in the flow loop at the Shell Technology Centre Amsterdam. The facility consists of a 100 m long horizontal flowline (with 50.8 mm diameter) followed by a 16.8-m vertical riser (with 44 mm diameter). Air and water are the working fluids, and operation is at atmospheric outlet pressure. Foam is created by adding "Dreft™" (a dishwashing detergent) in various concentrations to the water/air flow. Experiments were taken both without and with foam. Various measurement techniques were used: differential pressure sensors, flow visualization, and Distributed Acoustic Sensors (DAS) (fibre optics). The focus is on (growing) slugs in the horizontal flowline, and severe slugging in the flowline-riser. It can be concluded from the small-scale lab experiments that adding a surfactant mitigates (growing) slugs in (nearly) horizontal flowlines, whereas the severe slugging cycle in a flowline-riser configuration cannot be broken.

ABSTRACT Liquid entrainment fraction is an essential parameter in the calculation of the pressure drop and liquid holdup in multiphase flow lines, as well as in separator sizing. Liquid entrainment has been studied for several decades and several models and correlations have been published for its prediction. This study focuses on the evaluation of select entrainment fraction prediction models. Experimental data on liquid entrainment with hydrocarbon fluids in a 3-inch horizontal pipe at 8,375 kPa are presented and compared with predicted values. Two model improvements are proposed and the differences between predicted and experimental values are reduced by up to 17%. 1 INTRODUCTION This work represents the continuation of a study previously presented by Viana et al. 2016(1). This Part 2 paper provides additional data and covers the prediction of entrainment fraction. The data presented in this document include entrainment fraction values obtained from testing in a 3-inch diameter horizontal pipe using hydrocarbon fluids at 8,375 kPa operating pressure.

ABSTRACT Wax deposition in the onshore and offshore pipelines is a common occurrence and typically hinders the transportation of oil and gas. Accurate prediction of wax deposition in multiphase flow is of great importance to develop economical prevention/remediation strategies. However, reliable field and experimental data is lacking to understand the deposition process. In this work, statistically reliable two-phase gas-oil stratified flow wax deposition data have been obtained using Garden Banks condensate and natural gas in a 52.5 mm large-scale multiphase flow loop under the pressure of 2.41 MPa. Unprecedented local deposit analysis in the circumferential direction provides insights to understand the deposition behaviour at various experimental conditions. From the overall deposit analysis, it is evident that both deposit mass and wax content increase as superficial gas and liquid velocities increase. In addition, the deposit mass per unit area and wax content have a strong correlation with the actual liquid velocity. The deposit mass and wax content increase in two-phase is more predominant compared to single-phase deposition. It should be noted that without proper understanding in multiphase deposition, utilization of the pseudo single-phase modelling with mixture properties for the two-phase deposition process can yield significant errors in the prediction of wax deposition.

ABSTRACT Scaling up of wax deposition rates from small scale laboratory flow loops to field pipelines is generally problematic due to uncertainties in currently available state of the art models. Uncertainties in modeling include the right choice of mass transfer mechanism, the extent of shear effects, and the choice of aging/internal diffusion parameter. Traditional approach has been to use a model to tune to small scale flow loop data, and use that model to predict deposition rates in the field – with the understanding that the field predictions will usually be conservative. This approach, however, can lead to wasted CAPEX and/or OPEX, and may not work for economically challenged/ marginal projects. To address this problem, we have built a large-scale flow loop (4" pipe ID, and 120 ft long test section) for testing wax deposition. A variety of experiments were performed changing the operating conditions to observe the effect on the deposition rate. Different wax deposition software, both commercial and R&D, were tested against the data. Experiments were also performed on a small-scale loop to study scale up issues. Based on this work, a more robust approach to scale up of wax deposition rates has been developed.

ABSTRACT Design and operation of gas-condensate pipelines presents many challenges to the flow assurance engineer and pipeline operator. Important challenges are accumulation during low production and the removal of liquid during a ramp-up operation. Previously the connection between the behavior during a ramp-up scenario and multiple solutions for the steady-state fully-developed approximation was shown in [1],[3]. In this paper we will study the behavior during a ramp-down scenario where a pipeline or a well starts to accumulate liquid. The focus will be on understanding the phenomena. We will show how the steady-state fully-developed approximation can predict the behavior of a full transient simulation and we will make comparisons with laboratory experiments for different ramp-down scenarios. 1 INTRODUCTION Simulation of multiphase gas-condensate pipeline transport with low liquid loading is a challenging task. Important properties are the pressure drop for high rates and liquid accumulation at low rates which contribute to determining the operational envelope of the field.

ABSTRACT This study investigates experimentally and theoretically the existence slugs for vertical co-current two-phase flow in large diameter vertical tubing and risers. The slug flow pattern implies the presence of Taylor bubbles separated by hydraulically sealed liquid slugs. Previous experimental studies such as Ombere-Ayari and Azzopardi (2007) [1] showed evidence of the non-existence of Taylor bubbles for large diameter pipes. Models developed to predict this experimental behaviour (Kocamustafaogullari et al., 1984 [2]; Jayanti and Hewitt [3], 1990 and Kjølaas et al ., 2017 [4]) suggest that Taylor bubbles may disappear at large diameters and/or high velocities. However, the previously proposed models do not properly account for the simultaneous effect of velocity and diameter on the stability of a Taylor bubble.

ABSTRACT This paper presents new experimental data of gas/liquid flows in an annulus in the transition from annular to churn flow, which is relevant to the phenomenon of liquid loading in gas wells. The focus of the study has been on the effect of eccentricity to the onset of liquid back flow and pressure gradients. The results show that an increase in the eccentricity is triggering liquid loading already at larger gas flow rates. Pressure gradients seem to be affected most when gap bridging by waves is occurring. 2 INTRODUCTION Liquid loading is a well-known problem that occurs when wet gas wells approach their end-of-life. Due to the depletion of the reservoir, its pressure decreases. Consequently, the gas flow rate in the well is also reduced. When only dry gas is present in the well, the reduced gas flow rate has no further effect. However, when a liquid phase is present (condensate and/or water) the reduced gas flow rate may, at some point, no longer be effective in transporting the liquids to the well head and they will accumulate downhole; this is called liquid loading (or well liquification). The onset of liquid loading is defined here as the onset of liquid back flow.

ABSTRACT The slug capturing technique, which is incorporated in the industrial PROMPT code is used to explore the influence of entrainment of gas bubbles at slug fronts, on slug characteristics such as slug length and frequency. The computations are compared with available field data. It is found that the amount of gas entrained within the slug can substantially affect slug lengths, increasing them (almost doubling) in some cases, while in another it can decrease them. The paper presents the basic methodology, the models governing the motion of the gas entrained within slugs and the results of comparison with the field data. 1. INTRODUCTION Intermittent flow, especially the slug flow regime, occurs very frequently in the transport of hydrocarbon fluids in pipelines in the oil and gas industry. The slug flow regime is usually undesirable since the intermittency of slugs causes severe adverse conditions. Firstly, the flow rates of the gas and oil arriving at the receiving equipment (separators and slug-catchers) can fluctuate widely thereby undermining the functioning of the equipment (e.g. separator flooding). To cater for this situation, processing facilities are usually designed with generous "safety margins" at the cost of capital, weight and size. Secondly, the flow intermittency results in highly unsteady loading on the piping system and processing equipment, which can result in serious failure due to metal fatigue. It is therefore important to be able to predict the onset of slug flow; it would be even more beneficial if slug characteristics, such as slug length and frequency could be calculated reliably as well.

As the flow rate in gas wells decreases, the liquid content of the well increases (wellbore load up). This can eventually cause the well to die. For many years, people have used the Turner correlation to predict the minimum rate the well can produce without loading. The Turner correlation is based on the gas velocity being too low to lift droplets from the well. Several investigators have found that the droplet entrainment mechanism is wrong, and that the actual mechanism for wellbore load up is the reversal of the liquid film on the walls of the pipe. The film reversal point corresponds to the transition between annular flow and wavy annular/churn flow. Using this criterion as the basis, the film reversal point was well correlated by use of a modified Froude number, which was shown to give good results for pipes between 4 and 200 mm (1.5 and 8 in). Using this approach, a film reversal point was found for angles from near-horizontal to vertical. The model was compared with published wellbore load-up data and was found to give as good or better predictions than the Turner model. The new model also properly accounts for pipe inclination, diameter, and surface tension effects.

ABSTRACT Several oil and gas production facilities handle fluids with some wax content and Wachs et al. (1) have estimated that about 20% of the world petroleum reserves are waxy oils. As precipitation of wax is temperature driven, one mitigation is to not allow the crude to achieve temperatures below the wax appearance temperature (WAT) through the use of thick insulation to maintain heat from the reservoir. The most common approach for managing deposited wax is pigging to physically scrape and remove the wax. Due to cost reductions in design or operational challenges (e.g., decommissioned and or plugged flow lines) an increase in unpiggable flow lines is occurring. This challenge is exacerbated in aging fields where reservoir temperatures and rates have declined resulting in more of the flowline operating below the WAT. For these cases, robust surveillance approaches are needed to ensure the wax within the flowline is managed and that alternative methods, like chemical solutions, are optimized. One of the methodologies for assessing wax build-up is through the use of hydraulic models for understanding the effective flowline diameter as a function of time and relating that to wax deposit build-up. In this work, a comprehensive evaluation using field instrumentation and hydraulic modelling tools to measure wax buildup in a shallow water, single line tie-back system is presented. The methodology was then used to optimize well cycling frequency and chemical usage resulting in significant OPEX reduction.