Search

The tracer method has been adopted to study stratified–dispersed flow in a horizontal pipe, 80 mm in diameter and 50 m long, operating at 5 Bar with nitrogen–water mixtures. The use of the tracer method in a horizontal pipe required the development of a specially designed test section, the related electronics and a data acquisition system. It has also been necessary to develop a tracer injection system, which has been designed in order to obtain uniform tracer concentration in the liquid film immediately after its injection. The main flow parameters which can be measured with the present experimental set-up are the circumferential distribution of the film height, flow rate and tracer concentration, the rates of droplet entrainment and deposition and the split of the liquid phase between the wall layer and the entrained droplets. The average tracer concentration data have been interpreted with a new three-field model of the liquid phase in the stratified–dispersed flow pattern. In the present formulation, the model holds for steady, fully developed flow conditions and is based on a one-dimensional description of the flow system. The data cover a limited number of flow conditions.

A test section consisting of a circumferential array of conductance probes has been developed to measure the thickness distribution around the pipe wall of a liquid layer flowing in near horizontal pipes. When the film thickness is known, the array can be employed to measure the local film flow rate by injecting a high conductivity tracer into the liquid flowing at pipe wall. The test section consists of a short pipe made of a non-conducting material installed in a flow rig designed to operate at an appreciable pressure (40 bar). The flow loop is made of metallic pipes connected to the electrical earth. The conductance probes are made of three parallel, rigid wires spaced along the flow direction and are used to measure the height or the electrical conductivity of the liquid layer. The three-electrode geometry is aimed at minimizing current losses toward earth. The simultaneous operation of all the probes of the array, without multiplexing, allows a substantial reduction of current dispersion and a good circumferential resolution of film thickness or conductivity measurements. The probe geometry may generate an appreciable disturbance to the gas–liquid interface. This aspect of the proposed method has been studied with an experimental and numerical investigation relative to free falling liquid layers.

Air-water slug flow at atmospheric conditions in horizontal pipes of 31 and 53 mm i.d. has been characterized experimentally with local (optical) and cross-sectional (conductance) probes. The objective of the work is to improve the description of slug flow and the closure relations required in mean kinematic slug flow models. Local measurements include the radial void fraction distribution in the slugs, the size of the dispersed bubbles in the slug and the aeration of the liquid layer under the slug bubble. From cross-sectional holdup measurements, slug lengths and frequencies are determined, as well as the length of the highly aerated mixing zone in the front of the slugs. Measured values of film holdup at the tail of the slug bubbles and of slug frequency are compared with a physical model of slug flow derived from the work of Dukler & Hubbard.

Air/water slug flow in 53 and 90 mm i.d. horizontal pipes has been investigated for a large range of gas and liquid velocities. For this purpose a special instrumentation with simultaneous data acquisition and analysis has been developed. This has been used to determine the mean slug characteristics (length, holdup and velocity) and their statistical distributions.

For fully-developed two-phase flows, maps that correlate experimental and semi-empirical expressions for flow regimes are widely used. For calculations of the various important two-phase flow parameters, this in turn requires correlations for various interfacial and wall interaction effects that are flow regime dependent. For many systems of practical interest, however, the evolution of flow regimes (such as slug flow in oil–gas pipelines) is of interest because the development lengths are long and flow regimes may change in regions where pipeline inclination changes due to the terrain. It is shown here that for slow transients in near-horizontal pipes, the one-dimensional multi-field model, when solved with sufficient resolution, does not require flow regimes to be specified or flow regime dependent closure relationships. The formulation predicts the development of flow regimes and various flow parameters without the need for maps, or the need to change closure relationships. To accomplish this, the model includes four fields, i.e. continuous and dispersed liquid, continuous and dispersed gas, as well as a set of appropriate closure relationships from the literature. For the main application considered here, i.e. slow transients in oil–gas pipelines, order of magnitude analyses indicate that certain inertial terms in the model are very small and can be neglected in comparison to the others. Advantage is taken of this to simplify both the structure of the mathematical problem and the solution procedure, which is sufficiently accurate that mass is conserved for each of the four fields. Furthermore, the calculations require high spatial resolution, so a fast, easily-parallelizable numerical procedure has been applied. The results indicate that the development of certain flow regimes, including transitions from bubbly to stratified flow and vice versa, slug flow including slug frequency and length, and the evolution of these parameters along a pipeline are well predicted by the model when compared to experimental data. As part of the validation it is also shown that the model predicts, without need to change closure relationships, flow regimes in fully-developed near-horizontal two-phase flows in good agreement with existing flow regime maps. This suggests that for slow transients in flows for which one-dimensional effects dominate, predictions can be made without requirements for flow regime maps and closure relationships that depend on them.

In stratified gas–liquid horizontal pipe flow, growing long wavelength waves may reach the top of the pipe and form a slug flow, or evolve into roll-waves. At certain flow conditions, slugs may grow to become extremely long, e.g. 500 pipe diameter. The existence of long slugs may cause operational upsets and a reduction in the flow efficiency. Therefore, predicting the flow conditions at which the long slugs appear contributes to a better design and management of the flow to maximize the flow efficiency. In this paper, we introduce a wave transition model from stratified flow to slug flow or roll-wave regimes. The model tracks the wave crest along the pipe. If the crest overtakes the downstream wave end before hitting the top of the pipe, a roll-wave is formed, otherwise a slug. For model validation we performed measurements in air–water horizontal pipe flow facilities with internal diameters of 0.052 and 0.06 m. Furthermore, we made numerical calculations using a transient one-dimensional multiphase flow simulator (MAST) which adopts a four-field model. The model presented in this paper successfully predicts the evolution of waves and their transition into either slugs or roll-waves. It also predicts the formation time of slugs and roll-waves with a satisfactory agreement.

A set of experiments has been performed to analyse gas-phase distribution in horizontal bubble flow. The experimental data have been obtained by means of conductance probes and include measurements of the local void fraction and the bubble diameter and velocity. Present observations show large variations of the local void fraction and the bubble size moving from the top to the bottom of the pipe, with local maxima located close to the upper pipe wall. Measurements also show that, at the boundary of the dispersed bubble flow regime, the bubble size, and therefore the surface available for heat and mass transfer, changes abruptly for small variations of the liquid or gas flowrate. This change allows the transition between the dispersed and the elongated bubble regimes to be identified with good precision. Present data make it possible to develop a new correlation for the maximum bubble size in dispersed bubble flow.

In vertical annular flow, the motion of droplets in the gas core is dominated either by a diffusion mechanism, when the droplet size is small, or by inertial effects when droplets are large. These two mechanisms have to be considered when predicting deposition rates. Furthermore, since droplet-droplet interactions influence droplet motion, a deposition model should also account for collisions and coalescence among droplets. After reviewing the available deposition models, the effect of coalescence on droplet motion is theoretically analyzed. The results demonstrate that coalescence extends droplets residence time in the gas core thus decreasing the deposition coefficient. On the basis of these results, a new deposition model which accounts for the two deposition mechanisms and includes the effect of coalescence is proposed and compared against existing experimental data.

The dispersion of particles in turbulent duct flow under the influence of electrostatic fields is studied using direct numerical simulation. In this new approach, particles are moved in the temporally and spatially varying turbulent flow field under the influence of electrostatic and gravitational body forces, as well as fluid dynamic drag. The simulations agree well with previously performed experiments (done in geometries typical of wire-plate and plate-plate electrostatic precipitators) not only in the overall collection efficiency of particles, but in particle concentration profiles at various axial locations in the flow direction. This gives confidence in the technique that may be used to study different precipitator geometries and flow field configurations, supplementing costly and difficult experiments. Furthermore, information is obtained at a much more detailed level than is possible via experiments, allowing insights into the mechanisms dominating particle collection.

A linear stability analysis has been performed to investigate the stability of liquid flow down an inclined circular tube. To this purpose, approximate solutions which describe laminar and turbulent steady flow down an inclined tube have been developed first. The stability analysis has then been performed by an integral method. The results of the present investigation indicate that, in general, flow in a tube is more stable than in a channel and, in particular, there is a value of the liquid height at which the flow is always stable.

The aeration of a liquid body advancing over a slow moving liquid layer in a pipe has been experimentally investigated with a conductance method. The 50 mm i.d. pipe section was close to horizontal and the fluids were air and water at atmospheric conditions. It has been found that net gas entrainment only occurs when the relative velocity between the advancing front and the liquid layer is greater than a limiting value. The rate of gas entrainment is proportional to the relative velocity between the advancing slug and the liquid layer, and to the interfacial width of the layer. Measured pressure drops, film heights and gas volume fractions in the advancing slug agree well with available correlations or with similar measurements taken under conditions of fully developed slug flow.

The derivation of a correlation for drop size in annular flow based on a mechanistic model is presented. Optimum values of four constants were obtained by a fit to measured data over a wide range of gas and liquid flow rates, physical properties and pipe diameter.

The main characteristics of CWF (coal-water fuel) sprays produced by an internal mixing twin fluid atomizer are analyzed. Drop size and concentration have been determined by a laser light diffraction method (Malvern) and a high-speed photographic technique. The combined use of different experimental methods, along with the numerical inversion of line-of-sight measurements, allows a reliable characterization of the spray structure along radial and axial coordinates. The mean drop size depends essentially on the dynamic head of the atomizing air, the surface tension, and the coal particle size distribution of the slurry fuel. No clear effect of the CWF's apparent viscosity has been detected. The equation developed to correlate mean drop size takes all relevant factors into account and was derived for a nozzle operating with sonic flow in the exit ports.

The void fraction in liquid slugs has been determined for air—water fiow in horizontal and near-horizontal pipes by a newly-developed conductance probe technique. A semi-empirical correlation has been developed and compared with the present measurements and available data. This correlation predicts reasonably well the observed effects of diameter, inclination and physical properties.

Measurements of the effect of liquid viscosity on the initiation of roll waves in a horizontal gas-liquid flow are presented. These results are interpreted by an analysis based on the calculation of the growth of long wavelength disturbances.

A mathematical description of droplet entrainment and deposition is presented. This enables both unidirectional deposition and interchange experiments to be analysed. This approach permits description of deposition by both the diffusion like mechanism conventionally used and by the direct impaction mechanism which has recently been identified. A criterion has been derived to differentiate between the two mechanisms of deposition based on the balance between the initial drop momentum and the drag force of a turbulent eddy. The transition criterion has been substantiated with experimental data derived from unidirectional deposition experiments. A preliminary examination of interchange data has been carried out.

The point of onset of large disturbance waves and droplet entrapment has been determined for the annular downward flow of air-water mixtures in a 24 mm ID tube. Experimental results are in a good qualitative agreement with the previous work by Webb and Hewitt1. These results indicate that 1the critical gas velocity at the onset of entrainment is lower than the value predicted by available correlations 2the critical gas velocity continuously decreases at increasing the liquid Reynolds number 3below a liquid Reynolds number of 270 no entrainment occurs even at gas velocities as high as 80 m/sec A correlation for the critical gas velocity which is based on a simple physical model of atomization, is also proposed.

A liquid hold-up gauge based on the measurement of the electrical impedance has been developed for application in gas-liquid pipe flow. The gauge consists of two ring electrodes mounted flush to the pipe wall. The impedance (capacitance or conductance) seen by the electrodes depends on the distance between them and on the liquid hold-up. For distances above three tube diameters, the impedance is independent of the flow configuration for all separated flow patterns and, with good approximation, also for intermittent flows. Moreover, capacitance or conductance are linearly related to the liquid hold-up. The impedance under bubble flow conditions closely follows the theoretical predictions due to Maxwell. Also for the other flow configurations (annular, stratified, intermittent) the results of static and/or dynamic calibration agree closely with theoretical models.

New measurements are presented for film height and pressure drop for vertical gas-liquid annular flows. Improved methods for predicting the film height and interfacial friction factor are developed for situations in which the liquid film flow rate is known.

The main characteristics of diffusion flames formed on a liquid fuel surface with a parallel oxidizer flow have been studied both analytically and experimentally. In the present paper fuel burning rates and temperature and velocity profiles have been determined for various liquid fuels. Experimental results arc in good agreement with the theoretical model previously developed by one of the authors, except for the velocity profiles at low air velocity, which display a pronounced overshoot close to the flame zone. Present results also indicate that main flame characteristics can be controlled by means of the temperature of the liquid fuel contained in the burner and kept constant for long times. Finally, the burner studied in this research appears to be particularly suitable for the analysis of combustion kinetics.

Measurements of liquid hold-up and pressure drop are reported for stratified flow in a slightly inclined (0.65° and 2.1°), 5 cm pipeline. Velocity profiles in the gas phase have been determined for a limited number of flow conditions. Semi-empirical correlations are proposed for the transition to slug flow, the interfacial friction factor and the liquid hold-up.

Resultats de l'analyse de films cinematographiques obtenus avec la methode de la lumiere parallele de Hewitt et Whalley. Correlation des resultats par une equation simple qui peut etre reliee a l'analyse de la rupture d'onde de Taylor

A mechanistic model of slug flow based on a set of mass and momentum conservation equations and a number of empirical closure relations is presented. It is then shown that a simplified version of the model, which physically corresponds to the flow of long liquid slugs (long slug model, LSM), can be immediately derived from the general model. The LSM is much simpler than the general model; at the same time, the predictions obtained in the two cases are very similar, when the comparison is limited to the computation of the pressure gradient and the mean liquid holdup. The LSM successfully correlates a large set of experimental measurements relative to three pipe diameters (18, 50, and 90 mm), four inclinations (0[degree], [+-]3[degree], 0.3[degree]), and two pipe lengths (17 and 34 m).