Sunteți pe pagina 1din 233

Lukasz Piwoda

Metering of Two-Phase Geothermal Wells Using Pressure Pulse Technology

Diploma Thesis Norwegian University of Science and Technology Department of Petroleum Engineering and Applied Geophysics July 2003

Acknowledgement i __________________________________________________________________________ I wish to thank Professor Jon Steinar Gudmundsson for being my supervisor. I am grateful for his enthusiasm, an ocean of suggestions, and excellent supervision throughout this work. Thanks for my supervisor in Poland Dr. Ing. Czeslaw Rybicki to recommend me for Erasmus Link to Norway scholarship and his efforts to enable my take on the study at NTNU. I am thankful to ING AG Leipzig and Norwegian University of Science and Technology for financing my scholarship. I wish to thank Mr Wolfgang Laschet from Office of International Relations for his help to organize my stay in Norway. I also want thank to Professor Danuta Bielewicz and Professor Jan Falkus for their efforts and engagements into international cooperation between universities, and for their appreciate help to surmount the official adversity. In addition I want to thank Jon Rnnevig, Kjell Korsan and Harald Celius from Markland AS, for their guidance into the computer simulations and suggestions towards the obtained results.

ii List of Contents __________________________________________________________________________ List of contents Acknowledgement .i List of contents .ii Nomenclature.v Abstract..1 Introduction ...2 1. Metering of Multiphase Wells 3 1.1 Introduction.....5 1.2 Overview of multiphase metering.......................5 1.3 Challenges and accuracy.............7 1.4 Metering techniques............9 1.5 MFMs Projects..........12 2. Pressure Pulse Technology 13

2.1 Pressure Pulse method...............14 2.2 Water-hammer effect.............14 2.3 Theory and equation......................15 2.4 Pressure surge in wellbores...............16 2.5 Mass and volume flowrates...............................................................17 2.6 Flow condition analysis.....................19 2.7 Concluding remarks..................................20 3. Geothermal Applications 27

3.1. Geothermal energy...........................28 3.2 Geothermal well flow........................................................29 3.3 Well performance .....................................................................32 4. Multi-phase Flow in Wells 37

4.1. Introduction..38 4.2. Main difficulties.. 38 4.3. Phase behaviour... 39 4.4. Definition and variables.. 42 4.5 Fluid properties.... 44 4.6. Flow patterns... 44 4.7. Pressure gradient..47 4.8. Multiphase flow models...48 4.9. Duns and Ros correlation for multiphase flow in oil wells..51 4.10. Duns and Ros modifications..51 4.11. Orkiszewski correlation for multiphase flow in geothermal wells....52

iii List of Contents __________________________________________________________________________ 5. Sped of Sound in Two-Phase Mixtures 54

5.1 Introduction 55 5.2 Compressibility of two-phase mixtures. 55 5.3 Compressibility of steam-water system 57 5.4 Acoustic velocity models .. 62 5.5 Attenuation mechanisms of sound wave 66 5.6. Concluding remarks. 68 6. Case studies 74

6.1 Calculation purpose75 6.2 Water-hammer and line packing in oil wells..75 6.3 Water-hammer and line packing in geothermal well106 7. Discussion 156

7.1 Multiphase flow correlations... 157 7.2 Acoustic velocity profile...158 7.3 Line packing 159 7.4 Size of the pressure pulse..161 8. Conclusions 9. References Appendix A Multiphase Metering Projects Appendix B Duns and Ros, Orkiszewski - Multiphase Flow Correlations 162 164 174 183

B.1 Duns and Ross Correlation...184 B.2 Orkiszewski Correlation...192 Appendix C Sound Wave Propagation Process in Steam Water Mixture Appendix D PipeSim 2000-Multiphase Flow Simulator 195 198

D.1 PipeSim Well Performance Analyses.199 D.1.1. Fluid Properties Correlations...199 D.1.2 Advanced calibration data....203 D.2 Profile model..205 D.2.1. Detailed model205 D.2.2. Simplified model.206 D.3 IPR Data.207 D.4 Matching option.208 D.5 VLP correlations and applications.210

List of Contents iv __________________________________________________________________________ Appendix E HOLA 3.1-Multiphase Flow Simulator 213

E.1 Introduction....215 E.2 Governing equations .215 E.3 The computational models of HOLA 3.1..218 E.4 Heat loss parameters..218 E.5 Wellbore geometry219 E.6 Feedzone properties...220 E.7 Velocities of individual phases..221 E.8 Productivity Index estimation222 Appendix F Simulation results in oil wells 226

F.1 Well A1..227 F.2 Well A2......234 F.3 Well B240 F.4 Well C....246 Appendix G Simulation results in geothermal wells 252

G.1 Well D1.....253 G.2 Well D2.....256 G.3 Well E1......259 G.4 Well E2......262 G.5 Well F1......265 G.6 Well F2......268

Nomenclature v __________________________________________________________________________

a acoustic velocity

A cross section area


B volume factor Cp specific heat capacity at constant pressure CV specific heat capacity at constant volume

d diameter

f friction factor
g - absolute gravity h enthalpy H liquid holdup ID inner diameter k permeability

K slip ratio
KS isentropic compressibility KT isothermal compressibility

L length

m mass flow rates


p pressure
PI productivity index

q volumetric flow rates


R individual gas constant Re Reynolds number RS solution gas-oil ratio Sm3 gas/ Sm3 oil S entropy

t time
T temperature

u velocity
WC - water cut V volume x mass fraction z - direction opposite to gravity

Nomenclature vi __________________________________________________________________________ Greek letters: void fraction water-oil volumetric factor specific heats ratio dynamic viscosity

v specific volume

density

Abstract 1 __________________________________________________________________________ Multiphase flow measurement is of vital importance in petroleum and geothermal industry. Overview of currently available metering techniques has been made in present work. Pressure Pulse method is a new developed method which propose a different approach to measure two-phase flow in wells. The pressure effects after rapid valve closure that built up the method were illustrated. The inspection of the types of geothermal reservoirs allowed characterizing typical parameters of high enthalpy geothermal well. The difficulties to predict the multiphase flow in wells are presented together with description of the definitions and variables that need to be calculated. Multiphase flow models were examined and two most appropriate correlations have been selected for oil and geothermal wells. The speed of sound in two-phase mixtures was calculated. The available models to estimate acoustic velocity were studied and verified with respect to their limitations. The compressibility of steam-water system under the well flow conditions, required for calculations was derived from thermodynamics definitions. The simulations were performed in PipeSim 2000 and HOLA 3.1 programs for oil and geothermal wells respectively, in order to demonstrate the Pressure Pulse method. The case studies include three different North Sea oil wells and likewise three typical high enthalpy geothermal wells. Inflow performance and tubing performance calculations allowed extending the calculation for different diameters and flowrates. The results are presented in form of the tables and plots. Obtained results for oil and geothermal cases were compared to each other. All parameters that affect the acceleration pressure (pressure increase after rapid valve closure) and pressure built up in wells are discussed. The work ends with conclusions towards the performed calculations and gives the assessment for possible application of the Pressure Pulse method to meter the flow in two-phase geothermal wells.

Introduction 2 __________________________________________________________________________ Pipe-flow mixtures of crude oil, gas and water are common in petroleum industry, and yet their measurements nearly always present difficulties. The traditional solution is first to separate the components of the flow, and then measure the flow rate of each component using conventional single-phase flow meters. This method is both inconvenient and expensive to use for well monitoring. In addition the separation is not accurate, about 10% (Millington, 1999). Current multiphase meters have similar accuracy, they employ the complex techniques, and some of them contain the dangerous radioactive materials as discussed in Chapter 1. In geothermal wells producing steam and water mixture under various operating conditions the capability accurately measure the flow is also of value importance for several reasons similar to petroleum industry. These are general evaluation of the geothermal reservoir under proper reservoir management, optimalisation of the wellbore design from well deliverability considerations and minimization of scale deposits in the wellbore (Ragnarsson, 2000). The background for this thesis work is a new method to measure multiphase follow (Gudmundsson and Falk, 1999; Gudmundsson and Celius, 1999), developed at Norwegian University of Science and Technology (NTNU). The metering method is simple, requires little space, and is cost effective with at least the same accuracy as the competitors (Gudmundsson and Celius, 1999). The multiphase flow meter is based on measurements of pressure magnitude and pressure build-up. The output is velocity and density of the gasliquid flow. This thesis work concerns multiphase metering, specifically pressure transients caused by a rapid valve closure in oil and geothermal wells. Pressure propagation in fluids is closely related to sound velocity. The acoustic velocity in two phase mixtures varies significantly from this in single liquid or gaseous phase, and depends on physical properties of every mixture constituents. Available models for acoustic velocity in two-phase mixtures need to be verified according to their limitations in order to find the most appropriate for particular calculations.

Introduction 3 __________________________________________________________________________ Multiphase flow is a complex, turbulent and highly nonlinear process, which can not be fully described mathematically due to increased numbers of flow parameters. Computer simulations base on semi-empirical correlations were developed in order to predict the pressure and fluid parameters changes across the wellbore. Calculating flowing pressure profiles in oil wells, phase transfer between oil and gas requires a rather simple treatment, and is accomplished trough the use of solution gas-oil ratio Rs relationship. In geothermal wells, however phase transfer between water and steam attains critical importance and calculations must incorporate the steam tables accurately. Pressure profile calculations for geothermal wells vary from those for oil well in another important aspect in that the temperature of the fluid must be computed precisely. This thesis describes how the Pressure Pulse method can be used to meter the flow in high enthalpy two-phase steam-water geothermal wells similarly to oil wells. The calculations performed aim to estimate the size of pressure pulse after the valve closure and determine the parameters affecting the early pressure build-up.

Chapter 1
Metering of Multiphase Wells

1. Metering of Multiphase Wells 5 __________________________________________________________________________ 1.1 Introduction Multiphase means a single component existing in a variety of phases such as steam, water and ice. In the oil industry multiphase refers to a stream of fluid containing a liquid hydrocarbon phase (crude or condensate), a gaseous phase (natural gas, and non hydrocarbon gases), a produced water phase, and solids phase (sand, wax, or hydrates). In general the quantities of solids produced are minimal and thus have less impact than the liquid and gas phases. In present thesis work some simplification will be made and two phase, liquid phase and gas phase will be considered. The mixture of two immiscible fluids will be termed as liquid phase, regardless to components number. The mixture of gases flowing together will usually, unless there is a large density difference and little turbulence, diffuse together and can be treated as single homogenous phase (McNeil, 1990). Multiphase measurement is the measurement of the liquid and gas phases in a production stream without the benefit of prior separation of the phases before entering the meter. 1.2 Overview of multiphase metering While two-phase and multiphase flows have been common throughout petroleum industry for many years, there has until very recently been little or no demand for real-time metering of such flows. Traditionally the problem was circumvented by separating the flow into its constituent components, which allowed straightforward single phase metering techniques to be used (Theuveny et al., 2001). This approach was very practical and effective, but did give rise to processing systems which were quite inflexible in terms of their capability to handle fluctuating flowrates, varying water contents, and changes in the physical properties of production fluids. However in the early years of offshore North Sea production this was not a major problem, and at that time - pre 1980 - there was little or no impetus to develop more sophisticated metering technology that could perhaps dispense with separation equipment and expensive metering facilities (Falcone et al., 2002).

1. Metering of Multiphase Wells 6 __________________________________________________________________________ During the 1980s the process of gradually declining oil production from the major North Sea fields started, so in the interests of operational cost effectiveness, there was a move to use existing platform based process plant for other production roles (Steward, 2003). To maintain production levels, smaller satellite fields which were previously uneconomic to produce on a stand alone basis, were tied back to existing platform based infrastructure. From a technological point of view this introduced a step change in the complexity of production. There were now numerous fields, typically with quite different oil properties, water contents and gas fractions, all being produced through process plant designed for the early years of single-well production (Theuveny et al., 2001). Furthermore, the water contents and gas fractions started to increase, and this exacerbated the production problems even further. It began to emerge quite quickly that more operationally flexible multiphase technologies were going to be needed, if not immediately, certainly within five to ten years (Steward, 2003). For existing platforms the prime purpose of this new technology would be to improve processing flexibility, and for new field developments the aim would be to completely eliminate the need for costly and bulky platform based process plant (Falcone et al., 2002). The ultimate aim was of course to move towards remote subsea instrumentation. The key driver at all times being lower production costs through reduced initial capital expenditure, and reduced operating manpower. To take up these challenges, the growth in multiphase research and development since the early 1980s has been exponential, especially with regard to metering, and today there are a variety of multiphase flowmeter (MFM) installed onshore and offshore. It appears to be no reduction in new metering developments (Steward, 2003). However the actual growth rate of installations has been lower than initial industry forecast suggested (Falcone et al., 2002). Oil companies have been hesitant to invest in expense meters with limited tracks record. Figure 1.1 shows actual trend up, tied with very low level of utilisation MFM technology before the 2000. The reasons may be the fact that when operators decide between a traditional approach to the production facilities and one including MFM, must compare the capital and operating expenses of each solution. Very little operational history of MFM cause difficult to predict the operating costs (Jamieson, 1999). This difficulty results from relatively low number of MFM applications worldwide, allow claiming that widespread implementation of MFM

1. Metering of Multiphase Wells 7 __________________________________________________________________________ cannot take place until expertise is spread more widely trough oil industry (Falcone et al., 2002).
900
Number of MFM istalations

800 700 600 500 400 300 200 100 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 0 1 5 0 2 5 0 10 3 1 17 6 5 29 24 22 52 65 20 66 78 43 103 129 55 210 542

Offsore subsea Offshore topside Onshore

Figure 1.1 Grow rates of MFM installations (Falcone et al., 2002) 1.3 Challenges and accuracy The level of difficulties in accurately measuring the multiphase stream is increased dramatically over single phase measurement. Single phase fluids can be quantified by knowing about the pressure, fluid density, viscosity, compressibility and geometry of the measurement device (Williams, 1994). Unfortunately, multiphase fluids do not act in the same manner as single phase fluids and above variables of each phase would not quantify multiphase flow. Multiphase flow is a complex, turbulent, highly non linear process. Williams (1994), and also King (1999) give the brief description of the processes that may take place as the different phases flows simultaneously. The phases interact with each other gas may evolve out of the

1. Metering of Multiphase Wells 8 __________________________________________________________________________ solution, or is absorbed into the liquid, waxes and hydrates may precipitate etc. If the single component exists in the two phases there is significant mass transfer and thus mixture quality may be considered variable. The components do not mix homogenously, and tend to remain separate, the water does not mix well with the oil, and gas remains separate from the liquid phase. Both phases flow at the different velocities. It is common for gas and liquid to flow at the different rates. Very complex flow regimes can exist and are dependent on the relative velocity of the phases, fluid properties, pipe configuration and flow orientation. The mentioned above and other relevant to the multiphase flow parameters are described in Chapter 4, which deals with multiphase flow in wells. The parameters definitions and relationship between them are given together with the correlations developed in order to predict the flow behaviour. Expectations of MFM performance in the early days were concluded and sometimes in the fiscal range of accuracy. Such levels of accuracy were, and never will be achievable by present technology (Steward, 2003). Over the last ten years, a gradually more realistic assessment of uncertainty capabilities has evolved. To date, no international regulations for MFM accuracy has been delivered. Varying level of accuracy requirements exists in multiphase measurement depend on how the information will be utilized. Essentially, three main accuracy requirements exist for metering multiphase fluids (Falcone et al., 2002): approximately 5 -10% for reservoir management, approximately 2-5% for production allocation, and approximately 0.25-1% for fiscal metering, are anticipated to be required.

However because of high complexity of multiphase mixtures it may be optimistic to claim that the above ranges of accuracy apply to any regime and for any chemistry of the fluids.

1. Metering of Multiphase Wells 9 __________________________________________________________________________ 1.4 Metering Techniques Under the multiphase flow circumstances the following parameters are required to compute flowrates of each phase: the cross-sectional area of the pipe occupied by each phase the axial velocity of each phase density of each phase.

The cross-section area and phase velocities give the volumetric phase flowrates. The product of phase densities and phase volumetric flowrates gives the phase mass flow rate. Unfortunately, at the present time there is no method of measuring phase fraction directly, they are derived from two independent measurements, coupled with the continuity equation which requires the sum of oil water and gas phase fraction to equal unity. Typically, two variables independent are the density of the entire flow, and the water content in the liquid phase. Once these are measured, some simple mathematical analysis allows the individual phase fractions to be calculated. With these technology limitations, projects aimed at developing multiphase meters have tended to adopt one of two metering strategies (Millington 1999): A set of sensors that take volume measurements, which when combined are capable

of isolating the individual phase fractions. A combination of flow models and velocity measurements are used to derive the phase velocities as functions of time. To determine densities, temperature and pressure are measured and assumed equal in all phases. A set of sensors which again take volume measurements, but which also require flow

to be conditioned such that only one mixture velocity measured is assumed to be required . The overall mixture density is considered representative of the three individual phases. Phase fraction data is required as above.

1. Metering of Multiphase Wells 10 __________________________________________________________________________ Following sensors and techniques are commonly used: Gamma Densitometers consist of radioactive source and detector, placed so that the beam passes trough the flow and is monitored on the opposite side of the multiphase mixture. The amount of radiation that is absorbed or scattered by the fluid is a function of both fluid density and energy level of the source. Typical radioactive sources used include isotopes of caesium, barium or americium. Single energy gamma sensors are those that incorporate only one source or monitor only one energy level from source. These devices are often used to measure the density of the multiphase mixture. Dual energy gamma sensors measure the absorption of two separate energy levels. The two energy levels are provided either by two isotopes or by a single isotope that has two discernible levels. If two energy levels are far enough apart, these two independent absorption measurements can be used to determine the oil, gas, and water phase volume fractions. The densitometers are frequently calibrated by filling the device with known fluids, typically gas (or empty pipe) and water. Capacitance Sensors measure the dielectric properties of fluid. Each sensor consists of a pair of metal plates or electrodes. These are mounted on the pipe wall or are otherwise located so that the fluid occupies the space between them. The capacitance of the fluid is measured by varying the voltage difference between the plates and measuring the resulting electric current between them. From the capacitance, the dielectric constant of the mixture can be calculated. Since the dielectric constant of the mixture is a known function of the composition, this information can be used to calculate the volume fractions of oil, gas, and water phases. This technique will work for mixtures in which the liquid (oil/water mix) is oil continuous. Since the water phase is a much better conductor of electricity, water continuous mixtures will effectively "short" the capacitance plates rendering the measurement ineffective. For water continuous liquids, an approach based on conductance is used (see below). Conductance / Inductance Sensors use an electrical coil around the pipe to induce a current in the flowing multiphase mixture. The magnitude of this induced current is related to the dielectric constant of the mixture, which can be used to determine the (mixture composition as with the capacitance and microwave sensors.

1. Metering of Multiphase Wells 11 __________________________________________________________________________ Microwave Sensors measure the dielectric properties to help determine the phase fractions of the multiphase mixture. The sensor consists of emitters and receivers (antennae) of electromagnetic waves in the MHz or GHz range (microwaves). The dielectric constant of the mixture is a function of both the frequency of the waves and the mixture conductivity. The measured dielectric constant is a volume weighted average of the individual phase dielectric constants. The conductivity and dielectric constant of the water phase is a function of salinity. As such, meters that use this technique either need brine salinity as a calibration variable or have some other way of estimating it on-line. Cross Correlation Techniques use two similar measurements, each in a different axial location in the pipe. By comparing the two measurements, the velocity of the flow feature is determined, for example, the time required for a bubble to travel between the two sensors. Implicit in this technique is a measurable amount of non-homogeneity in the multi phase flow. For this reason, many available meters require the Gas Volume Flow (GVF) to be within certain limits, far enough from the pure liquid (GVF = 0) and pure gas limits (GVF = l) that the flow does not appear homogeneous to the sensors. Gamma densitometers, microwave sensors, and capacitance sensors are used in MPM systems for cross correlation. Venturi Meters consist of a gradual restriction in the flow path, followed by a gradual enlargement. For single phase flows, the pressure drop across the restriction is a straightforward function of the velocity and density of the fluid. For multiphase flows, the analysis is more complicated. The gradual restriction in the flow path makes the Venturi meter slightly intrusive to the flow. Positive Displacement Meters (PD) rely on the metered fluid to rotate mechanical gears or rotors in the flow path. Each rotation of the rotor corresponds to a known amount of volume passing through the meter. PD meters are commonly used in single phase service. For full well stream production, risks due to erosion and blockage should be considered.

1. Metering of Multiphase Wells 12 __________________________________________________________________________ 1.5 MFMs Projects A very limited amount of information is available on MFMs performance. In the oil and gas sector where competition is always intense a black box MFMs packages are usually offered, where very little is unveiled. A brief description of some MFMs projects that are now commercially available is given in Appendix A, together with the tables containing comparison of the methods with regard to the techniques that are used for measurement purposes.

Chapter 2
Pressure Pulse Technology

2. Pressure Pulse Technology 14 __________________________________________________________________________ 2.1 Pressure Pulse method Multiphase metering in oilfield operation is of considerable interest in petroleum industry as described in Chapter 1. As the response for these needs new method called Pressure-Pulse has been developed at NTNU by Professor Gudmundsson. The method is based on the propagation properties of pressure waves in gas-liquid media. Waves generated in gas-liquid mixture flowing in a pipe at a speed of sound will propagate as pressure pulses (Gudmundsson and Celius, 1999). These effects called water-hammer and line packing are described precisely below. The method has been tested in several offshore platforms including Gullfalks A, Gullfalks B, and Oseberg B, with positive repeatable results similar to the theoretical models (Gudmundsson, Falk, 1999). Total of 800 tests were run on 12 different gravel packed wells. No negative effects were observed on the production system or the reservoir during the 11-month test period. The method has the advantage of being simple, low-cost, and gives the same accuracy as the competitors (Gudmundsson and Celius, 1999). Pressure is the easiest parameter to measure in the production of oil and gas. It can be measured in pipelines, flowlines and wellbores; at wellhead, chokes, manifolds, and separators. The widespread use of the quick acting valves in the oil industry to open, close, and control pipeline and wellbore flow, has made it possible to harness the information contained in the rapid pressure transients when a valve is activated (Gudmundsson et al., 2002). 2.2 Water-hammer effect The water-hammer effect can be caused by a rapid closure a valve in pipe line with flowing liquid. The immediate pressure increase created by the valve is referred as the acceleration pressure-pulse pa. Wylie and Streeter (1993) described how this increase in pressure travels in the pipe with the velocity of sound, and stop the flow as it passes. The instant the valve is closed, the fluid immediately adjacent to it is brought to rest by the impulse of the higher pressure developed at the face of the valve. As soon as the first layer is stopped, the same action is applied to the next layer of fluid bringing it to rest. In this manner a pulse wave of

2. Pressure Pulse Technology 15 __________________________________________________________________________ high pressure is visualised as travelling upstream at same sonic velocity. However in long pipe flows with high frictional pressure loss the accelerational pressure-transient is attenuated and does not stop the flow completely. Yet, since the fluid must stop by the valve, there is a continuous pressure increase near the valve also after is wholly closed. The name of these phenomena is line packing. Figure 2.1 illustrates the water-hammer effect. 2.3 Theory and equation Water-hammer phenomena, line packing and pressure pulse velocities are essential for the new multiphase method. Water-hammer pressure transient can be found using homogenous continuity equation at high pressure well conditions fluids are well mixed and thus homogenous continuity equation can be applied (Falk, 1999). Continuity equation
p p u +u =0 + a2 x x t

(2.1)

The equation may be rewritten in form


p t p u t +u =0 + a2 t x t x t

(2.2)

The characteristic pressure pulse velocity running upstream the valve is Thus,
p a 2 u u p =0 + + t a u t a u t p u = a t t

x = a u. t

(2.3)

(2.4)

2. Pressure Pulse Technology 16 __________________________________________________________________________ During quick valve closure the velocity jump is u = -u in a short period of time t. The water-hammer due this retardation is

p a = a u
This equation is generally known in literature as Joukowski equation. Momentum conservation principle is given as

(2.5)

u u u u 1 p dz +u + = f g t x x 2d dx

(2.6)

In steady-state turbulent pipe flow frictional pressure gradient is represented by DarcyWeisbach equation
p f

f u2 2d

(2.7)

where, f is the dimensionless friction factor. The frictional pressure gradient is made available to measure when the flow is brought to the rest after valve closure. The line-packing pressure increase, in liquid-only flow represents the pressure drop with distance in the pipeline. In two-phase flow line-packing is more complicated and in addition to frictional pressure gradient it contains also increase in waterhammer with upstream distance. In vertical gas-liquid wells pressure increase with depth and hence the water-hammer changes with depth (Gudmundsson and Celius, 1999).
2.4 Pressure surge in wellbores

Using a high sampling rate and high resolution pressure gauge, pressure buildup is possible to record. A typical pressure-pulse technology set up is shown in Figure 2.2. It contains a quick-

2. Pressure Pulse Technology 17 __________________________________________________________________________

acting valve and two pressure transducers A and B upstream of the valve taking samples in micro to mili seconds time period. Today technology can definitely provide such high sampling gauges. A valve is termed quick-acting if it closes completely before waves are reflected from up-stream or downstream. If there are reflections before valve is closed, the pressure on closing will be affected (Gudmundsson and Falk, 1999). The example of measured pressure from two transducers is shown on Figure 2.3. Pressure Pulse is measured at two locations spaced 83.35m up-stream a quick acting valve. The speed of sound may be estimated from cross correlation between the signals. In this example figure this is 170 m/s. By knowing the mixture density, acoustic velocity for the mixture and pressure increase due to acoustic term during a quick shut-in, mixture velocity may be calculated at the wellhead (Gudmundsson, 1999). The studies of Khokhar (1994) suggest that the phenomena like wellbore storage, skin effect, and phase redistribution that occur after the well shut have no effect on the pressure technique. Whereas, the pressure-pulse method dependents more on mixture composition and gas-liquid ratio of the well fluids which influence the acoustic velocity. The speed of sound in two-phase mixtures, and its dependency on fluid properties and PVT conditions will be investigated in Chapter 5 of this work. In the Pressure-Pulse method the sound speed can be determined from cross-correlation of two pressure signals from locations A and B, as indicated in Fig. 2.4. The testing of the Pressure-Pulse method on several North Sea fields has resulted in measurements that make this possible (Gudmundsson, Falk, 1999).
2.5 Mass and volume flowrates

The mass flowrate in a pipe of constant cross-sectional area can be obtained directly from the Joukowski water-hammer equation, when the sound speed is also determined from crosscorrelation of the measured delay time between two signals from transducers A and B. A kg a s

m = p a

(2.8)

2. Pressure Pulse Technology 18 __________________________________________________________________________

The continuity principle dictates that the mass flow rate at the valve is the same as the mass flow rate at other locations. Mixture density and the mixture velocity can be also obtained from the measurements f L p a
2 2

mix =

2 d a p f

kg m3

(2.9)

v mix =

2 d a p f kg 3 f L p a m

(2.10)

Knowing the density of individual phases of the fluid mixture, void fraction can be calculated

L mix L g

(2.11)

Flow rates in petroleum industry are traditionally expressed in volumetric quantities. The mass flowrate and the volumetric flowrate of the liquid are related trough relationship
m Sm 3 s

q=

(2.12)

Treating about volumetric flow rates requires volumetric factor to be taken into consideration. Volumetric factor B(p,T), indicates the effects of pressure and temperature changes, from reservoir to stock-tank conditions. Thus, volumetric flowrates for oil can be calculated as
Sm 3 q o ( p, T ) = o Bo ( p , T ) s

(2.13)

where oil density given by relationship

2. Pressure Pulse Technology 19 __________________________________________________________________________

o ( p, T ) =

o + g R s ( p, T ) kg
Bo ( p, T ) m3

(2.14)

The volumetric flowrates of gas


Sm 3 q g ( p, T ) = q g Bo ( p , T ) s where gas volume can be also expressed as

(2.15)

Sm 3 q g ( p, T ) = q o GOR R s ( p, T ) B g ( p, T ) s

(2.16)

where:
Sm 3 gas R( p, T ) - amount of dissolved gas in oil 3 Sm oil Sm 3 GOR gas oil ratio at standard conditions 3 m

If the oil produced contains water, also watercut WC [%] need to be known.
2.6 Flow condition analysis

The pressure profile in a pipeline can be used to detect and monitor solid deposits as shown on Figure 2.5. Deposits will change the frictional pressure drop in the affected interval both by change pipe roughness and by reducing the tubing diameter. This will show up as increase in the line packing gradient in the affected region. When the valve is activated the pressure is measured resulting in a pressure time log. The pressure - time log is then converted into pressure - distance log. Those give the location and extend of the deposits in a pipeline. Pressure Pulse testing can be also used in gas lift wells for flow rate metering and flow conditions analysis. An examination of the line packing pressures makes it possible to identify the location of gas injection points, and asses the status gas lift valves (Gudmundsson

2. Pressure Pulse Technology 20 __________________________________________________________________________

et al., 2002). Figure 2.6 shows an example of the simulations for three different valve locations. The bubble point depth may be identified from line packing as it appears with the peak on the time derivative plot. Figure 2.7 shows typical bubble point pressure response experienced during the fields tests.

2.7 Concluding remarks

The water-hammer theory treats pressure-pulse propagation in single phase flow, and has also been directly extended to multiphase flow. This theory is important for the new multiphase meter. Multiphase flow models like the drift flux model, the homogenous model and certain forms of the two-fluid models could predict pressure pulse. However, the assumptions of the one pressure in the one-dimensional two-fluid model are not appropriate. This thesis uses the homogenous model, where due to large pressure surge fluid homogeneity and continuity may be assumed. The multiphase models are described in Chapter 4 of this work that treats about multiphase flow in wells.

2. Pressure Pulse Technology 21 __________________________________________________________________________

Figure 2.1 Water Hammer Effect

Figure 2.2 Pressure Pulse setup

2. Pressure Pulse Technology 22 __________________________________________________________________________

Figure 2.3 Pressure Pulse measurements (Gudmundsson and Celius, 1999)

2. Pressure Pulse Technology 23 __________________________________________________________________________

Figure 2.4 Pressure Pulse technology principles

2. Pressure Pulse Technology 24 __________________________________________________________________________

Figure 2.5 Deposit appearances on line packing

2. Pressure Pulse Technology 25 __________________________________________________________________________

Figure 2.6 Simulation results for three different valve locations (Gudmundsson et al, 2001)

2. Pressure Pulse Technology 26 __________________________________________________________________________

Figure 2.7 Typical bubble point pressure response during the field tests.

Chapter 3
Geothermal Applications

3. Geothermal Applications 28 __________________________________________________________________________ 3.1 Geothermal energy

Geothermal energy is one of the cleaner forms of energy now available in commercial quantities. The use of this alternative energy source, with low atmospheric emissions, has a beneficial effect on our environment by displacing more polluting fossil and nuclear fuels. Thermal energy carried in the produced fluid can be used for direct heating in residential, agricultural, and industrial applications; or the thermal energy of higher temperature systems can be used to produce electricity. Rapidly growing energy needs around the world will make geothermal energy exceedingly important in several countries. For example in Iceland provides 50% of the total power supply, and 86% energy used for space heating (Ragnarsson, 2000). The production of electricity requires a greater concentration of energy than other applications. If hot fluid is available in great enough quantities, a geothermal power plant can be installed that uses the produced steam directly to drive a turbine generator system. Geopressured geothermal reservoirs are closely analogous to the geopressure oil and gas reservoirs. Fluid caught in stratigraphic trap may be raised to litostratic pressure due to overburden pressure. Such reservoirs are given fairly deep (over 2,000 m), so that the geothermal gradient can give temperature over 100oC (Grand, 1982). A number of such reservoirs have been found in drilling for oil and gas. These reservoirs derive their heat from the terrestrial heat flux, and are widespread throughout the world. It occurred not economic to exploit even the most favourable reservoirs for a long time, but over the last decade many projects arise to utilise their energy (Dickson and Fanelli, 2001). In some places over the world high temperature over 250 oC geothermal reservoirs occur. That heat source may be either an abnormal high geothermal gradient or volcanism nature. Those fields usually display surface activity when high temperature fluid systems transfer heat to the surface from crustal rocks heated by magmas and are mainly located in six countries: United States, Mexico, New Zealand, Philippines, Iceland and Italy (Gudmundsson and Ambasth, 1986).

3. Geothermal Applications 29 __________________________________________________________________________ 3.2 Geothermal well flow

High temperature geothermal reservoirs can be liquid and vapour dominated this mean that can have liquid only or steam-water feedzone (Gudmundsson, 1989). Steam dominated reservoirs are relatively rare, and most geothermal fields are water-dominated, where liquid water at high temperature, but also under high (hydrostatic) pressure, is the pressurecontrolling medium filling the fractured and porous rocks. When liquid water flows into a geothermal well, the water will remain liquid up the wellbore until reaching a depth where the pressure is equal to the saturation pressure. The pressure decreases as the water moves toward the surface and at this depth the liquid water will start to flash to form a steam. It will continue to flash until reaching the wellhead, surface pipeline, and eventually the steam separator. Beginning with liquid water, the first flashing results in comparatively small amounts of stream that flows as a bubbles trough a continuous column of water. With pressure drop towards the surface more steam evaporates and thus flow changes the regime into slug and steam continuous annular flow (Gudmundsson, 1989). The two-phase output from geothermal wells is piped to a separator to produce steam for electric power generation. The liquid water separated from the steam is disposed of at the surface or injected back into the reservoir. Reinjection of the geothermal liquid back into reservoir after use has a number of purposes. The most important ones being (Eliasson, 2001): disposal being use liquid without polluting the environment, sustenance of the reservoir pressure to counteract with drawn down and surface subsidence, mining of heat stored in hot formations simultaneously extend the useful life of the reservoir. The most common approach for measuring flow rate similarly to the petroleum industry where gas is separated from oil, in geothermal applications separator is also used, where steam-water mixture is separated into a flow of water and steam at the pressure of separator. The flow of each phase can be then measured individually using pressure differential devices.

3. Geothermal Applications 30 __________________________________________________________________________

High temperature wells are typically drilled in four stages: (Figure 3.1) a wide hole to a depth of 50-100 [m] into which is cemented the surface casing. a narrower hole to a depth of 200-600 [m] into which the anchor casing is cemented. a narrower hole still to a depth of 600 to 1,200 [m] which carries a cemented casing called the production casing finally the production part of the wells drilled into and/or trough the active aquifer. This part carries a perforated liner that is hung from the production casing reaching almost to the well bottom. On top (wellhead) the well is fitted with expansion provision and a sturdy sliding plate valve (master valve). It is also commingled to a muffler, usually of a steel cylinder fitted with an expanding steam inlet pipe to low down the fluid on entry. Steam capacity of these wells commonly range between 3 30 [kg/s] (1.5 MWe 15 MWe) (Eliasson, 2001). The drilling programs are of two types: Standard: Wide: Surface casing 22 nominal diameter in a 24 hole. Anchor casing 18 nominal diameter in a 20 hole. Production casing 13 3/8 nominal diameter in a 17 1/2 hole Liner 9 5/8 nominal diameter in a 12 1/2 hole. Surface casing 18 nominal diameter in a 20 hole. Anchor casing 13 3/8 nominal diameter in a 17 1/2 hole. Production casing 9 5/8 nominal diameter in a 12 1/2 hole Liner 7 nominal diameter in an 8 1/2 hole.

Wide tubing configurations has been initially implemented in order to cut down the frequency of wellbore cleaning due to calcium carbonate scale depositions (Gudmundsson, 1986). These wide 13 3/8 production casing has been particularly developed in Iceland and

3. Geothermal Applications 31 __________________________________________________________________________

the narrower 9 5/8 production casing is reported by many authors as the typical (Uphady et al., 1977), (Gudmundsson and Thrainsson, 1988) . In present work, data about the wells was taken from Icelandic sources for 13 3/8 production casing, from two fields Reykjanes and Svartsengi. Nevertheless the deliverability considerations, presented in next section allowed finding the operating parameters assuming the 9 5/8 production casing and thus simulations covered the both typical tubing sizes. Geothermal wells have total mass flowrates are greater than oil and gas wells, primarily due to width casing configuration presented above that allows yield such high mass flowrates. The calculations made in this work confirmed that diameter change from 13 3/8 to 9 5/8 allows yield almost double output. Typical exploitation parameters gained from literature are placed in the Table 3.1 Table 3.1 Typical exploitation parameters
Variables Range

total mass flowrate wellhead pressure wellhead temperature wellhead enthalpy well depth

12.9 - 68.6 [kg/s] 2.3 - 56.5 [bar] 150 - 250 [oC] 965 - 1966 [kJ/kg] 913 - 2600 [m]

Geothermal reservoir data inspection presented here is required to characterise typical geothermal well that can be used in pressure pulse simulations. Simulations will be done for various mass flowrates, and wellhead pressures. In order to predict the above parameters, deliverability method developed in petroleum industry and widely applied also for geothermal reservoir engineering is necessary.

3. Geothermal Applications 32 __________________________________________________________________________ 3.3 Well performance

The production of liquid water from a geothermal reservoir depends on the reservoir pressure, the flow of fluid trough the feedzone into the well, and then up the wellbore to the surface. These three elements of deliverability are called reservoir, inflow and vertical lift performance respectively. The production output test gives the deliverability at the time of testing. As the production proceeds with the time the deliverability is like to change because of drawdown in reservoir pressure. The prediction of the reservoir pressure with time is the subject of reservoir modelling, and is not necessary to be discussed here. The fluid entering flowing well in liquid dominated reservoir contains pressurized water. Nevertheless when well flowing pressure pwf decrease below saturation pressure psat a twophase mixture of steam vapor and liquid water flows into the wellbore as a result of flashing outside the wellbore. The flashing occurs over a relatively short distance near the wellbore. This indicates that rapid pressure drop and radial-flow effects in the wellbore region may control the output characteristics of geothermal well (Gudmundsson, 1986). The inflow performance curve for geothermal well is composed of two forms of flow behavior, depending upon whether the flowing pressure is above or below the saturation pressure of the geothermal fluid. Above the saturation pressure a linear relationship as assumed between the mass flowrate m and the well flow pressure pwf. The inflow performance curve for geothermal well is composed of two forms of flow behavior, depending upon whether the flowing pressure is above or below the saturation pressure of the geothermal fluid. Above the saturation pressure a linear relationship is assumed between the mass flowrate m and the well flow pressure pwf. In general the mass flowrate increase when pressure difference enlarges as can be expressed:

m = PI ( p r p wf
where:

kg ) s

(3.1)

3. Geothermal Applications 33 __________________________________________________________________________

kg m - mass flowrate s
p r - average reservoir pressure [bar ]

p wf - well flow pressure [bar ]


kg PI - productivity index bar s

This equation applies for single-phase Darcy flow into the wellbore. If the pressure in a near wellbore distance decreases below bubble point pressure the slope of inflow performance curve is assumed to become more negative. This indicates that when steam-water mixture enters the wellbore, the resistance to flow is grater than for liquid only flow for the same flowrate. It is like a solution-gas drive reservoir in petroleum industry, and thus equation from petroleum industry can be adopted, after some modifications. The orginal form of the equation for oil and gas is given below
1 2 k h B o o pr re 3 ln r +s w 4 kr o 2 k h B o o pb re 3 ln r +s w 4

qo =

( p r pb ) +

(p

2 b

p wf

) sm

2 pb

s
3

(3.2)

where:
k permeability [m2] h thickness [m] dynamic viscosity [Pas] re effective radious [m] rw well radious [m]

Using HOLA 3.3 simulator described in Chapter 6, it is possible to estimate Productivity Index PI for given wellhead flow conditions.

3. Geothermal Applications 34 __________________________________________________________________________

In Darcy law for two phase flow the fluids are assumed to flow practically independently of each other. The fundamental law is then applied to the two phase flow individually. In geothermal simulation studies of two phase reservoir flow, the relative permeability for steam and water need to be defined. The following expression gives the total mass flowrates:

k k dp kg m = A k rw + rs w s dL s w s psat

(3.3)

Thus, to calculate curve for liquid only feedzone, when well flow pressure pwf above saturation pressure, the following equation can be used:

m = PI ( p r p wf

kg ) s

(3.4)

And as the well flow pressure pwf decrease below saturation pressure the equation (3.3) can be used in form

2 2 p sat p wf kr w kr s m = PI ( p r p sat ) + PI + w s 2 p sat w s psat

sm 3 s

(3.5)

where: psat saturation pressure [bar] Gudmundsson et al. (1986) in their study of relative permeabilities give the necessary relations. The relative permeability ratio of vapor and water can be calculated from equation
Sw 1 w = K s 1 Sw

kr w kr s

(3.6)

3. Geothermal Applications 35 __________________________________________________________________________

where: K - slip ratio,


Sw water saturation;

Assuming that there is no interaction between the flowing phases, that is steam and water are assumed to flow independently, the retaliations can be made
kr w + kr s = 1

(3.7)

Relative permeability for steam and water can be found from following functions:
S w < 0.4,
kr w = Sw kr w = Sw
0, 6

For for and for

; ;

0.2 < S w < 0.4,

0,7

S w < 0.2,

kr w = Sw

0 , 77

The other way to calculate the two-phase performance curve part, below the saturation pressure is to use the Vogel empirical relationship obtained for the situation when gas is coming out of the solution (Gudmundsson, 1986).The equation has form
2

p wf m = 1 0 .2 p m max sat

p 0.8 wf p sat

(3.8)

Where mmax is the ideally maximum flowrate obtained assuming pwf = 1[bar] Vertical lift performance curves were calculated for particular wellhead pressures pwh, using HOLA 3.1 wellbore simulator. Then MATHLAB 6.5 program has been used to calculate the deliverability curves. The mass flowrate of steam and water from geothermal reservoirwellbore system is given by well operating point, determined by the intersection of the IPR and VLP curves.

3. Geothermal Applications 36 __________________________________________________________________________

Figure 3.1 Casing stage types

Chapter 4
Multiphase Flow in Wells

4. Multiphase Flow in Wells 38 __________________________________________________________________________ 4.1 Introduction

Two-phase flow occurs commonly in the petroleum, geothermal, chemical, civil, and nuclear power industries. In the petroleum and geothermal industry, two-phase flow is encountered in well production, transportation, processing systems. The complex nature of two-phase flow challenges production engineers with problems of understanding, analyzing, and modelling two-phase-flow systems. The calculation and prediction methods that are discussed in this chapter were developed for petroleum industry. Geothermal applications also base on this method, however due to different water and crude nature a different approach is required in some cases.
4.2. Main difficulties

When two or more phases flow simultaneously in pipes, the flow behaviour is much more complex than for single-phase flow. Phases tend to separate because of differences in density. Shear stresses at the pipe wall are different for each phase as a result of their different densities and viscosities. Expansion of the highly compressible gas phase with decreasing pressure increases the in-situ volumetric flow rate of the gas. As a result, the gas and the liquid phases normally do not travel at the same velocity in the pipe, upward flow the less dense, more compressible, less viscous phase tends to flow at a higher velocity than the liquid phase, causing a phenomenon known as slippage. However, for down flow, the liquid often flows faster than the gas. Perhaps the most distinguishing aspect of multiphase flow is variation in the physical distribution of the phases in the flow conduit characteristic known as flow pattern or flow regime (Brill, 1999). During multi-phase flow through pipes, the flow pattern that exists depends on the relative magnitudes of the forces that act on the fluids. Buoyancy turbulence, inertia, and surface-tension forces vary significantly with flow rates, pipe diameter, inclination angle, and fluid properties of the phases (Brill, 2001). Several different flow patterns can exist in a given well result of the large pressure and temperature changes the fluids encounter (Manabe et al., 2001). Especially important is the significant variation in

4. Multiphase Flow in Wells 39 __________________________________________________________________________

pressure gradient with flow pattern. Thus, the ability to predict flow pattern as a function of the flow parameters is of primary concern. Analytical solutions are available for many single-phase flow problems. Even when empirical correlations were necessary (i.e., for turbulent-flow friction factors), the accuracy of prediction was excellent. The increased complexity of multiphase flow logically resulted in a higher degree of empiricism for predicting flow behaviour. Many empirical correlations have been developed to predict flow pattern, slippage between phases, friction factors, and other such parameters for multi-phase flow in pipes. Virtually all the existing standard design method relies on these empirical correlations. However, since the mid-1970s, a dramatic advance have taken places that improve understand the fundamental mechanisms that govern multiphase flow. These have resulted in new predictive methods that rely much less on empirical correlations. This chapter introduces and discusses basic definitions for parameters unique to multiphase flow in pipes. Flow patterns are described in detail, including methods available to predict their occurrence. The use of empirical correlations based on dimensional analysis and dynamic similarity performed by software used in this work are presented.
4.3. Phase behaviour

Two-phase can be interpreted as a single component like a water and its vapour steam, and a complex mixture of various components like a hydrocarbons composition. Geothermal fluid or complex mixture of hydrocarbon compounds or components can exists as a single-phase liquid, a single-phase gas, or as a two-phase mixture, depending on the pressure, temperature, and the composition of the mixture (Campbell, 1994). Unlike to a single component or compound, such as water-steam system, when two phases exist simultaneously a multicomponent mixture will exhibit an envelope rather than single line on a pressure/temperature diagram. Figure (3.1) gives a typical phase diagram for a multicomponent hydrocarbon system. Shapes and ranges of pressure and temperature for

4. Multiphase Flow in Wells 40 __________________________________________________________________________

actual envelopes vary widely with composition. Figure (3.1) permits a qualitative classification of the types of reservoirs encountered in oil and gas systems. Typical oil reservoir has temperatures below the critical temperature of the hydrocarbon mixture. Volatile oil and condensate reservoirs normally have temperatures between the critical temperature and the cricondentherm for the hydrocarbon mixture. Dry gas reservoirs have temperature above the cricondentherm (Campbell, 1994). Many condensate fluids exhibit retrograde condensation, a phenomena in which condensation occurs during pressure reduction rather than with pressure increase, as for most gases (Firoozabadi, 1999). This abnormal or retrograde behaviour occurs in a region between the critical and the cricondentherm, bounded by the dewpond curve above and, a curve below formed by connecting the maximum temperate for each liquid volume percent. As pressures and temperatures change, mass transfer occurs continuously between the gas and the liquid phases within the phase envelope of Fig. 3.1. All attempts to describe mass transfer assume that equilibrium exists between the phases. Two approaches have been used to simulate mass transfer for hydrocarbons the "black-oil" or constant-composition model and the (variable) compositional model (Brill, 1999). Each is described in the following sections.

Figure 3.1 Typical phase diagram (Campbell, 1994)

4. Multiphase Flow in Wells 41 __________________________________________________________________________

Black-Oil Model

The term black oil is a misnomer and refers to any liquid phase that contains dissolved gas, such as hydrocarbons produced from oil reservoirs. These oils are typically dark in colour, have gravities less than 40 API (824.97 kg/m3), and undergo relatively small changes in composition within the two-phase envelope (William and McCain, 2002). A better description of the fluid system is a constant-compositional mode. For black oils with associated gas, a simplified parameter Rs has been defined to account for gas that dissolves (condenses) or evolves (boils) from solution in the oil. This parameter, Rs can be measured in the laboratory or determined from empirical correlations. Because the black-oil model cannot predict retrograde condensation phenomena, it should not be used for temperatures approaching the critical-point temperature. A second parameter, called the oil formation volume factor Bo also has been defined to describe the shrinkage or expansion of the oil phase. Oil volume changes occur as a result of changes in dissolved gas and because of the compressibility and thermal expansion of the oil. Dissolved gas is by far the most important factor that causes volume change. Oil formation volume factor can be measured in the laboratory or predicted with empirical correlations (Brill and Mukherjee, 1999). Once the black-oil-model parameters are known, oil density and other physical properties of the two phases can be calculated. When water also is present, solution gas/water ratio, Rsw, and water formation volume factor, Bw, can be defined. Brill and Mukherjee (1999) also give correlations for these parameters and physical properties of the water. The amount of gas that can be dissolved in water and the corresponding possible changes in water volume are much smaller than for gas/oil systems (William and McCain, 2002).
Compositional Model

For volatile oils and condensate fluids, vapour-liquid equilibrium (VLE) or "flash" calculations are more accurate to describe mass transfer than black-oil-model parameters. Brill and Mukherjee (1999) provide a description of VLE calculations. Given the composition of a fluid mixture or "feed," a VLE calculation will determine the amount of the feed that exists in the vapour and liquid phases and the composition of each phase. From

4. Multiphase Flow in Wells 42 __________________________________________________________________________

these results, it is possible to determine the quality or mass fraction of gas in the mixture. Once the composition of each phase is known, it also is possible to calculate the interfacial tension and densities, enthalpies, and viscosities of each phase. Brill and Mukherjee (1999) also give methods to predict these properties. VLE calculations are considered more rigorous than black-oil model parameters to describe mass transfer. However, they also are much more difficult to perform. If a detailed composition is available for a gas/oil system, it is possible to generate black-oil parameters from VLE calculations. However, the nearly constant compositions that result for the liquid phase and the increased computation requirements make the black-oil model more attractive for non-volatile oils (Brill, 1999).
4.4. Definition and variables

When performing multiphase calculations, single-phase flow equations often are modified to account for the presence of a second phase. This involves defining mixture expressions for velocities and fluid properties that use weighting factors based on either volume or mass fraction (King, 1990). When gas and liquid flow simultaneously up a well, the higher mobility of the gas phase tends to make the gas travel faster than the liquid. This is a result of the lower density and viscosity of the gas. The slippage between both phases in defined as the ratio of the gas velocity to the liquid velocity
uG uL

K=

(4.1)

where, uG gas velocity [m/s], uL liquid velocity [m/s]. The mass fraction of flowing phases is defined as the ratio of gas mass flowrate to the total mixture flowrate:

4. Multiphase Flow in Wells 43 __________________________________________________________________________


x= mG mG + m L

(4.2)

where, mG gas mass flowrate [kg/s], mG liquid flowrate [kg/s]. The gas mass flowrate is related to the volume flowrates with expression
mG = u G G AG

(4.3)

and similarly the liquid phase


m L = u L L AL

(4.4)

where AG and AL are the cross sectional area occupied by gas and liquid phase respectively. Under steady state condition the slippage between both phases result in a disproportionate amount of the slower phase being present at any given location in the well. Gas void fraction can be defined as the fraction of pipe cross sectional area occupied by gas. Substitution of the equations (4.3) and (4.4) to the equation (4.2) results in the void fraction given by
x

x + K G (1 x ) L

(4.5)

The opposite value to the gas void fraction is the liquid holdup defined similar way as the cross section area occupied by liquid or volume increment that is occupied by the liquid phase
H L = (1 )

(4.6)

The gas void fraction and liquid holdup can be distinguished in horizontally oriented pipes where stratification occurs due to gravity. In vertical wellbore two-phase turbulent flow under

4. Multiphase Flow in Wells 44 __________________________________________________________________________

high velocities, both phases may be considered as a homogenous mixture (King, 1990). Two phases may be assumed to flow at the same mixture velocity with no slippage between.
4.5. Fluid properties

A numerous equations have been proposed to describe the physical properties of gas/liquid mixtures. The following expression has been used to calculate in multi-phase flow mixture density

M = G + (1 ) L

(4.7)

The two phase viscosity is the property expressed per mass unit and thus was calculated from equation

1 x

(4.8)

When performing the temperature change calculations for multi-phase flow in geothermal wells, it is necessary to predict the enthalpy of the multiphase mixture. Also most VLE calculation method for oil wells includes a provision to predict the enthalpies of the gas and liquid phases. Enthalpy of the mixture was calculated from equation
ht = x hG + (1 x ) hL
4.6. Flow patterns

(4.9)

Prediction the flow pattern that occurs at a given location a well is extremely important. The empirical correlations or mechanic model used to predict flow behaviour varies with flow pattern (Gomez, 2001). Essentially all flow pattern predictions are based on data from lowpressure systems, with negligible mass transfer between the phases and with a single liquid phase (Brill, 1999). Consequently, these predictions may be inadequate for high-pressure,

4. Multiphase Flow in Wells 45 __________________________________________________________________________

high production-rates, evidently high-temperature geothermal wells, or for wells producing oil and water or crude oils with foaming tendencies, respectively (Manabe et al., 2001), (Gudmundsson and Ambastha, 1984), (Aggour, 1996). A consensus exists on how to classify flow patterns (Brill, 1999). For upward multi-phase flow of gas and liquid, most investigators now recognize the existence of four flow patterns: bubble flow, slug flow, churn flow, and annular flow. These flow patterns, shown schematically in Fig. (4.2) and Figure (4.3) are described next. Slug and churn flow are sometimes combined into a flow pattern called intermittent flow. It is common to introduce a transition between slug flow and annular flow that incorporates churn flow. Some investigators have named annular flow as mist or annular-mist flow. Flow in vertical and horizontal or inclined pipes exhibits different behaviour. The distribution of the multiphase contents across the pipe in vertical flow regimes is randomly chaotic, and the phases show no preferences for the one side of the pipe or another. The exception to the random distribution is annular flow where at very high flow rates gas occupies the centre of the pipe. There may be large discontinuities that pass along the vertical pipe or wellbore, as when gas flows much faster than liquid in slug and churn flow regime. In non vertical flow random distribution of the phases across the pipe is replaced by gravity segregation by the phases.
Bubble Flow

Bubble flow is characterized by a uniformly distributed gas phase and discrete

bubbles in a continuous liquid phase. Based on the presence or absence of slippage between the two phases, bubble flow is further classified into bubbly and dispersed - bubble flows. In bubbly flow, relatively fewer and larger bubbles move faster than the liquid phase because of slippage. In dispersed bubble flow, numerous tiny bubbles are transported by the liquid phase, causing no relative motion between the two phases.
Slug Flow

Slug flow is characterized by a series of slug units. Each unit is composed of a

gas pocket a plug of liquid called a slug, and a film of liquid around the bubble flowing downward relative to the Taylor bubble. The Taylor bubble is an axially symmetrical, bullet-

4. Multiphase Flow in Wells 46 __________________________________________________________________________

shaped gas pocket that occupies almost the entire cross-sectional area of the pipe. The liquid slug, carrying distributed gas bubbles, bridges the pipe and separates two consecutive Taylor bubbles.
Churn Flow

Churn flow is a chaotic flow of gas and liquid in e which the shape of both the

Taylor bubbles and the liquid slugs are distorted. Neither phase appears to be continuous. The continuity of the liquid in the slug is repeatedly destroyed by a high local gas concentration. An oscillatory or alternating direction of motion in the liquid phase is typical of churn flow.
Annular Flow Annular flow is characterized by the axial continuity of the gas phase in a

central core with the liquid flowing upward, both as a thin film along the pipe wall and as dispersed droplets in the core. At high gas flow rates more liquid becomes dispersed in the core, leaving a very thin liquid film flowing along the wall. The interfacial shear stress acting at the core/film interface and the amount of entrained liquid in the core are important parameters in annular flow.

Figure 4.2 Vertical flow patterns

4. Multiphase Flow in Wells 47 __________________________________________________________________________

Figure (4.3) Horizontal and inclined flow patterns


4.7. Pressure gradient

The pressure gradient equation for multi-phase flow can be modified from single-phase flow. Considering the fluids to be a homogenous mixture the equation may be written
du dp f M u M = + M g sin + M u M M dL dL 2d
2

(4.10)

For vertical flow = 90o, dL =dz and the equation for pressure gradient can be written as
dp dp dp dp = + + dz dz f dz el dz acc

(4.11)

4. Multiphase Flow in Wells 48 __________________________________________________________________________

The pressure-gradient equation for single-phase flow in pipes was developed by use of the principles of conservation of mass and linear momentum. The same principles are used to calculate pressure gradient for multiphase flow in pipes. However, the presence of an additional phase makes the development much more complicated. The pressure-drop component caused by friction loses requires evaluation of two phase friction factor. The pressure drop caused by elevation change depends on the density of the two phase mixture which may be calculated from equation (4.7). The pressure drop caused by acceleration component in normally negligible as is considered only for cases of very high flow velocities (King 1990).
4.8. Multiphase flow models

Early investigators treated multiphase flow as a homogeneous mixture of gas and liquid. This approach did not recognize that gas normally flows faster than liquid. The no slip approach tended to under predict pressure drop because the volume of liquid predicted to exist in the well was too small (Brill, 1999) Improvements to the no-slip methods used empirical liquid holdup correlations to account for slippage between the phases. Although liquid holdup and friction effects were often dependent on the flow pattern predicted by empirical flow-pattern maps, in general these methods still treated the fluids as homogeneous mixture (Falk 1999). Treating the fluids as a homogeneous mixture is often unrealistic, resulting in poor predictions of flow behaviour (Brill, 1999). A trend to improve flow-behaviour predictions has emerged that is a compromise between the empirical correlations and the two-fluid approach. The methods used to predict pressure gradient can be classified as empirical correlations and mechanistic models (Gomez, 1999). The empirical ones are based on experimental data, and are suitable for preceding steady-state flow. The mechanistic multiphase flow models include the two-fluid model, the drift-flux model and the homogenous model (Manabe et al., 2001). These can be developed from physical relationship with mass, momentum and energy conservation of each phase resulting in local, instantaneous equations. The mechanistic modelling approach still requires use of some

4. Multiphase Flow in Wells 49 __________________________________________________________________________

empiricism, but only to predict specific flow mechanisms or closure relationships (Gomez, 1999). The conservation laws are connected by interaction laws between the phases and between the fluid and wall. A popular approach is to average the conservation equations over the cross sectional area to get a one-dimensional model. The mechanistic models differ from each in how they implement the conservation laws (Falk, 1999). The two-fluid method uses one conservation equation for each phase, the drift-flux method uses the sum of the momentum equations in addition to energy and mass conservation for each phase, while the homogenous flow model uses only the sum of all phases for each conservation law. The homogenous flow model is a simplification, assuming the same flow velocity for all phases (Gould, 1970), thus needs neither interfacial friction nor drift flux terms. The empirical correlations can be placed in one of three categories (Brill,1999): I category - no slip, no flow pattern consideration. The mixture density is calculated based on the input gas/liquid ratio. That is, the gas and liquid are assumed to travel at the same velocity. The only a correlation required is for the two-phase friction factor. No distinction is made for different flow patterns. II category - slip considered, no flow pattern considered. A correlation is required for both liquid holdup and friction factor. Because the liquid and gas can travel at different velocities, a method must be improved to predict the portion of the pipe occupied by liquid at any location. The same correlations used for liquid holdup and friction factor are used for all flow patterns. III category - slip considered, flow pattern considered. Not only are correlations required to predict liquid holdup and friction factor, but methods to predict which flow pattern exists are necessary. Once the flow pattern is established, the appropriate holdup and friction factor correlations are determined. The method used to calculate the acceleration pressure gradient also depends on flow pattern.

4. Multiphase Flow in Wells 50 __________________________________________________________________________

The following list presented on Figure (4.3) gives the published empirical correlations for vertical upward flow and the categories in which they belong.

Method Poettmann and Carpenter Baxendell and Thomas Fancher and Brown Hagedorn and Brown Gray Asheim Duns and Ros Orkiszewski Aziz Chierici Beggs and Brill Mukherjee and Brill

Category I I I II II II III III III III III III

Figure 4.3 Published Vertical Flow Correlations Categories The following sections of this chapter present method to predict pressure gradients and presents the methods applied for calculations that were performed. In present work Duns and Ros, 1963 for oil wells and also Orkiszewski, 1967 for geothermal well methods was applied. Those authors summarized numerous investigations that have described flow patterns in wells and made attempts to predict when occur. Both correlations were verified many times

4. Multiphase Flow in Wells 51 __________________________________________________________________________

since the time that was developed by other authors and by the industry elaborating the software commercially available. Those methods are recommended for vertical wells calculations and contain modifications that improve accuracy. (PipeSim Manual), (Gudmundsson and Oritz, 1984) (Uphady, 1977)
4.9. Duns and Ros correlation for multiphase flow in oil wells

Duns and Ros method was chosen for oil wells calculations performed in present work. This

method is ranged to III group from Figure (4.3), which is assumed to give the most appropriate issues. The method is a result of an extensive laboratory study in which liquid holdup and pressure gradients were measured. About 4,000 two-phase-flow tests were conducted in a 185-ft (56,39m)-high vertical-flow loop. Pipe diameters ranged from 1.26 to 5.60 in. and included two annulus configurations. Most of the tests were at near-atmospheric conditions with air for the gas phase and liquid hydrocarbons or water as the liquid phase. Liquid holdup was measured by use of a radioactive-tracer technique. A transparent section permitted the observation of flow pattern. For each of three flow patterns observed, correlations were developed for friction factor and slip velocity, from which liquid holdup can be calculated. Duns and Ros performed the first dimensional analysis of two-phase flow in pipes. They identified 12 variables that were potentially important in the prediction of pressure gradient. Performing a dimensional analysis of these variables resulted in nine dimensionless groups. Through a process of elimination, four of the groups were identified as being important and were used to select the range of variables in the experimental program. Equations presented in Appendix B for this method gives those four groups.
4.10. Duns and Ros modifications

Two proprietary modifications of the Duns and Ros method have been developed but are not available in the literature. The first, known as the Ros field method, involved modifications based on carefully obtained data from 17 high-GOR vertical oil wells. In a joint Mobil-Shell study undertaken between 1974 and 1976, a modification resulted in the Moreland-MobilShell method (MMSM) (PipeSim Manual). In this study, 40 vertical oil wells, including the

4. Multiphase Flow in Wells 52 __________________________________________________________________________

17 used in the Ros field method, and 21 directional wells were selected as the basis for the modifications. The MMSM method includes liquid- holdup correlations derived from the data for bubble and slug flow that are simpler in form than those used in the original Duns and Ros method (Brill, 1999). Possible discontinuities at flow-pattern-transition boundaries also were removed.
4.11. Orkiszewski correlation for multiphase flow in geothermal wells

For Geothermal Two-phase flow calculations Orkiszewski correlation was used. This method was recommended by Uphadhay and Hartz (1977). Their work contains comparison of calculated and observed flowing pressure profiles for geothermal wells located in the United States and Philippines. Comparisons were included for tubular flow as well as flow trough the casing-tubing annulus. Their work revealed that for tubular flow, the Orkiszewski correlation makes the best prediction, whereas for annular flow, no clear choice of correlation can be made. Similar work has been done by Ambastha and Gudmundsson (1986), they measured the flowing pressure profile data from many geothermal wells around the world, covering a vide range of flowrate, fluid enthalpy and wellhead pressures. The authors reported a good accuracy of Orkiszewski correlation in estimating the downhole conditions. The capability to accurately predict flowing pressures in geothermal wells producing steam and water a mixture under various operating conditions is of value importance for several reasons similar to petroleum industry. These are general evaluation of the geothermal reservoir under proper reservoir management, optimalisation of the wellbore design from well deliverability considerations and minimization of scale deposits in the wellbore (Ragnarsson, 2000). The predictive capability is especially important because of the difficulty of running flowing pressure surveys in geothermal wells. These wells are characterized by very high fluid velocities, which sometimes makes impractical for pressure recorders to traverse downward in the well. There have been the cases of pressure recorders thrown out of the wellbore due to fluid velocities (Uphadhay, 1977). In calculating flowing pressure profiles in oil wells, phase transfer between oil and gas

4. Multiphase Flow in Wells 53 __________________________________________________________________________

requires a rather simple treatment, and is accomplished trough the use of solution gas-oil ratio Rs relationship. In geothermal wells, however phase transfer between water and steam attains critical importance and calculations must incorporate the steam tables accurately. Pressure profile calculations for geothermal wells vary from those for oil well in another important aspect in that the temperature of the fluid must be computed precisely. Orkiszewski tested several published correlations with field data and concluded that none was sufficiently accurate for all flow patterns (Orkiszewski, 1967) He then selected what he considered to be most accurate correlations for bubble and mist flow and proposed a new correlation for slug flow. Orkiszewski used the Duns and Ros flow-pattern transition for the boundaries between slug flow and mist flow, including the transition region between them. Equation defined these are given in Appendix B together with Duns and Ros correlation description. For the boundary between bubbly flows, he chose these criteria established by Griffith and Wallis.

Chapter 5
Speed of Sound in Two-Phase Mixtures

5. Speed of Sound in Two-Phase Mixtures 55 __________________________________________________________________________ 5.1. Introduction

A book on physical acoustic (Trusler, 1991), defined sound as infinitesimal pressure waves propagating trough a medium with a characteristic speed; the velocity of sound depending on media. In present work a pressure wave caused by a rapid valve closure travels at the velocity of sound trough the two phase mixture. This velocity is dependant on the compressibility and densities of both phases. These flowing phases may have different structures; usually it is of the gas bubbles or slug surrounded by the liquid phase. The gas present in liquid phase cause a marked increase of damping. The examination of this phenomena showed that this damping is due to increase of distortion of the liquid separating the bubbles. The pressure variations act almost entirely on the volume of gas and scarcely at relative incompressible liquid (Firoozabadi, 1999)
5.2. Compressibility of two-phase mixtures

The velocity of sound is defined as the square root of the derivative of pressure with respect to the density at constant entropy (Henry et al. 1977)

p a= S

(5.1)

The second law of the thermodynamics tells that process must be isentropic due to there is no temperature gradient except inside the wave itself. Therefore instead of differentiate the density, the sonic velocity can be related to the properties of the fluid. Using isentropic compressibility Ks the speed of sound formula can be written
1 KS

a2 =

(5.2)

5. Speed of Sound in Two-Phase Mixtures 56 __________________________________________________________________________

The isentropic compressibility of a single fluid Ks is defined as


1 v K S = v p S ,n

(5.3)

where v is the specific volume [m3/kg] and n is the composition vector, which is defined by n = (n1, n2, n3, , nc), where c is the total number of components and ni is the number of moles of each component i of the mixture. The expression for the two-phase gas liquid mixture compressibility will be 1 v M K S M = v M p S ,n

(5.4)

vM is the total specific volume of the gas liquid phases in the mixture. Similarly, the isothermal compressibility of a two-phase multicomponent system is defined by 1 v M K T M = v M p T ,n

(5.5)

The isothermal compressibility represents the volume change caused by small change in the pressure of the closed system at constant temperature. The isothermal and isentropic compressibility in the single phase state are related by a simple expression (Firoozabadi, 1999) KT = Cp CV KS (5.6)

Where Cp and CV are the heat capacities [kJ/kgK] at constant pressure and constant volume, respectively. The derivation of (5.6) can be found in (Firoozabadi, 1999).

5. Speed of Sound in Two-Phase Mixtures 57 __________________________________________________________________________

Since Cp Cv, then KT KS. In reservoir engineering applications isothermal compressibility is often used to describe the fluid compressibility away from the wellbore. Inside the well, due to expansion, the fluid may undergo heating or cooling and process may become nonisothermal. If the heat loss can be neglected, the isentropic compressibility may better represent the pressure and volume changes. Practicing engineers in order to obtain the two-phase compressibility often use the following relationship (Firoozabadi, 1999)

KT M = KT G SG + KT L SL

(5.7)

where SG and SL are the volumetric fractions of the gas and liquid components, respectively. The equation is invalid where there is mass transfer between the phases (Firoozabadi, 1999) what is of vital importance for water and its vapor system discussed in further section.
5.3. Compressibility of steam-water system

Calculations of the sound speed in two-components two-phase systems is an easy procedure if adiabatic equation of state data are available to calculate compressibility, because pressure and temperature may be considered to be independent variables in such systems. Calculation of the sound speed in one-component two-phase system is more difficult matter because the pressure and temperature are not independent variables and are related by the Gibbs equation for equilibrium between both phases (Kieffer, 1977). The complex physical process which occurs during propagation of sound wave in watervapor two-phase system is given in Appendix C. Propagation of the sound wave in the fluids is accomplished by compression or rarefaction. If steam-water system remains in thermal equilibrium where variables characterizing the system may be considered to follow the saturation line, there must be mass transfer between the phases, since the fraction of steam in mixture changes due to evaporation or condensation process. When the pressure wave passes the adiabatic compression causes pressure increase in both phases; as a result the water

5. Speed of Sound in Two-Phase Mixtures 58 __________________________________________________________________________

becomes subcooled, and the steam becomes superheated. The induced temperature difference between the steam ad water phases leads to the heat transfer from superheated steam to subcooled water (Kieffer, 1977). Depending on the original phase composition, whether steam or water was the dominant phase heat transfer will cause water evaporation or steam condensation. These heat and mass transfer cause that both water and steam are restored to the saturation line. Similar process takes place in cause of pressure reduction. In general consideration the evaporation and condensation can not take place instantaneously, since transportation of heat and mass can only occur at a finite speed. The time period in flashing water to steam or condensing steam to water is important in determination the degree of equilibrium obtained in the sound wave. Since condensation and evaporation generally proceed at different rates, it should be expected that compression and rarefaction waves behave differently. Experiments have confirmed this theory that finite amplitude rarefaction waves in steam water mixtures have lower velocities than compression waves because rarefaction waves tend to maintain continuous equilibrium (McWilliam and Duggins, 1969). A mixture of liquid and its vapor may respond to pressure disturbances by equilibrium and nonequilibrium state. Nearly high frequencies waves the process of pressure wave propagation follows fast and may be considered adiabatic where there is no equilibrium between the phases. In this case the mass transfer can be neglected and thus calculations of speed of sound are greatly simplified. The pressure pulse waves are low frequency (Gudmundsson and Celius, 1999; Falk 1999) thus equilibrium response needs to be considered and the mass transfer between the liquid and its vapor occurs in a time short with comparison to the acoustic wave period. Due to complex physical process related to the acoustic wave propagation in one-component system different approach than for two-component need to be considered in order to find steam-water compressibility. Experimental data on sonic velocity in steam-water system at geothermal wellbore conditions are not readily available. To solve this problem theoretical calculation need to be carried out to find the compressibility of this system for ideal thermodynamics conditions. The method to derive the compressibility of a steam-water mixture in contact with reservoir rock was elaborated by Grant and Sorey (1979). Similar

5. Speed of Sound in Two-Phase Mixtures 59 __________________________________________________________________________

approach was adapted to geothermal wells by Gudmundsson et al., (1998). The studies below are based on the work of these authors. Two-phase fluid in geothermal well can be considered to be in thermodynamics equilibrium where the flow is homogenous, steady-state and one dimensional. Assuming adiabatic conditions no heat loss and gain a balance equation can be written for steam-water mixture flowing from one infinitesimal cross section to another (Gudmundsson et al., 1998) x1 hG 1 + (1 x1 ) hL 1 = x 2 hG 2 + (1 x 2 ) hL 2 (5.8)

where x represents the mass fraction of the steam, and hG and hL [kJ/kg] are the enthalpy of steam and vapor and liquid water, respectively. The equation can also be written as hL 1 + x1 hLG 1 = hL 2 + x 2 hLG 2 (5.9)

the hLG [kJ/kg] is the latent heat of vaporization which may be assumed to change negligible between adjacent cross-sections; from one infinitesimal cross-section to another. Thus the equation (5.17) can be rewrite in form hL 1 + hL 2 = (x 2 x1 ) hLG (5.10)

Since for liquid water a change in enthalpy is equal to the heat addition at constant pressure h Cp = T p

(5.11)

the equation (5.17) can be written as


C p L T = ( x 2 x1 ) hLG

(5.12)

5. Speed of Sound in Two-Phase Mixtures 60 __________________________________________________________________________

When pressure is lowering the dominant volume change is that caused by phase change; flashing liquid into steam vapor. As water before changed phase has occupied a volume mvW, after that occupies larger volume being in gas phase mvG. This increase of volume can be written as V = m (vG v L ) The mass of water that changed phase is simply
m = ( x1 x 2 ) m

(5.13)

(5.14)

Therefore the change in mixture volume becomes V = m ( x 2 x1 ) (vG v L ) (5.15)

As shown in previous section the compressibility may be defined in several ways depending on what physical property is assumed to be constant, temperature, enthalpy or entropy. The question arises what conditions may be found as the fluid flows in the wellbore. Usually as defining compressibility the temperature is assumed constant, giving an isothermal compressibility. For two-phase wellbore flow without heat loss or gain adiabatic process seems to be most appropriate. From thermodynamics we know that an isentropic process is both adiabatic and reversible. The flow in pipes, pipelines and wells, frictional pressure loss makes the process non-reversible. This aspect of fluid flow is particularly important in situation where rapid pressure drop occur for example in nozzles (Watters, 1978). It may be considered less important in situations where the pressure changes gradually with distance for example if wellbores and long pipelines. Since the sonic velocity of steam-vapor flowing in the well may be approximated to isentropic conditions for which the acoustic velocity is defined in equation (5.1).

5. Speed of Sound in Two-Phase Mixtures 61 __________________________________________________________________________

Taking the compressibility from equation (5.3), and eliminating (x2-x1) from equation (5.23) by substituting equation (5.20), the compressibility may be written in form (Gudmundsson et al., 1998) 1 m C p L T (vG v L ) VM p hLG

Ks =

(5.16)

The fraction represents the total mixture density

m = M defined in equation (4.7). VM

At all condition fluid maintain equilibrium between the phases and follows the saturation line thus
p p = T T sat

(5.17)

Substituting from equation (4.7) the compressibility is

Ks =

C p L (vG v L ) ( G (1 ) L ) p hLG T sat

(5.18)

p may be calculated using steam tables what bring considerable inconvenience in T sat

evaluating compressibility numerically. This can be avoided using Clausius-Clapeyron equation


hLG p = T sat (T + 273.15) (vG v L )

(5.19)

5. Speed of Sound in Two-Phase Mixtures 62 __________________________________________________________________________

Compressibility expressed in this form is convenient for numerical calculations. The thermodynamics properties values can be obtained from wellbore simulator output file and then sonic velocity can be calculated.
5.4 Acoustic velocity models

The formulas derived for acoustic velocity base on experiments performed and varies with respect to the components that was applied and flow patterns encountered during the experiments. The flow regimes are described in chapter 4 of this work which deals with multiphase flow in wells. Also the fact is important, whether homogeneity can be assumed or not. For homogenous flow the slippage between the phases may be neglected. The study of different models that was made here has the purpose to find the most appropriate model for sonic velocity in one component steam-water mixture and two component gas-oil mixtures. In present thesis acoustic velocity in steam-water mixture was calculated from Wood (1941) equation. He derived the equation for the velocity of sound in a homogenous two component media, based on air-water experiments (Wood, 1944). The author proposes instead of differentiating the density of liquid mixture as shown in (5.1) relate the sonic velocity directly to the properties of the gases and liquids. This model assumes that overall compressibility of the mixture is related to the compressibility of the constituents by the relation K S M = K S G + K S L (1 ) (5.20)

Substituting the compressibility of single phases from equation (5.2) yields 1

aWood =

G + (1 ) L G aG 2
+ 1

(5.21)

L aL 2

5. Speed of Sound in Two-Phase Mixtures 63 __________________________________________________________________________

where is the void fraction and is the density, with the subscripts G and L indicating the gas and liquid, respectively. Nakoryakov et al. (1993) provided that Woods equation can be derived from equation (5.1) assuming the mass fraction of each phase reminded constant what is not such a good assumption for steam-water system. Another theoretical limiting case for vapor-liquid had the changes in density exclusively included by variation in the gas-mass fraction due to flashing. In experiments however practically nobody has manage to observe disturbances propagating with this flashing velocity (Falk, 1999). Semenow and Kostern (1964) found experimentally sonic velocity in steam water flow agreeing well with Woods estimate, and Noryakov et al. (1993) claimed that acoustic of Wood was the low frequency limit for bubbly flow when the wave process were isothermal. For high frequencies the wave process was adiabatic without energy transfer leading to the frozen velocity not treated here. (Nakoryakov et al., 1993). The dependence of propagation velocity on flow regimes demonstrates analytical expression derived by Henry for bubbly flow using a slip flow model (Falk, 1999). 1

a Henry =

(1 (1 K ))

K G + (1 ) L

1 K + 2 G aG L aL 2

(5.22)

The author claimed that interfacial momentum transfer exhibited a strong influence on the propagation velocity, and that different flow structures would have different interfacial drag during the passage of the wave. Assuming homogenous flow with sleep K=1, equation (5.29) reduces to Woods equation (5.28). However formula for acoustic velocity in slug flow overestimates the results giving velocities close to the velocity of sound in gas phase (Falk, 1999). As the slug flow is mostly encountered flow pattern in flowing geothermal wells, this model may bring the wrong results. There are a few examples of analytical models for sonic velocity in gas liquid mixtures in literature (McWilliam and Duggins, 1969; Kiefer, 1977; Firoozabadi 2000). These models

5. Speed of Sound in Two-Phase Mixtures 64 __________________________________________________________________________

are based on equation (5.1) and relate the density of the mixture to the densities gases and liquids. To make the differentiation in equation (5.1) possible, the ideal gas law was employed for the gas and adiabatic state equation for the liquid. The derived expression for the density of mixture was then differentiated with respect of the pressure. Henry et al. (1974) in the book of thermodynamics properties of hydrothermal systems showed that steam express the significant deviation from the perfect gas under high pressure conditions. Such high pressures may be expected in geothermal wells treated in this work, thus these models also do not occur to be the most applicable. It is well known that properties of the natural gas also deviate from perfect gas for high pressures, thus from the same purpose acoustic velocity in oil - gas mixture was calculated using the formula developed by Gudmundsson and Dong (1993). The authors similarly to Wood relate the sonic velocity directly to the properties of the gases and liquids (Gudmundsson and Dong, 1993). Their formula presented below, was developed for gas-oil mixture and assumes that liquid phase may contain water. The thermodynamics compressibility and densities of mixture components are related assuming the phases are homogenously distributed in the liquid. From equation (4.7) the density of gas/liquid mixture is given by

M = G + (1 ) L

(5.23)

where is the gas/liquid void fraction, and the subscripts G and L stand for gas and liquid, respectively. If the liquid phase contains both water and oil, the density of the liquid phase can be similarly obtained from

L = W + (1 ) O

(5.24)

where is the water-oil volumetric fraction and the subscripts W and O stands for water and oil, respectively.

5. Speed of Sound in Two-Phase Mixtures 65 __________________________________________________________________________

Substituting (5.24) into (5.23) gives the density of gas-oil-liquid mixture

M = G + (1 ) [ W + (1 ) O ]
The relationship (5.7) can be written for: gas-oil-water mixture as

(5.25)

KTM = KTG SG + KTO SO + KTW SW

(5.26)

where the S volumetric fraction of the component is in fact the void fraction and SO and SW is the liquid holdup (1-). Thus (5.26) can be alternatively written as

KT M = KT G + (1 )KT L

(5.27)

If the liquid contains water and oil, its isothermal compressibility can be expressed as K T L = K T W + (1 ) K T O (5.28)

For gas-oil-water mixture, the heat capacities at constant temperature and constant volume may be expressed as

CpM = xCpG +(1 x) CpL = xCpG +(1 x) yCpW +(1 y) CpO


and

(5.29)

CVM = xCVG +(1 x) CV L = xCVG +(1 x) [yCVW +(1 y) CVO]

(5.30)

5. Speed of Sound in Two-Phase Mixtures 66 __________________________________________________________________________

Where x is the gas-liquid mass fraction and y is the water-oil mass fraction. It should be noted that mass fraction instead void fraction is employed in these equations because the ratio of specific heats is a property based on mass unit (Gudmundsson, 1993). From equations (5.2) and (5.6) speed of sound for mixture is given as

a2 =

M KT M

(5.31)

where is the ratio of specific heats Cp CV

(5.32)

Substituting (5.25), (5.26) and (5.32) to equation (5.31), speed of sound for gas-oil-water mixture can be written as

a=

[x C

[ G + (1 ) ( W

pG

+ (1 x ) y C p W + (1 y ) C p O /[x CV G + (1 x ) ( y CV W + (1 y ) CV O )] + (1 ) O )] [ K T G + (1 )( K T W + (1 )K T O )]

)]

(5.33)
5.5 Attenuation mechanisms of sound wave

When sound spreads trough the media there are a diminution mechanisms of intensity or attenuation as the distance from source increase due to loses of mostly frictional character. All media capable of transmitting sound, are limited in extend and sooner or later the wave must stop or change from a medium to another. Thermal conductivity and shear viscosity attenuate the wave motion. Both the speed and attenuation of the sound waves depend on the frequency. General considerations say that pressure waves will pass through structures if they are of the frequency for which the

5. Speed of Sound in Two-Phase Mixtures 67 __________________________________________________________________________

wavelength is larger than the structure. If the structure is a bubble or a slug in gas-liquid flow, a pressure pulse needs to have a wavelength greater than the size of the bubble and slug to propagate through the gas-liquid flow (Gudmundsson and Celius 1999). In multiphase flow there are additional attenuation mechanisms (Falk, 1999) which include: viscous drag steady interfacial drag added mass (interior effect) boundary layer around a particle/drop (Basset force) interfacial heat exchange compressibility of each phase concentration gradient effects phase transitions deformation and fragmentation of bubbles and drops reflections at the interface

Knowledge of which of these attenuation mechanisms are the most important can greatly simplify the models for sound wave propagation. The relative importance, however, depends on the media, flow pattern and the frequency. In approach towards zero frequency where all processes take place slowly, the compression and expansion of the fluid occur reversibly and adiabatically. Trusler (1991) claimed that boundary-layer absorption was the most important attenuation mechanism for low-frequency waves in tubes with single phase flow. However, at high frequencies the fluid cannot maintain local equilibrium. Consequently, some of the energy is dissipated and high frequency waves will be highly attenuated by reflections of the interfaces and can not propagate very far in bubbly and slug flow (Falk, 1999). The dominant frequency in lowpressure air-water flow in pipelines has been shown to be in the range 1-10 Hz (Dong and Gudmundsson 1993). Similar results were reported by Falk (1998). Therefore, pressure

5. Speed of Sound in Two-Phase Mixtures 68 __________________________________________________________________________

waves in gas-liquid flow are infrasonic (lower frequency than audible sound which is about 20,000Hz). Numerical calculations in bubbly flow showed that the slippage between the phases has little influence on the wave structure (Falk, 1999), though it can only be neglected when the liquid viscosity is considerable higher than of water, or if the bubbles are small. In bubbly flow the main attenuation mechanism was heat exchange between the gas bubbles and surrounding liquid. Kiefer (1977) showed addition theory that added liquid mass is important for the pressure wave propagation. Calculations of speed of sound in bubbly flow (McWilliam and Duggins, 1970) and (Kiefer, 1977) showed that surface tension is important for small bubbles, while at large bubbles only liquid compressibility is important. Firoozabadi (2000) also revealed that that capillary pressure may affect the two-phase compressibility only in porous media at reservoir conditions outside the wellbore where phases interface is curved.
5.6 Concluding remarks

1. The acoustic velocity in steam-water two phase systems is more complicated than for twocomponent systems. In two-component system gas being dissolved under high pressure comes out of solution as the pressure decrease and is characterized by solution gas-oil ratio Rs. In one component system gas emerges due to evaporation process and if the pressure change is not of the high frequency the equilibrium is maintained between the phases. As the system responses with equilibrium for pressure change the temperature effects and mass transfer between both phases is essential for calculations. 2. The study of models made in this chapter showed that there is not one best model for acoustic velocity in one component steam-water mixture. The Wood equation was chosen for calculations. The choice of this model is mainly due to Semenow and Kostern (1964) experiments that showed a sound velocity in steam water flow agreeing well with those estimated from Woods model. However the other authors show the limitation of this model in many areas. The speed of sound in oil gas mixtures was calculated from formula

5. Speed of Sound in Two-Phase Mixtures 69 __________________________________________________________________________

Developed by Gudmundsson and Dong (1993). This formula is reported by authors to be in good agreement with measured data. Both chosen models require the properties of each constituent phase to be known. The procedure necessary in order to calculate these properties was fully described in this chapter for both gas-oil and steam-water cases. For oil and gas two-phase flow thermodynamics compressibility can be readily obtained from simulations performed. Unfortunately software available for geothermal simulations is not such sophisticated as those available for petroleum applications and thus steam compressibility need to be calculated from equation derived in this chapter. 3. The acoustic velocity in liquid single-phase is higher from acoustic velocity in pure gas phase due to significantly lower compressibility of liquid phase. At the presence of only one percent by volume gas in the form of gas bubbles the acoustic velocity decreases dramatically and two-phase system presents different character from each of the constituent phases. The explanation is the fact that such two-phase system has density of liquid and compressibility of gas. The calculated acoustic velocities for gas-oil and steam-water mixtures for wide range of void fraction values are presented on Figure 5.1 and Figure 5.2. The plots also shows that depression in the sonic velocity is less if the pressure is increased while retaining the same void fraction, although the sensitivity to change in pressure becomes progressively diminished. 4. The plots of acoustic velocity vs. void fraction for oil-gas and steam-water systems revealed different behavior for high void fraction values. For hydrocarbons two-phase fluids mixture the plot has less slope which becomes more flat as the pressure is higher. For steamwater mixture this shape is more inclined towards vertical. The calculations made showed out that this effect may be important for lower pressure values where water dryness has higher values and steam occupies main volume of the pipe. This high value void fraction changes has more effect on acoustic velocity than pressure changes and thus acoustic velocity decrease with pressure increase across the wellbore. This problem will be presented precisely together with calculations made in Chapter 8.

5. Speed of Sound in Two-Phase Mixtures 70 __________________________________________________________________________

5. Gudmundsson and Dong model relate the sonic velocity directly to the properties of the gases and liquids in the same manner as Woods model. In fact assuming that mixture do no contain water and = 0, the equation may be expressed in the same form as Woods. The difference is in the approach towards the thermodynamic process that occurs during the sound propagation. Compressibility is often defined as the small volume change than occur in closed system at constant temperature. The second law of the thermodynamics tells that sound propagation process must be isentropic due to there is no temperature gradient except the wave itself. Woods proposed to calculate speed of sound separately for both phases from isentropic compressibility. Gudmundsson and Dong equation use the isothermal compressibility and then transform the equation to the isotropic condition using the specific heats ratio as shown in equation (5.31). The possible error due to assuming isothermal process instead isentropic is up to 7% for gas-oil mixtures and even up to 14% for steam-water depending on the pressure. Table 5.1 contains calculated results for pressure equal 52.4 bar, for gas-oil mixture and Table 5.2 shows similar calculations for steam-water mixture at the 45.1 bar. The differences between calculated sound speed values assuming whether isothermal or isentropic process are also shown in form of plots by Figure 5.3 and 5.4 for gas-oil and steam-water respectively.

5. Speed of Sound in Two-Phase Mixtures 71 __________________________________________________________________________

1200 1100 1000 900 acoustic velocity [m/s] 800 700 600 500 400 300 200 100 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

p = 100.47 p = 52.36 p = 19.45

0.9

void fraction

Figure 5.1 Calculated acoustic velocities Vs void fraction for oil-gas mixture

1200 1100 1000 900 acoustic velocity [m/s] 800 700 600 500 400 300 200 100 0 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7 0.8

p = 45.2 [bar] p = 30 [bar] p = 15 [bar]

0.9

Figure 5.2 Calculated acoustic velocities Vs void fraction for steam-water mixture

5. Speed of Sound in Two-Phase Mixtures 72 __________________________________________________________________________

Table 5.1 Calculated acoustic velocities for gas-oil mixture.


oil-gas mixture, p = 70.2 [bar] 0.1 0.3 0.5 0.8 0.9 a [m/s] isentropic 268.3 176.4 158.5 178.4 207.0 a [m/s] isothermal 250.8 164.8 148.5 169.0 198.6 difference [m/s] 17.5 11.6 10.0 9.4 8.4 % 6.8 6.7 6.5 5.4 4.1

800

700

isentripic p = 45.1 [bar] isothermal p = 45.1 [bar]

600 acoustic velocity [m/s]

isentropic p = 70.2 [bar] isothermal p = 70.2 [bar]

500

400

300

200

100

0 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7 0.8 0.9 1

Figure 5.3 Calculated acoustic velocity for gas-oil mixture assuming isothermal and isentropic process of sound propagation

5. Speed of Sound in Two-Phase Mixtures 73 __________________________________________________________________________

Table 5.2 Calculated acoustic velocity for steam-water mixture.


steam-water mixture, p = 19.4 [bar] 0.1 0.3 0.5 0.8 0.9 a [m/s] isentropic 282.7 187.3 170.9 205.5 258.0 a [m/s] isothermal 247.1 163.2 148.8 179.6 226.8 difference [m/s] 35.6 24.1 22.1 25.9 31.2 % 13.4 13.8 13.8 13.5 12.9

800

700

isentropic p = 19.4 [bar] isothermal p = 19.4 [bar]

600 sonic velocity [m/s]

isentropic p = 52.4 [bar] isothermal p =52.4 [bar]

500

400

300

200

100

0 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7 0.8 0.9 1

Figure 5.4 Calculated acoustic velocity for steam-water mixture assuming isothermal and isentropic process of sound propagation

Chapter 6
Case Studies

6. Case Studies 75 __________________________________________________________________________ 6.1 Calculation purpose

The purpose of these calculations is to estimate water-hammer and line packing effects when valve has been fully closed. The results from these calculations should give a picture of these phenomena in two-phase wells. Then similar calculations for oil and geothermal wells may be compared witch each other. A description of the effect should give a result which can be linked to field measurements to see if any other factors might affect the pressure gradient after the valve closure. Offshore Pressure Pulse tests have been made in several North Sea wells to validate the theoretical simulation. The tests issues were described in two confidential reports and then published (Gudmundsson and Falk, 1999), (Gudmundsson and Celius, 1999). Geothermal wells are planed to be tested for Pressure Pulse method in summer 2003. The water-hammer effect is caused by rapid closure a valve in flowing pipe. The pressure increase is dependant on fluid density, velocity and sonic velocity in the flowing well as shown in Chapter 3. These will be calculated here. In liquid flow, line packing is the increase in pressure caused by a wall friction when a valve is closed and the fluid is stopped by a pressure wave which is emitted from the valve in the instant of closure. In liquid/gas twophase flow, the increase in pressure is the sum of pressure drop due to wall friction and the pressure drop due to interfacial friction. The interfacial friction is an unknown quantity, while the frictional pressure drop can be considered a known quantity.

6. Case Studies oil wells 76 __________________________________________________________________________ 6.2 Water-hammer and line packing in oil wells

The three programs used were PipeSim 2000, GOW 3.0 and Excel. PipeSim is a multiphase flow simulator. The futures and modes of the program used in present work are described in Appendix D. PipeSim simulates the flow of an oil/gas/water mixture in a well or pipeline. The program takes the well fluid data and uses them to calculate various properties of the fluid and system. PipeSim was also used for inflow performance (IP) and tubing performance (VLP) in order to estimate the flowrates for different wellhead pressures. The GOW program is published by Gulf Publishing Co. Houston, Texas. It allows calculating the various parameters of many substances including oil, gas and water. The PipeSim output file does not give the values of oil and gas compressibility and these were found from this program for given PVT conditions taken from well profile. The Excel spreadsheet was used in present work for additional calculations of acoustic velocity. The acoustic velocity was calculated using formula reported by Dong and Gudmundsson (1993) and given in equation (5.40) in this work. Excel was also used to plot the results. The above programs may be substituted by similar programs. Prosper Multiphase Flow Simulator, designed by Petroleum Experts Ltd. Edinburgh gives similar output files suitable to present calculations. Only the PipeSim was available at universitys computers thus calculations placed in this thesis work hail from PipeSim. Excel may be replaced by almost any spreadsheet that can plot and calculate. Also GOW has equivalent programs. To run PipeSim the fluid, completion and production data are necessary. The data used in these simulations for one well was obtained from work of Jonsson (1995). This contains the data about production, completion and fluid properties from the Draugen oil field. Two other wells were simulated based on data from work of Falk (1999). These data were gained during the Pressure Pulse Method tests on Gullfaks and Oseberg platforms. In addition some

6. Case Studies oil wells 77 __________________________________________________________________________

reference values for these oil fields were taken from the Skjveland and Kleppe Monograph (1992) that contains the characteristics of the most North Sea fields. PipeSim was used to determine the properties of the oil and gas at different depths in the producing well. Simulations gave the values at 50 meter intervals. Oil pressure, temperature, density, void fraction, gas and oil heat capacities and compressibility are the properties required to calculate the acoustic velocity. The last two were computed in GOW and then entered into the Excel spreadsheet. The void fraction is also not given directly in PipeSim output file but as water cut is 0 (assumed in all simulations done), the void fraction and the liquid holdup add to one. Excel was used to plot the results. The depth, acoustic velocity and the pressure drop due to friction were entered into spreadsheet. The frictional pressure drop is available on the PipeSim output file. These values are then used to calculate travelling time for a pressure pulse down to a certain point and up again. Calculated results were plotted against each other. Depth and time are plotted on the x axis and total pressure drop and acoustic velocity on the y axis. The total pressure drop versus time gives the effect of line packing. PipeSim gives the output file compatible with excel format thus the results are presented in two forms; Excel spreadsheet and Excel plots. Excel spreadsheet contains the calculated values of various properties affecting the acoustic velocity and line packing listed in columns and is available from all simulated oil wells in Appendix F of this work. These values were directly used to plot the results. Table 6.1 contains the data about the wells that was chosen to illustrate the pressure pulse method in oil wells. These data include well depth, tubing inner diameters, SCSSV depth and inner diameters, reservoirs pressures and temperatures necessary to run simulations in PipeSim. Table 6.2 contains the molecular composition of the well fluids used in simulations. PipeSim base on the reservoir data to predicts the well flowing pressure at the bottom of the well and then simulate the flow in vertical well. Table 6.3 presents the calculated parameters. The wellhead pressures, temperatures and flowrates are presented in the table together with parameters that make up water-hammer: density, acoustic velocity and flow velocity.

6. Case Studies oil wells 78 __________________________________________________________________________

Table 6.1 Well geometry and reservoir data for computer simulations

WELL SYMBOL

A1 1893 650 165 71

A2 1893 650 165 71

B 1952 4.5 none none 254 73

C 1924 5.125 none none 314 75

Depth H (m) Tubing ID (inch) SCSSV depth (m) SCSSV ID (inch) Reservoir pressure pres (bar) Reservoir temperature Tres (oC)

6.184 6.184 5.963 5.963

Table 6.2 Molecular composition of the well fluids (measured as a mole fraction of the gas phase at stock-tank conditions)

WELL SYMBOL

A1

A2

B 0.307 0.996 4.045 0.881 0.556 0.511 0.662 0.293 1.014 2.869 4.064 3.257

C 0.320 0.620 3.83 0.94 0.57 0.56 0.6 0.28 0.92 2.28 4.05 3.36

N2 CO2 C1 C2 C3 iC4 nC4 iC5 nC5 C6 C7 C8 C9 C10+

0.090 0.090 0.280 0.280 6.14 4.60 0.92 2.31 0.99 1.35 2.04 3.15 3.35 2.18 6.14 4.60 0.92 2.31 0.99 1.35 2.04 3.15 3.35 2.18

46.49 46.49 46.323 44.63

25.54 25.54 34.222 37.04

6. Case Studies oil wells 79 __________________________________________________________________________

Table 6.3 Calculated values at the inlet and wellhead condition

WELL SYMBOL

A1 156 71 7300.0 75.144 8.95 0.827 142.5 18.438 87.7 2.30


3 3

A2 158.5 71.0 5125.0 52.755 20.35 0.601 322.4 8.782 102.7 2.91

B 258.7 72.0 2165 22.708 123.5 0.210 616.2 2.906 288.4 5.16

C 228.97 72.85 1024.9 10.794 65.12 0.252 623.5 1.607 192.9 1.93

Inlet pressure - pwf (bar) Inlet temperature Twf (oC) Stock tank oil flow rate - QL (Sm /day) Total mass flowrate - m (kg/s) Wellhead pressure pwh (bar) Void fraction Mixture density M (kg/m3) Fluid mean velocity u (m/s) Acoustic velocity a (m/s) Water-hammer pa (bar)

Stock tank gas flow rate - QG (million m /day) 0.3796 0.26651 0.20352 0.0641

Figure 6.1 contains two plots that present estimated water-hammer and line packing for oil wells taken into considerations in present work. The first pressure increase on the plots is the water hammer effect after valve closure, and then the long line packing shows the pressure build up due to friction. The rapid pressure increase in line packing starts at the time where bubble point is reached.

6. Case Studies oil wells 80 __________________________________________________________________________

180 160 140 120 pressure [bar] 100 80 60 40 20 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 tim e elapsed [s] Well A2 Well A1 Well C Well B

35.0 32.5 30.0 27.5 25.0 22.5 pressure [bar] 20.0 17.5 15.0 12.5 10.0 7.5 5.0 2.5 0.0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 tim e elapsed [s] Well C Well A1 Well A2 Well B

Figure 6.1 Estimated water-hammer and line packing in oil wells

6. Case Studies oil wells 81 __________________________________________________________________________

Calculated results are presented for each well in the form of plots. The list of plots is given below:
mixture density Vs depth, ..... Figure 6.2 void fraction Vs depth. Figure 6.3 densities ratio Vs velocities ratio........ Figure 6.4 logarithm of densities ratio Vs logarithm of velocities ratio.............. Figure 6.5 acoustic velocity Vs depth.. Figure 6.6 acoustic velocity Vs time.... Figure 6.7 sum frictional pressure drop Vs depth........ Figure 6.8 sum frictional pressure drop Vs time...... Figure 6.9 pressure (line packing) Vs time...... Figure 6.10 acoustic velocity Vs void fraction (for increasing pressures)......... Figure 6.11 elevation Vs pressure.......... Figure 6.12

The depth is the depth of the point in question below the wellhead. The time elapsed is the time period it takes for a sound wave to travel to the point in question from the wellhead and back again, i.e. down and up. The sum frictional pressure drop is the total pressure drop between the point in question and the wellhead.

6. Case Studies well A1 82 __________________________________________________________________________

800

700 600 mixture density [kg/m3]

500 400

300 200

100 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.2 A1, mixture density fraction Vs depth

90 80 70 60 50 40 30 20 10 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

void fraction [%]

Figure 6.3 A1, void fraction Vs depth

6. Case Studies well A1 83 __________________________________________________________________________

90 80 70 densities ratio DL/DG 60 bubble II 50 40 30 20 10 0 0.0 0.2 0.4 0.6 0.8 velocities ratio (slip) vG/vL 1.0 1.2 1.4 bubble I Transition of flow pattern flow revealed on this plot, it w as not reported in simulator output file.

liquid only

Figure 6.4 A1, densities ratio Vs velocities ratio

10

densities ratio log(D L/DG)

bubble II bubble I 1 0.0 0.1 velocities ratio (slip) log(vG/vL) 1.0

Figure 6.5 A1, logarithm of densities ratio Vs logarithm velocities ratio

6. Case Studies well A1 84 __________________________________________________________________________

1200 1100 1000 900 acoustic velocity [m/s] 800 700 600 500 400 300 200 100 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.6 A1, acoustic velocity Vs depth

1200

1000

acoustic velocity [m/s]

800

600

400

200

0 0 1 2 3 4 5 6 7 8 tim e [s] 9 10 11 12 13 14 15

Figure 6.7 A1, acoustic velocity Vs time

6. Case Studies well A1 85 __________________________________________________________________________

12

10 sum frictional pressure drop [bar]

0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.8 A1, sum of frictional pressure drop Vs depth

12

10 sum frictional pressure drop [bar]

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 tim e [s]

6. Case Studies well A1 86 __________________________________________________________________________

Figure 6.9 A1, sum of frictional pressure drop Vs time

24 22 20 18 16 pressure [bar] 14 12 10 8 6 line packing time differential

5 5 4 4 3 3 2 2 1

time differential dp/dt

4 2 0 0 1 2 3 4 5 6 7 8 9 10 11 tim e elapsed [s] 12 13 14 15 16 17 18 1 0

Figure 6.10 A1, line packing and time differential

6. Case Studies well A1 87 __________________________________________________________________________


400

750m 600m

350

450m 300m

300 sonic velocity [m/s]

150m 0m

250

path

200

150

100

50 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7 0.8 0.9

Calculated acoustic velocities for subsequent pressures across the w ellbore

Figure 6.11 A1, acoustic velocity Vs void fraction (for increasing pressure)

Figure 6.12 A1, elevation Vs pressure

6. Case Studies well A2 88 __________________________________________________________________________

800 700 600 500 400 300 200 100 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

mixture density [kg/m3]

Figure 6.2 A2, mixture density fraction Vs depth

70

60

50 void fraction [%]

40

30

20

10

0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.3 A2, void fraction Vs depth

6. Case Studies well A2 89 __________________________________________________________________________

40 35 30 densities ratio DL/DG 25 bubble 20 15 10 liquid only 5 0 0.0 0.2 0.4 0.6 0.8 velocities ratio (slip) vG/vL 1.0 1.2 1.4 Transition betw een liquid, bubble and slug flow slug

Figure 6.4 A2, densities ratio Vs velocities ratio

10

densities ratio log(DL/DG)

1 0.0 0.1 velocities ratio (slip) log(vG/vL) 1.0

Figure 6.5 A2, logarithm of densities ratio Vs logarithm velocities ratio

6. Case Studies well A2 90 __________________________________________________________________________

1200 1100 1000 acoustic velocity [m/s] 900 800 700 600 500 400 300 200 100 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.6 A2, acoustic velocity Vs depth

1200

1000

acoustic velocity [m/s]

800

600

400

200

0 0 1 2 3 4 5 6 tim e [s] 7 8 9 10 11 12

Figure 6.7 A2, acoustic velocity Vs time

6. Case Studies well A2 91 __________________________________________________________________________

12

10 sum frictional pressure drop [bar]

0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.8 A2, sum of frictional pressure drop Vs depth

12

10 sum frictional pressure drop [bar]

0 0 1 2 3 4 5 6 tim e [s] 7 8 9 10 11 12

Figure 6.9 A2, sum of frictional pressure drop Vs time

6. Case Studies well A2 92 __________________________________________________________________________

40 line packing 35 time differential 30 25

time differential dp/dt

pressure [bar]

4 20 3 15 2 10 5 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1

time elapsed [s]

Figure 6.10 A2, line packing and time differential

6. Case Studies well A2 93 __________________________________________________________________________

480 430 380 sonic velocity [m/s] 330 280 230 180 130 80 0 0.1 0.2 0.3 0.5 0.6 void fraction Calculated sonic velocities 0.4 0.7 0.8

600m 500m 400m 300m 200m 100m 0m path

0.9

Figure 6.11 A2, acoustic velocity Vs void fraction (for increasing pressure)

Figure 6.12 A2, elevation Vs pressure

6. Case Studies well B 94 __________________________________________________________________________

800 700 600 mixture density [%] 500 400 300 200 100 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.2 B, mixture density fraction Vs depth

25

20 void fraction [%]

15

10

0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.3 B, void fraction Vs depth

6. Case Studies well B 95 __________________________________________________________________________


8 7 6 densities ratio DL/DG 5 4 liquid only 3 2 1 0 0.0 0.2 0.4 0.6 0.8 1.0 velocities ratio (slip) vG/vL 1.2 1.4 1.6

bubble flow

Figure 6.4 B, densities ratio Vs velocities ratio

1 0.1

1.0

densities ratio log(D w /Ds)

0 velocities ratio (slip) log(vs/vw )

Figure 6.5 B, logarithm of densities ratio Vs logarithm velocities ratio

6. Case Studies well B 96 __________________________________________________________________________

1300 1200 1100 1000 acoustic velocity [m/s] 900 800 700 600 500 400 300 200 100 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.6 B, acoustic velocity Vs depth

1400

1200

acoustic velocity [m/s]

1000

800

600

400

200

0 0 1 2 3 4 tim e [s] 5 6 7 8

Figure 6.7 B, acoustic velocity Vs time

6. Case Studies well B 97 __________________________________________________________________________

30

25 sum frictional pressure drop [bar]

20

15

10

0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.8 B, sum of frictional pressure drop Vs depth

30

25 sum frictional pressure drop [bar]

20

15

10

0 0 1 2 3 4 tim e [s] 5 6 7 8

Figure 6.9 B, sum of frictional pressure drop Vs time

6. Case Studies well B 98 __________________________________________________________________________

190 180 170 160 150 pressure [bar] 140 130 line packing time differential

12

10

time differential dp/dt

6 120 110 100 90 80 70 60 0 1 2 3 4 5 6 tim e elapsed [s] 7 8 9 10 0 bubble point pressure 2 4

Figure 6.10 B, line packing and time differential

6. Case Studies well B 99 __________________________________________________________________________


1200 1700m 1000 sonic velocity [m/s] 1400m 1100m 800m 800 500m 200m 600 0m path 400

200

0 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7 0.8 0.9 1

Calculated acoustic velocities for subsequent pressures across the w ellbore

Figure 6.11 B, acoustic velocity Vs void fraction (for increasing pressure)

Figure 6.12 B, elevation Vs pressure

6. Case Studies well C 100 __________________________________________________________________________

900 800 700 mixture density [kg/m3] 600 500 400 300 200 100 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.2 A1, mixture density fraction Vs depth

30

25

void fraction [%]

20

15

10

0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.3 A1, void fraction Vs depth

6. Case Studies well C 101 __________________________________________________________________________

18 16 14 densities ratio DL/DG 12 10 8 6 4 2 0 0.0 0.2 0.4 0.6 0.8 1.0 velocities ratio (slip) vG/vL 1.2 1.4 1.6 liquid only bubble

Figure 6.4 A1, densities ratio Vs velocities ratio

10.0

densities ratio log(DL/DG )

flow transition

bubble

1.0

0.000001

0.000010

0.000100 0.001000 0.010000 velocities ratio (slip) log(vG/vL)

0.100000

0.1 1.000000

Figure 6.5 A1, logarithm of densities ratio Vs logarithm velocities ratio

6. Case Studies well C 102 __________________________________________________________________________

1200 1100 1000 900 acoustic velocity [m/s] 800 700 600 500 400 300 200 100 0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.6 A1, acoustic velocity Vs depth


1200

1000

acoustic velocity [m/s]

800

600

400

200

0 0 1 2 3 4 tim e [s] 5 6 7 8 9

Figure 6.7 A1, acoustic velocity Vs time

6. Case Studies well C 103 __________________________________________________________________________

sum frictional pressure drop [bar]

0 0 200 400 600 800 1000 depth [m ] 1200 1400 1600 1800 2000

Figure 6.8 A1, sum of frictional pressure drop Vs depth

5 sum frictional pressure drop [bar]

0 0 1 2 3 4 tim e [s] 5 6 7 8 9

Figure 6.9 A1, sum of frictional pressure drop Vs time

6. Case Studies well C 104 __________________________________________________________________________

74 73 72 71

4.5 4.0 3.5 3.0

line packing time differential

time differential dp/dt

pressure [bar]

70 2.5 69 2.0 68 1.5 67 66 65 64 0 1 2 3 4 5 6 7 8 9 10 11 tim e elapsed [s] 1.0 0.5 0.0

Figure 6.10 A1, line packing and time differential

6. Case Studies well C 105 __________________________________________________________________________


1000 900 800 sonic velocity [m/s] 700 600 500 400 300 200 100 0 0.5 0.6 0.7 0.8 0.9 void fraction Calculated acoustic velocities for subsequent pressures across the w ellbore 0.1 0.2 0.3 0.4 1 1050m 900m 700m 500m 300m 100m 0m path

Figure 6.11 A1, acoustic velocity Vs void fraction (for increasing pressure)

Figure 6.12 A1, elevation Vs pressure

6. Case Studies geothermal wells 106 __________________________________________________________________________ 6.3 Water-hammer and line packing in geothermal wells

Two programs used in this case were HOLA 3.1 wellbore simulator, and Excel (also MATHLAB 6.0 and GOW 3.0 but in limited single tasks). The features and modes of HOLA program that was used in present work are described in Appendix E. The program requires the well and feedzone parameters in order to calculate the various parameters of fluid and system in wellbore. This program was also used for inflow performance (IP) and tubing performance (VLP) purposes in order to estimate the flow parameters for different tubing size (13 and 9). The wellhead pressure was assumed to have the same value for both tubing diameters, thus the mass flowrate must be different. Calculation confirmed that increase in tubing size from 9 to 13 can bring the almost double output what justify economically drilling such wide wells. The 13 wide configuration is widely adopted in Iceland but literature shows that 9 may be considered as typical pipe diameter for high enthalpy liquid dominated geothermal wells worldwide. That was the purpose of including the both tubing sizes in simulations. The IP (inflow performance) curves could not be calculated directly from HOLA and some special approach was used that is described in Appendix E. The Excel spreadsheet was used to perform the calculations and plot the results. The data for simulations were assumed base on information about two Iceland fields Reykjanes and Svartsengi, (Gudmundsson, 2003). The data include information about average reservoir pressure and temperature, well and producing fractures depth, casing program and wellhead production parameters. The additional like enthalpies and productivity-indexes was computed using HOLA simulator. The wells D and E are representing the same field thus input parameters and simulated results are similar. The wells indicated with F are based on data from a different field thus properties and results may be expected to vary from the previous wells. The HOLA simulator gives the limited number of fluid properties as compared to PipeSim, but those obtained were sufficient to continue the necessary calculations in Excel spreadsheet. HOLA output text file gives calculated well profile containing values of pressure, temperature, enthalpy, dryness, density, and velocity of each phase separately. The steam

6. Case Studies geothermal wells 107 __________________________________________________________________________

void fraction was calculated from equation (4.5), and then the mixture density, viscosity and enthalpy was computed from relations (4.7), (4.8), (4.9) respectively. Acoustic velocity was obtained from Woods formula given as (5.21) in present work. In single liquid phase flow sonic velocity depends mostly on water compressibility whether in two phase region the sonic velocity is highly dependant on steam compressibility and water compressibility has much less impact. The steam properties in geothermal well follows the saturation line as the system tends to remain in equilibrium between both phases. These properties are given is steam tables, what is obviously inconvenient for computer calculations. Michaelides (1981) proposed the polynomial expressions to solve this inconvenience problem. In present work CalcSoft 3.0 the shareware program written by M.L. McGuire was used in computations of steam and water heat capacities at constant pressure and volume. Isothermal compressibility KTW of the water phase was possible to compute from GOW program and the isothermal compressibility of steam fraction was calculated from formula suitable for computer calculations, derived in chapter 5 and finally given by the equation (5.24). The acoustic velocity allows then calculate the travelling time of pressure pulse down to a certain point and back again. Frictional pressure drop across the wellbore is the required value for line packing calculations. From HOLA program only total pressure drop across the wellbore is available, thus the friction factor f was computed for given pipe roughness from empirical formula proposed by Haalad (Stetfjerding, 1998)
n

1 1.8 ks 6.9 = log + f n Re 3.75 d

1.11n

(6.1)

where n = 1 for gradual transition between the smooth and rough flow, or n = 3 that is suitable for gas pipelines. The Reynolds number is given by the equitation (4.26), ks is the wall roughness and d pipe diameter. Pressure drop due to friction may be calculated from

6. Case Studies geothermal wells 108 __________________________________________________________________________

Darcy-Weisbach equation given as (2.6). The total pressure drop versus time gives the effect of line packing. Calculated values were then plotted in Excel. Depth and time are plotted on the x axis and total pressure drop and acoustic velocity on the y axis. The results are presented it two forms Excel spreadsheet and Excel plots. Excel spreadsheet contains the calculated values of various properties affecting the acoustic velocity and line packing listed in columns and is available in Appendix G for all simulated geothermal wells. These values were directly used to plot the results. Table 6.4 contains geothermal wells data used in present calculations. The table include well depths, reservoir parameters and two stage tubing diameters. Calculated parameters are presented in Table 6.5 which includes wellhead conditions and parameters that make up water-hammer at the wellhead. Estimated water-hammer and line packing in geothermal wells are shown in the form of plots. The Figure 6.13 presents the results for the wells taken into consideration in present work, the water-hammer pressure increase starts from the actual calculated wellhead pressure. Line packing in the case of geothermal wells contains both bubble point and diameter change effects. These effects are discussed precisely in the next chapter. Figure 6.14 shows the same plots but the pressure on the y axis is the only pressure increase due to water-hammer and friction effects, beginning from the 0 pressure value. Figure 6.15 presents only the waterhammer pressure increase after a valve closure.

6. Case Studies geothermal wells 109 __________________________________________________________________________

Table 6.4 Reservoir and geometry input data for computer simulations

WELL Depth (m) Reservoir Pressure pr (bar) Reservoir Temperature Tr (oC) Downhole pressure pwf (bar) Enthalpy at the feedzone h (kJ/kg) Productivity Index PI (E-12kg/s/m3) Production casing ID (inch) Slotted liner ID (inch) Liner from depth (m)

D1 2028 197.8 309.6 176.3 1389 2.39 13 9 750

D2 2028 197.8 309.6 179.3 1389 2.39 9 7 750

E1 2228 198 309.7 174.3 1390 1.25 13 9 750

E2 2228 198 309.7 183.7 1284 1.25 9 7 750

F1 1450 100 263.1 83.1 1150 3.61 13 9 750

F2 1450 100 263.1 90.2 1129 3.61 9 7 750

Table 6.5 Calculated values at the wellhead conditions

WELL Wellhead pressure pwh (bar) Wellhead temperature Twh (oC) Wellhead enthalpy hwh (kJ/kg) Total mass flow m (kg/s) Void fraction Fluid mean velocity (m/s) Mixture density (kg/m3) Acoustic velocity (m/s) Water-hammer pa (bar)

D1 45.17 257.6 1275.7 42.61 0.615 1.48 317.1 152.6 0.716

D2 44.7 257.0 1255.4 25.98 0.628 1.18 307.6 152.6 0.848

E1 31.94 237.3 1198.7 49.87 0.640 1.80 304.4 128.1 0.703

E2 31.93 237.3 1168.9 30.28 0.722 2.21 239.1 136.0 0.880

F1 15.02 198.3 1082.1 47.15 0.878 4.61 112.6 123.4 0.641

F2 15.06 198.5 1051.0 28.46 0.879 5.46 111.4 123.8 0.754

6. Case Studies geothermal wells 110 __________________________________________________________________________

49 Well D2 46 43 40 37 pressure [bar] 34 31 28 25 22 19 16 13 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 tim e elapsed [s] Well D1

Well E2

Well E1

Well F2

Well F1

Figure 6.13 Estimated water-hammer and line packing in geothermal wells

6. Case Studies geothermal wells 111 __________________________________________________________________________


3.50 3.25 3.00 2.75 2.50 2.25 2.00 1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 tim e elapsed [s] Well D1 Well F1 Well D2 Well E1 pressure [bar] Well F2 Well E2

Figure 6.14 Estimated water-hammer and line packing in geothermal wells

6. Case Studies geothermal wells 112 __________________________________________________________________________


1.10 0.99 0.88 0.77 0.66 0.55 0.44 0.33 0.22 0.11 0.00 0.0 0.1 0.2 0.3 0.4 tim e elapsed [s] 0.5 0.6 0.7 0.8

water - hammer pressure [bar]

Well D1 Well D2 Well E1 Well E2 Well F1 Well F2

Figure 6.15 Estimated water-hammer magnitudes in geothermal wells (valve is assumed to close completely in 0.5 s)

6. Case Studies geothermal wells 113 __________________________________________________________________________

Calculated results are presented for each well in the form of plots. The list of plots is given below:
VLP deliverability curves... Figure 6.16 pressure and temperature well profile......... Figure 6.17 mixture density Vs depth.Figure 6.18 void fraction Vs depth. Figure 6.19 densities ratio Vs velocities ratio.... Figure 6.20 logarithm of densities ratio Vs logarithm of velocities ratio.. Figure 6.21 acoustic velocity Vs depth.. Figure 6.22 acoustic velocity Vs time ... Figure 6.23 frictional pressure drop per 10 (m) Vs depth.. Figure 6.24 sum frictional pressure drop Vs depth Figure 6.25 sum frictional pressure drop Vs time.. Figure 6.26 pressure (line packing) Vs time . Figure 6.27 acoustic velocity Vs void fraction (for increasing pressures) .... Figure 6.28 acoustic velocity Vs void fraction (for increasing pressures),

scaled up plot (only F1 and F2 wells) .... Figure 6.29

The depth is the depth of the point in question below the wellhead. The time elapsed is the time period it takes for a sound wave to travel to the point in question from the wellhead and back again, i.e. down and up. The frictional pressure drop per 10 meters is the pressure loss due to wall friction at this distance. The sum frictional pressure drop is the total pressure drop between the point in question and the wellhead.

6. Case Studies well D1 114 __________________________________________________________________________

320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0
0 20 40 60 80 100 120 140

95/8 '' 133/8 ''

Pwf [bar]

160

180

200

220

240

260

280

300

320

340

m [kg/s]

Figure 6.16 D1, well inflow and tubing performance

Figure 6.17 D1, simulated well pressure and temperature profile

6. Case Studies well D1 115 __________________________________________________________________________


800 700 600 mixture density [kg/m3] 500 400 300 200 100 0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.18 D1, mixture density Vs well depth

0.7 0.6 0.5 void fraction 0.4 0.3 0.2 0.1 0.0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.19 D1, void fraction Vs well depth

6. Case Studies well D1 116 __________________________________________________________________________


40 35 30 densities ratio Dw/Ds 25 20 15 10 5 0 0.0 0.5 1.0 1.5 2.0 2.5

velocities ratio (slip) vs /vw

Figure 6.20 D1, densities ratio Vs velocities ratio

10

Transition between bubble and slug flow


densities ratio log(D w /Ds)

1 0.0 0.1 1.0

velocities ratio (slip) log(vs /vw )

Figure 6.21 D1, logarithm of densities ratio Vs logarithm of velocities ratio

6. Case Studies well D1 117 __________________________________________________________________________


1600 1400 1200 acustic velocity [m/s] 1000 800 600 400 200 0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.22 D1, acoustic velocity Vs well depth

1600 1400 1200 acustic velocity [m/s] 1000 800 600 400 200 0 0 1 2 3 4 5 tim e [s] 6 7 8 9 10

Figure 6.23 D1, acoustic velocity Vs time

6. Case Studies well D1 118 __________________________________________________________________________


0.006

0.005 frictional pressure drop per 10m [bar/m]

0.004

0.003 discontinuity 0.002

diameter change from 133/8 to 95/8

0.001

first bubble appears

0.000 0 200 400 600 800 1000 1200 1400 1600 1800 2000 w ell depth [m ]

Figure 6.24 D1, frictional pressure drop at 10 m distance Vs well depth

0.9 0.8 Sum frictional pressure drop [bar/m] 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.25 D1, sum of frictional pressure drop Vs well depth

6. Case Studies well D1 119 __________________________________________________________________________


0.90 0.80 0.70 Sum frictional pressure drop [bar] 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0 1 2 3 4 5 tim e [m ] 6 7 8 9 10

Figure 6.26 D1, sum of frictional pressure drop Vs time

46.8 46.6 46.4 46.2 Pressure [bar]

1.6 1.4 1.2 1

line packing time differential

time differential dp/dt

46.0 0.8 45.8 0.6 45.6 45.4 45.2 45.0 0 1 2 3 4 5 6 7 tim e [m ] 8 9 10 11 12 0.4 0.2 0

Figure 6.27 D1, estimated water hammer and line packing at the wellhead, dp (together with time differential ) dt

6. Case Studies well D1 120 __________________________________________________________________________

450 410 acoustic velocity [m/s] 370 330 290 250 210 170 130 0 0.1 0.2 0.3 0.4 0.5 0.6 void fraction 0.7 0.8 0.9

0m 100 m 200 m 300 m 400 m 500 m 600 m 640 m 660 m path

Figure 6.28 D1, Calculated sonic velocities for varying pressure across the well depth
Plot presents void fraction Vs acoustic velocity changes across the well depth. Sudden change in void fraction is caused by computational discontinuity at the 230m depth

6. Case Studies well D2 121 __________________________________________________________________________

320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0
0 20 40 60 80 100 120 140

95/8 '' 133/8 ''

Pwf [bar]

160

180

200

220

240

260

280

300

320

340

m [kg/s]

Figure 6.16 D2, well inflow and tubing performance

Figure 6.17 D2, simulated well pressure and temperature profile

6. Case Studies well D2 122 __________________________________________________________________________

800 700 600 mixture density [kg/m3] 500 400 300 200 100 0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.18 D2, mixture density Vs well depth

1.0 0.9 0.8 0.7 void fraction 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.19 D2, void fraction Vs well depth

6. Case Studies well D2 123 __________________________________________________________________________


40 35 30 densities ratio Dw/Ds 25 20 15 10 5 0 0.0 0.5 1.0 1.5 2.0 2.5

velocities ratio (slip) vs /vw

Figure 6.20 D2, densities ratio Vs velocities ratio

10

Transition between bubble and slug flow


densities ratio log(D w /Ds)

1 0.0 0.1 1.0

velocities ratio (slip) log(vs /vw )

Figure 6.21 D2, logarithm of densities ratio Vs logarithm of velocities ratio

6. Case Studies well D2 124 __________________________________________________________________________


1600 1400 1200 acustic velocity [m/s] 1000 800 600 400 200 0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.22 D2, acoustic velocity Vs well depth

1600 1400 1200 acustic velocity [m/s] 1000 800 600 400 200 0 0.0 2.0 4.0 tim e [s] 6.0 8.0 10.0

Figure 6.23 D2, acoustic velocity Vs time

6. Case Studies well D2 125 __________________________________________________________________________


0.012

0.010 frictional pressure drop per 10m [bar/m]

0.008

0.006

diameter change from9 3/8'' to 7''

0.004

0.002 first bubble appears 0.000 0 200 400 600 800 1000 1200 1400 1600 1800 2000 w ell depth [m ]

Figure 6.24 D2, frictional pressure drop at 10 m distance Vs well depth

1.6 1.4 Sum frictional pressure drop [bar/m] 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.25 D2, sum of frictional pressure drop Vs well depth

6. Case Studies well D2 126 __________________________________________________________________________


1.60 1.40 Sum frictional pressure drop [bar] 1.20 1.00 0.80 0.60 0.40 0.20 0.00 0 1 2 3 4 5 tim e [m ] 6 7 8 9 10

Figure 6.26 D2, sum of frictional pressure drop Vs time

47.5 line packing time differential 46.5 Pressure [bar]

1.8 1.6 1.4


time differential dp/dt

47.0

1.2 1

46.0 0.8 45.5 0.6 0.4 45.0 0.2 44.5 0 1 2 3 4 5 6 tim e [m ] 7 8 9 10 11 0

Figure 6.27 D2, estimated water hammer and line packing at the wellhead dp (together with time differential ) dt

6. Case Studies well D2 127 __________________________________________________________________________

290

250 sonic velocity [m/s]

210
0m 100 m 200 m 300 m 400 m 500 m 550 m 570 m 590 m path

170

130 0 0.1 0.2 0.3 0.4 0.5 0.6 void fraction 0.7 0.8

0.9

Figure 6.28 D2, calculated sonic velocities for varying pressure across the well depth

6. Case Studies well E1 128 __________________________________________________________________________

300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360

9 5/8 '' 13 3/8 ''

Pwf [bar]

m [kg/s]

Figure 6.16 E1, well inflow and tubing performance

Figure 6.17 E1, simulated well pressure and temperature profile

6. Case Studies well E1 129 __________________________________________________________________________

800 700 600 mixture density [kg/m3] 500 400 300 200 100 0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.18 E1, mixture density Vs well depth

1.0 0.9 0.8 0.7 mixture density [kg/m3] 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.19 E1, void fraction Vs well depth

6. Case Studies well E1 130 __________________________________________________________________________

60

50

densities ratio Dw/Ds

40

30

20

10

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

velocities ratio (slip) vs /vw

Figure 6.20 E1, densities ratio Vs velocities ratio

10

Transition between the bubble flow and slug flow (discountinuity at the flow patterns boundary)
densities ratio log(D w /Ds)

1 0.0 0.1 1.0

velocities ratio (slip) log(vs /vw )

Figure 6.21 E1, logarithm of densities ratio Vs logarithm of velocities

6. Case Studies well E1 131 __________________________________________________________________________

1600 1400 1200 acustic velocity [m/s] 1000 800 600 400 200 0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.22 E1, acoustic velocity Vs well depth

1600 1400 1200 acustic velocity [m/s] 1000 800 600 400 200 0 0 1 2 3 4 5 6 7 tim e [s] 8 9 10 11 12 13

Figure 6.23 E1, acoustic velocity Vs time

6. Case Studies well E1 132 __________________________________________________________________________


0.010 0.009 0.008 frictional pressure drop per 10m [bar/m] 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0.000 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000 diameter change from 13 3/8" to 9 5/8" first bubble appears

Figure 6.24 E1, frictional pressure drop at 10 m distance Vs well depth

1.4

1.2 Sum frictional pressure drop [bar/m]

1.0

0.8

0.6

0.4

0.2

0.0 0 200 400 600 800 1000 1200 1400 w ell depth [m ] 1600 1800 2000 2200 2400

Figure 6.25 E1, sum of frictional pressure drop Vs well depth

6. Case Studies well E1 133 __________________________________________________________________________

1.40

1.20 Sum frictional pressure drop [bar]

1.00

0.80

0.60

0.40

0.20

0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 tim e [m ]

Figure 6.26 E1, sum of frictional pressure drop Vs time

34.5 line packing time differential dp/dt 33.5 Pressure [bar]

1.6 1.4 1.2 1.0 0.8 0.6 0.4

34.0

time differential dp/dt

33.0

32.5

32.0

0.2 0.0 0 1 2 3 4 5 6 7 8 tim e [m ] 9 10 11 12 13 14 15

31.5

Figure 6.27 E1, estimated water hammer and line packing at the wellhead dp (together with time differential ) dt

6. Case Studies well E1 134 __________________________________________________________________________

430 390 350 310 270 230 190 150 110 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7 0.8 0.9

sonic velocity [m/s]

0m 100 m 200 m 300 m 400 m 500 m 600 m 700 m 750 m 790 m path

Figure 6.28 E1, calculated sonic velocities for varying pressure across the well depth

6. Case Studies well F1 135 __________________________________________________________________________

300 280 260 240 220 200

9 5/8 '' 13 3/8 ''

Pwf [bar]

180 160 140 120 100 80 60 40 20 0 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360

m [kg/s]

Figure 6.16 E2, well inflow and tubing performance

Figure 6.17 E2, simulated well pressure and temperature profile

6. Case Studies well F1 136 __________________________________________________________________________

800 700 600 500 400 300 200 100 0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

mixture density [kg/m3]

Figure 6.18 E2, mixture density Vs well depth

1.0 0.9 0.8 0.7 void fraction 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.19 E2, void fraction Vs well depth

6. Case Studies well F1 137 __________________________________________________________________________

60

50

densities ratio Dw/Ds

40

30

20

10

0 0.0 0.5 1.0 1.5 2.0

velocities ratio (slip) vs /vw

Figure 6.20 E2, densities ratio Vs velocities ratio

10

densities ratio log(D w /Ds)

1 0.0 0.1 1.0

velocities ratio (slip) log(vs /vw )

Figure 6.21 E2, logarithm of densities ratio Vs logarithm of velocities ratio

6. Case Studies well F1 138 __________________________________________________________________________

1400

1200

1000 acustic velocity [m/s]

800

600

400

200

0 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

Figure 6.22 E2, acoustic velocity Vs well depth

1400 1200 1000 acustic velocity [m/s] 800 600 400 200 0 0 1 2 3 4 5 6 tim e [s] 7 8 9 10 11

Figure 6.23 E2, acoustic velocity Vs time

6. Case Studies well F1 139 __________________________________________________________________________


0.016

0.014

0.012 frictional pressure drop per 10m [bar/m]

0.010 diameter change from 13 3/8" to 9 5/8"

0.008

0.006

0.004 first bubble appears 0.000 0 200 400 600 800 1000 1200 w ell depth [m ] 1400 1600 1800 2000

0.002

Figure 6.24 E2, frictional pressure drop at 10 m distance Vs well depth


2.5

2.0 Sum frictional pressure drop [bar/m]

1.5

1.0

0.5

0.0 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 w ell depth [m ]

Figure 6.25 E2, sum of frictional pressure drop Vs well depth

6. Case Studies well F1 140 __________________________________________________________________________


2.50

2.00 Sum frictional pressure drop [bar]

1.50

1.00

0.50

0.00 0 1 2 3 4 5 6 7 8 9 10 11 tim e [m ]

Figure 6.26 E2, sum of frictional pressure drop Vs time


35.5 35.0 34.5 2.0 1.8 line packing time diferential 1.6 1.4
time differential dp/dt

Pressure [bar]

34.0 33.5 33.0 32.5

1.2 1.0 0.8 0.6 0.4

32.0 31.5 0 1 2 3 4 5

bubble point

0.2 0.0

6 7 tim e [m ]

10

11

12

Figure 6.27 E2, estimated water hammer and line packing at the wellhead dp (together with time differential ) dt

6. Case Studies well F1 141 __________________________________________________________________________

430 390 sonic velocity [m/s] 350 310 270 230 190 150 110 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7 0.8

0m 100 m 200 m 300 m 400 m 500 m 550 m path

0.9

Figure 6.27 E2, calculated sonic velocities for varying pressure across the well depth

6. Case Studies well F1 142 __________________________________________________________________________

160 140 120 100 Pwf [bar] 80 60 40 20 0 0 20 40 60 80 100 120 140 160 180 200 220 240 260 m [kg/s]

13 3/8 '' 9 5/8 ''

Figure 6.16 F1, well inflow and tubing performance

Figure 6.17 F1, simulated well pressure and temperature profile

6. Case Studies well F1 143 __________________________________________________________________________

900 800 700 mixture density [kg/m3] 600 500 400 300 200 100 0 0 200 400 600 800 w ell depth [m ] 1000 1200 1400 1600

Figure 6.18 F1, mixture density Vs well depth

1.0 0.9 0.8 0.7 void fraction [kg/m3] 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 200 400 600 800 w ell depth [m ] 1000 1200 1400 1600

Figure 6.19 F1, void fraction Vs well depth

6. Case Studies well F1 144 __________________________________________________________________________

120

100

densities ratio Dw/Ds

80

60

40

20

0 0.0 0.5 1.0 1.5 2.0 2.5

velocities ratio (slip) vs /vw

Figure 6.20 F1, densities ratio Vs velocities ratio

10

Transition between bubble and slug flow and discontinuity at the flow patterns boundary
densities ratio log(D w /Ds)

1 0.0 0.1 1.0

velocities ratio (slip) log(vs /vw )

Figure 6.21 F1, logarithm of densities ratio Vs logarithm of velocities ratio

6. Case Studies well F1 145 __________________________________________________________________________

1400 1200 1000 acustic velocity [m/s] 800 600 400 200 0 0 200 400 600 800 w ell depth [m ] 1000 1200 1400 1600

Figure 6.22 F1, acoustic velocity Vs well depth

1400 1200 1000 acustic velocity [m/s] 800 600 400 200 0 0 2 4 6 8 10 tim e [s] 12 14 16 18 20

Figure 6.23 F1, acoustic velocity Vs time

6. Case Studies well F1 146 __________________________________________________________________________


0.016

0.014

frictional pressure drop per 10m [bar/m]

0.012 Diameter change from 13 3/8'' for 9 5/8''

0.010

0.008 0.006 first bubble appears

0.004

0.002

0.000 0 200 400 600 800 1000 1200 1400 1600 1800 2000 w ell depth [m ]

Figure 6.24 F1, frictional pressure drop at 10 m distance Vs well depth


1.0 0.9 0.8 Sum frictional pressure drop [bar/m] 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 200 400 600 800 1000 w ell depth [m ] 1200 1400 1600

Figure 6.25 F1, sum of frictional pressure drop Vs well depth

6. Case Studies well F1 147 __________________________________________________________________________

1.00 0.90 0.80 Sum frictional pressure drop [bar] 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 tim e [m ]

Figure 6.26 F1, sum of frictional pressure drop Vs time

16.6 16.4 16.2 16.0 Pressure [bar] 15.8 15.6 15.4 15.2 15.0 14.8 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 tim e [m ] diameter change bubble point line packing time differential

1.4 1.2

time differential dp/dt

1.0 0.8 0.6 0.4 0.2 0.0

Figure 6.27 F1, estimated water hammer and line packing at the wellhead dp (together with time differential ) dt

6. Case Studies well F1 148 __________________________________________________________________________

240 220
sonic velocity [m/s]

200 180 160 140 120 100 80 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7

0m 100 m 200 m 300 m 400 m 500 m 600 m 700 m 800 m 900 m 960 m path

0.8

0.9

Figure 6.28 F1, calculated sonic velocities for varying pressure across the well depth

Calculated sonic velocities 140

135

sonic velocity [m/s]

130

125
0m 100 m 200 m 300 m 400 m 500 m 600 m 700 m 800 m path

120

115

110 0.5 0.6 0.7


void fraction

0.8

0.9

Figure 6.29 F1, calculated sonic velocities for varying pressure across the well depth scaled up.
The same plot scaled up. Acoustic velocity decreases although pressure increases across wellbore. It is caused by void fraction decrease which more affects acoustic velocity value than pressure for right-hand void values above = 0

6. Case Studies well F2 149 __________________________________________________________________________

160 140 120 100 Pwf [bar] 80 60 40 20 0 0 20 40 60 80 100 120 140 160 180 200 220 240 260 m [kg/s]

13 3/8 '' 9 5/8 ''

Figure 6.16 F2, well inflow and tubing performance

Figure 6.16 F2, simulated well pressure and temperature profile

6. Case Studies well F2 150 __________________________________________________________________________

900 800 700


mixture density [kg/m3]

600 500 400 300 200 100 0 0 200 400 600 800 1000 w ell depth [m ] 1200 1400 1600

Figure 6.17 F2, mixture density Vs well depth

1.0 0.9 0.8 0.7


void fraction

0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 200 400 600 800 1000 w ell depth [m ] 1200 1400 1600

Figure 6.18 F2, void fraction Vs well depth

6. Case Studies well F2 151 __________________________________________________________________________

120 Flow patern changes from liquid only to bubble flow and then to slug flow , at the diameter change flow comes back to bubbly flow and then transforms into slug again

100

densities ratio Dw/Ds

80

60

40

20

0 0.0 0.5 1.0 1.5 2.0

velocities ratio (slip) vs /vw

Figure 6.19 F2, densities ratio Vs velocities ratio

10

densities ratio log(D w /Ds)

1 0.0 0.1 velocities ratio (slip) log(vs /vw ) 0.0 1.0

Figure 6.20 F2, logarithm of densities ratio Vs logarithm of velocities ratio

6. Case Studies well F2 152 __________________________________________________________________________

1400 1200 1000


acustic velocity [m/s]

800 600 400 200 0 0 200 400 600 800


w ell depth [m ]

1000

1200

1400

1600

Figure 6.21 F2, acoustic velocity Vs well depth

1400 1200 1000


acustic velocity [m/s]

800 600 400 200 0 0 2 4 6 8 10 tim e [s] 12 14 16 18 20

Figure 6.22 F2, acoustic velocity Vs time

6. Case Studies well F2 153 __________________________________________________________________________


0.018 0.016 0.014
frictional pressure drop per 10m [bar/m]

Diameter change from 13 3/6'' for 9 5/8''

0.012 0.010 0.008 0.006 0.004 0.002 0.000 0 200 400 600 800 1000 1200 1400 1600 1800 2000
w ell depth [m ]

first bubble appears

Figure 6.23 F2, frictional pressure drop at 10 m distance Vs well depth

1.6 1.4 Sum frictional pressure drop [bar/m] 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0 200 400 600 800 1000 w ell depth [m ] 1200 1400 1600

Figure 6.24 F2, sum of frictional pressure drop Vs well

6. Case Studies well F2 154 __________________________________________________________________________

1.60 1.40
Sum frictional pressure drop [bar]

1.20 1.00 0.80 0.60 0.40 0.20 0.00 0 1 2 3 4 5 6 7 8


tim e [m ]

10

11

12

13

14

15

16

Figure 6.25 F2, sum of frictional pressure drop Vs time

17.5 line packing time differential 16.5


Pressure [bar]

1.6 1.4 1.2 1


time differential

17.0

16.0 single phase

0.8 0.6

15.5 0.4 15.0 diameter change 0.2 0 0 1 2 3 4 5 6 7 8 9 10 tim e [m ] 11 12 13 14 15 16 17

14.5

Figure 6.26 F2, estimated water hammer and line packing at the wellhead dp (together with time differential ) dt

6. Case Studies well F2 155 __________________________________________________________________________

240 220 200 180 160 140 120 100 80 0 0.1 0.2 0.3 0.4 0.5 void fraction 0.6 0.7

sonic velocity [m/s]

0m 100 m 200 m 300 m 400 m 500 m 600 m 700 m 800 m 830 m path

0.8

0.9

Figure 6.27 F2, calculated sonic velocities for varying pressure across the well depth
140

135

sonic velocity [m/s]

130

125
0m 100 m 200 m 300 m 400 m 500 m 600 m 700 m 800 m 830 m path

120

115

110 0.5 0.6 0.7 void fraction 0.8 0.9

Figure 6.28 F2, calculated sonic velocities for varying pressure across the well depth - scaled up.
The same plot scaled up. Acoustic velocity decreases although pressure increases across wellbore. It is caused by void fraction decrease which more affects acoustic velocity value than pressure for right-hand void values above 0.7.

Chapter 7
Discussion

7. Discussion 157 __________________________________________________________________________

This chapter discuss the results of the calculations performed in present work and shown in the form of plots. The theoretical explanation is given to the all effect observed on estimated line packing and conclusions are linked to the field measurements. The discussion includes the theoretical models used in calculations and verifies their limitations. The results obtained for both oil and geothermal wells are compared to each other in order to find whether there are other factors that may influence water-hammer and line packing in geothermal wells.
7.1 Multiphase flow correlations

In all well examples taken in considerations in present work both oil and geothermal, fluid enters the wellbore as single liquid phase. As the average pressure and temperature existing in tubing decrease towards the wellhead a second phase emerges due to gas coming out of the solution in the oil wells or steam flashing in geothermal wells at the bubble point pressure. For single phase flow physical scale and fluid property effects can be easy investigated using mathematical modelling. For two-phase flow this is not the case. Because of the increased number of flow conditions and flow parameters can only be obtained using wholly empirical and semi-empirical correlations and great care must be exercised when considering the results. In present work two correlations was used in order to predict the pressure gradient and flow patterns in simulated wells. The Duns and Ros correlation for oil wells and the Orkiszewski correlation for geothermal wells. These methods were recommended in literature to give the good pressure estimation in vertical upward flow for two-phase oil and geothermal well (Brill and Mukherjee, 1999) and (Uphadhay and Hartz, 1977) respectively. The Duns and Ros correlation used in PipeSim performs well and gives the results without discontinuities. The Orkiszewski method used in HOLA appeared to have a convergence problem in computing algorithm which results with discontinuities at the flow pattern transitions from bubble flow to slug flow. A sudden peak observed on the plots made for geothermal wells may be considered as the effect of these discontinuities. Some authors

7. Discussion 158 __________________________________________________________________________

suggest that coefficients in equations (B.41) and (B.45) may be modified so that the slopes of the curves are retained but discontinuities are eliminated. This solves the convergence problem but also may affect the accuracy of results (Brill and Mukherjee, 1999).

7.2 Acoustic velocity profile

As shown in Chapter 5 the density, compressibility and acoustic velocity are the fluid properties closely related with each other. As the properties changes across the wellbore with respect to the pressure and temperature changes it should be expected that acoustic velocity will be different from point to point in vertical well. The computed profile of acoustic velocity across the wellbores depth confirms this statement. In order to predict the sonic velocity calculations was made using formula developed by Gudmundsson and Dong (1993) and Woods (1944) for oil and geothermal two-phase wells respectively. Both formulas assume the homogeneity of the mixture and were reported to give a good estimation with measurements. The acoustic velocity in steam-water system is more complicated due to temperature effects and mass transfer between the phases. This brings theoretical limitations towards the Woods formula, however experiments of Semenov and Costern (1964) showed velocity in steam-water mixture agreeing with those estimated from Woods, and other sound speed models have also many limitations. Gudmundsson and Dong presented their method to be in very good agreement with measurements for gas-oil two-phase mixture (Gudmundsson and Dong, 1993). Speed of sound varies significantly across the two-phase length of the wellbore. In simulations these disparities are from 100 (m/s) at the wellhead to 1150 (m/s) at the bottom for oil wells, and likewise from 120 (m/s) to 1350 (m/s) for simulated geothermal wells. These differences are mainly due to different densities of oil, natural gas, water and steam. Also the fact that pressures are in different range for each of the wells is important as analysing the estimated acoustic velocities.

7. Discussion 159 __________________________________________________________________________

There are more discrepancies between acoustic velocity profiles estimated for oil and geothermal wells. The plots for oil wells show that there is not significant drop in sonic velocity as the crude enters the two-phase region. For geothermal wells these decrease in sound speed is less monotonically from pure water to the points of increasing void fraction of steam at lower pressure. Moreover in the case of well F1 and F2 where pressure at the wellhead is of relative lower value compared to other simulated wells, acoustic velocity decrease despite the pressure increase downwards the wellbore. In order to explain this behaviour this was plotted on the picture were the acoustic velocity estimated for varying pressures across the well profile was placed versus void fraction. These plots made for gas-oil and for geothermal wells compared each other reveal different behaviour at high void fraction values. For hydrocarbons two-phase mixture the acoustic velocity curves have less slope which becomes more flat as the pressure is higher. For steamwater mixture this shape is getting to be more inclined towards vertical. This more steep shape affects the sonic velocity. In the high void fraction region in the well, where steam occupies the main volume of the pipe, changes of void fraction value affect more acoustic velocity that pressure changes. This makes that acoustic velocity is decreasing downwards the well profile despite pressure increase. This effect is presented by the path on the mentioned plots. The acoustic velocity decreases until reach the region of the lower values of the void fraction, where the curves on the plot are more flattened. It is = 0.751 for well F1, and = 0.749 for F2.
7.3 Line packing

Line packing gives the information about pressure pulse propagation in the well. In order to estimate the line packing, frictional pressure drop across the wellbore needs to be calculated. The sum of the frictional pressure drop versus time gives the effect of line packing after valve closure. PipeSim gives the values of pressure drop for the every 50m distance in the case of oil wells. For geothermal wells this was calculated with 10m step, from Darcy-Weisbach equitation. It is obvious that this model will not give reasonable results over all flow patterns but if the actual flow is nearly homogenous, which is the case at high volume flowrates, then

7. Discussion 160 __________________________________________________________________________

predictions may be considered reasonable. Cornish (1976) applied the homogenous model to 10 different vertical oil wells and found that the percentage difference between calculated and measured total frictional pressure drop was on average around 2%. The time o the plots is the time it takes for pressure wave to travel from the wellhead to the point in question and back again i.e. up and down. This pressure wave travels at the speed of sound which is significantly lower in two phase region where also depends on void fraction. It makes the line packing unlinearized in the two phase region. These none linearly behaviour of line packing is clearer for low void fractions because of higher changes in sonic velocity (I refer again to the pots of calculated sonic velocities for varying pressure across the well depth). The frictional pressure drop is significantly higher for two-phase flow and increase as more gas phase arises. The plots also reveal the dependence the frictional pressure drop on diameter size. Diameter reduction causes the increase of frictional pressure drop. Both above effect may be observed on the plots with estimated line packing. For oil wells there is no tubing change thus only bubble point response affect calculated line packing. For geothermal wells two stage tubing was assumed in calculations and in this case response on both effect is visible. These effects are most distinguishable on the plots performed for well F1. The different line inclination may be observed. The first indicates the diameter change from 13 to 9 and the second is the response of bubble point. These effects are obviously visible on the all plots made in this work, but as calculations revealed bubble point in simulated geothermal wells tends to appear near the diameter change, and thus both effects overlap each other what makes them less distinguishable. The time differential of the pressure dp/dt placed together with line packing on the plots allows observing this effect. The transitions observed on the line pacing may be important for flow condition analysis in geothermal wells, to estimate the depth at which scale precipitation would commence for various wellbore diameters and mass flowrates. This can assist the engineer in the selection of operating conditions that will tend to cause scaling at shallower depth, thus required easer

7. Discussion 161 __________________________________________________________________________

clean-up operations (Ragnarsson, 2000). In the case of oil wells flow condition analysis may be used in order to optimize gas lift operations (Gudmundsson at al, 2001).
7.4 Size of the pressure pulse

Closure of a valve will create a pressure pulse. The water-hammer theory estimate the magnitude of the pressure pulse as: the product of the sonic velocity, the density, and flow velocity. In producing gas-liquid wells the pressure increase with depth and hence the mixture density and speed of sound also will increase. The hypothetical valve placed at an increasing up-stream distance will experience an increasing water-hammer. Theoretical water-hammer at the wellhead was examined from the calculations. The results are placed in the tables. The acceleration pressure (water-hammer) values calculated for oilgas two-phase wells are in range from 1.9 to 5.2 (bar) depending on the wellhead conditions. For geothermal well these values are significantly lower and are in range of 0.6 to 0.9 (bar), This is mainly due to wide tubing diameter that effect in decrease of fluid velocity which directly influence water-hammer.

Chapter 8
Conclusions

8. Conclusions 163 __________________________________________________________________________

1. Water-hammer and line packing were estimated for oil and geothermal wells from calculations performed in present work. PipeSim 2001 and HOLA 3.1 multiphase flow simulators were used to predict two phase flow in the oil and geothermal wellbores respectively. The simulations in present work incorporate correlations developed by Duns and Ros for oil wells and Orkiszewski for geothermal wells, coupled with equations for phase behaviour and wellbore heat-loss. Both models assume the changes in flow regimes and slippage between the phases. 2. The results confirmed the possibility to adapt the information contained in the rapid pressure transients when valve is activated to determine the mixture mass flowrate, density, velocity and gas void fraction. Additional calculations allow converting this data to the volume flowrates of each phase. Pressure Pulse Technology may be also used for flow condition analysis as the effect of bubble point and diameter change affect the line packing and are readily distinguishable in both oil and geothermal cases. 3. Pressure pulse propagation is closely related to sound velocity. Speed of sound varies significantly across the two-phase length of the wellbore. Calculations showed that velocities are from 100 (m/s) at the wellhead to 1150 (m/s) at the bottom for oil wells, and likewise from 120 (m/s) to 1350 (m/s) for simulated geothermal wells. These differences are mainly due to different densities of oil, natural gas, water and steam. Performed calculations revealed that there is not significant drop in sonic velocity as the crude enters the two-phase region in oil wells. For geothermal wells these decrease in sound speed is less monotonically from pure water to the points of increasing void fraction of steam at lower pressure. Moreover in the case of well F1 and F2 where pressure at the wellhead is of relative lower value compared to other simulated wells, acoustic velocity decrease despite the pressure increase downwards the wellbore. This effect is presented on the plots. Acoustic velocity decreases until reach the region of the lower values of the void fraction, where the curves in the plots are more flattened. It is = 0.751 for well F1, and = 0.749 for F2.

8. Conclusions 164 __________________________________________________________________________

4. The geothermal wells studied in present work are characterized by large tubing diameters. Typical values assumed in present work are 13 , 9 of producing casing, what may be considered as very unusual compared to 6.184, 5.125 and 4.5 taken for oil wells. This wide diameters result with low velocities of the flowing fluid. The water-hammer theory estimate the size of the pressure pulse as: the product of the sonic velocity, the density, and flow velocity. This acceleration pressure (water-hammer) values calculated for oil-gas twophase wells are in range from 1.9 to 5.2 (bar) depending on the wellhead conditions. For geothermal well due to large tubing size these values are significantly lower and are in range of 0.6 to 0.9 (bar). This fact may affect the accuracy of the measurements, as the waterhammer and line packing shape may be less distinguishable from usual pressure fluctuations at the well head.

165

References

9. References 166 __________________________________________________________________________

Agar J., Farchy D., Wet Gas Metering Using Dissimilar Flow Sensors: Theory and Field Trial Results, Agar Corporation, SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, 29 September-2 October 2002. SPE 77349 Aggour M.A. Al-Yousef H.Y., Al-Muralkhi A.J. Vertical Multiphase Flow Correlations for High Production Rates and Large Tubulars, SPE Production and Facilities, New Orleans, 1996, SPE 28465 Al-Taweel A.B, Barlow S.G Field Testing of Multiphase Meters, SPE Annual Technical Conference and Exhibition, Houston, 1999. SPE 56583 Ansari A.M., Sylvester N.D., Brill J.P. - Supplement to SPE 20630, A Comprehensive Mechanistic Model for Upward Two-Phase Flow in Wellbores- Pakistan Petroleum Ltd, SPE Journal,1994. SPE - 28671 Bjornnson G. Multi-Feedzone Geothermal Wellbore Simulator, Lawrence Berkley, report LBL -23546, USA, 1987 Brill and Mukherjee J.P., Mukherjee H. Multiphase Flow in Wellbores, SPE Monograph, 1999 Busaidi K., Bhaskaran H., New Development in Water Cut Meter with Salinity CompensationPetroleum Development Oman (PDO), Oman, J. Chen, Haimo Technologies, China, 2002. SPE 77894 Campbell Gas Conditioning and Processing Vol.1 The basic Principles, Campbell Petroleum Series, USA, 1994 Cornish R. E. The Vertical Multiphase Flow of Oil and Gas at High Rates Abu Dhabi Petroleum Co. Ltd., SPE Journal, 1976. SPE 5791

9. References 167 __________________________________________________________________________

Dickson M.H., Fanelli M. What is Geothermal energy?. Istituto di Geoscienze e Georisorse , Pisa, Italy, 2001. Dowty E. L., Hatton G. L., Durrett M.G., Dean T.L., The Genesis of a Three-Phase Subsea Metering System Texaco Inc., 1991 SPE 22983 Eliasson E. T. Power generation from High-Enthalpy Geothermal Resources, National Energy Authority-GHC Bulletin, Iceland 2001 Falcone G, Hewitt, G. F. Alimonti C. Multiphase Flow Metering: Current Trends and Future Developmnets, SPE Annual Technical Conference and Exhibition held in New Orleans, 2002. SPE 74689 Falk K. Bergant A. Dudlik I. Pressure Surges in Multiphase Media Norwegian Institute of Science and Technology, Trondheim, 1999 Firoozabadi A. Thermodynamics of Hydrocarbon Reservoirs, The McGraw-Hill Companies Inc. ISBN 0-07-022071-9, USA, 1999 Firoozabadi A. and Pan H. Two-Phase Isentropic Compressibility and Two-Phase Sonic Velocity for Multicomponent Hydrocarbon Mixtures Annual Technical Conference and Exhibition, San Antonio, Texas, 2000 SPE 65403

Glas O. Generalized Pressure Volume-Temperature Correlations, SINTEF/NTH,


Trondheim, 1980. SPE - 8018 Gomez L. E., Shoham O., Schmidt Z., Northug T. A Unified Mechanistic Model for Steady-State Two-Phase Flow in Wellbores and Pipelines, SPE Annual Technical Conference and Exhibition held in Houston, Texas, 1999. SPE 56520

9. References 168 __________________________________________________________________________

Gould T.L., Tek M.R. Steady and unsteady State Two-Phase Flow Trough Vertical Flow Strings, American Institute of Mining, Metallurgical, and Petroleum Engineers, 1970. SPE 2084 Grant M. A. and Sorey M. The Compressibility and Hydraulic Diffusivity of a WaterSteam Flow, Department of Scientific and Industrial Research, Vol. 15 No. 3, US, 1979 Grant M.A., Donaldson I.G., Bexley P.F. Geothermal Reservoir Engineering, Academic Press Inc., London, 1982 Gudmundsson J. S. Two-Phase Wells, Petroleum Engineering Department, Stanford University, California, 1986. Gudmundsson J.S., Notes about wells on Reykianes and Svartsengi fields, private communication, 2003. Gudmundsson J. S., Horne R. N., A. J. Menzies Streamtube Relative-Permeability Functions for Flashing Steam/Water Flow in Fractures, SPE Reservoir Engineering, 1986. Gudmundsson J. S., Thrainsson H. Power Potential of Two Phases Geothermal Wells, Geothermal Division, National Energy Authority, Reykjavik, 1988. Gudmundsson J.S, Ambastha A.K. Geothermal Two-Phase Wellbore Flow, Petroleum Engineering Department, Stanford University, California, 1986. Gudmundsson J.S. and Dong L. Model for Sound Speed in Multiphase Mixtures Department of Petroleum Engineering and Applied Geophysics, 3rd Lerkendel Petroleum Engineering Workshop, Trondheim, 1993. Gudmundsson J.S., Ambastha A.K. Pressure Profiles in Two Phase Geothermal Wells: Comparison of Field Data and Model Calculations, Petroleum Engineering Department,

9. References 169 __________________________________________________________________________

Stanford University, 1986 Gudmundsson, J.S., Brock, D.C. Reservoir Compressibility from Water Influx Modelling of Liquid Dominated Systems, Petroleum Engineering Department, Stanford University, California, 1989 Gudmundsson J.S., Celius H.K. Gas-Liquid Metering Using Pressure-Pulse Technology SPE Annual Technology Conference and Exhibition, Housto, Texas, 1999. SPE 56584. Gudmundsson J.S., Durgut I., Rnnevig J., Korsan K, Celius H.K., Pressure Pulse Analysis of Gas Lift Wells, ASME/API Gas Lift Workshop, Aberdeen, 2002 Gudmundsson J.S., Falk. K, Multiphase Pressure Pulses for Quick-Acting Valves: Offshore Testing, SPE, Department of Petroleum Engineering and Applied Geophysics, NTNU, 1999. SPE 56526. Gudmundsson J.S., Oritz R. Two-Phase Flow and Calcite Deposition SPE AIME, California, 1984. SPE 12741 Gudmundsson J.S., Sveinson S., Jonsson V. California, 1988 Hansen B. V. Operational Experience with Subsea Multiphase Flow Meters, Framo Engineering A.S, Multiphase Metering Conference, Aberdeen, 1997. Henry J. Ramey Jr., William E. Brigham H., Chen K., Atkinson D. - Thermodynamics and hydrodynamic Properties of Hydrothermal Systems, Stanford University, California, 1977 HOLA 3.1 Program Manual, Earth Science Division, Lawrence Berkley University of California, USA, 1993 Compressibility And Sonic Velocity in

Steam/Water Mixtures 13th workshop Geothermal Reservoir Engineering, Stanford

9. References 170 __________________________________________________________________________

Horne R.N. Modern Well Test Analysis A Computer Aided Approach, Stanford University, Petroway Inc., USA, 1990. Jamieson A.W. High Performance Multiphase Metering, Shell U.K., Exploration and Norflow Seminar, http://www.iceweb.com, 1999 Jonsson G. Simulations of line packing in an oil well, Department of Petroleum Engineering and Applied Geophysics, Trondheim, 1995. Khokhar A.A. Pressure Surge At High Velocity Wellbore Flow Dynamics, The Norwegian Institute of Technology, Department of Petroleum Engineering and Applied Geophysics, Trondheim, 1994 Kiefer S.W. Sound Speed in Liquid-Gas Mixtures: Water-Air and Water-Steam, Journal of Geophysical Research, Vol. 82, No. 20, 1977 King N.W. The Basis of the Multiphase-Flow, National Engineering Laboratory, Glasgow, 1990. King. N.W, Multiphase Flow in Pipeline System Its Transfer Measurement and Handling, National Engineering Laboratory, East Kilbride, Glasgow, 1999 Manabe et al. R.,Zhang H.-Q., Brill and MukherjeeJ.P. Crude Oil-Natural Gas Two Phase Flow Pattern Transition Boundaries at High Pressure Conditions, SPE Annual Technical Conference and McNeil D. A. Pipeline Design for Two - Phase Flow, National Engineering Laboratory, East Killbride, Glasgow, 1990. McWilliam D. and Duggins R. K. Speed of Sound in Bubbly Liquids, Department of Mechanical Engineering, University of Nottingham, Vol. 184, Pt 3C, UK, 1969

9. References 171 __________________________________________________________________________

Michaelides E.E., - Thermodynamics Properties of Geothermal Fluids, Department of Mechanical and Aerospace Engineering, University of Delaware, Newark, 1981. Millington B. Multiphase Flow Metering: A Review, National Engineering Laboratory, Oilfield Product Ltd. UK, 1999. Nakaryakow V.E., Pokusayev B.G., Shreiber I.R. Wave propagation in Gas Liquid Media, second edn. CRC Press Inc., Florida, 1993 Orkiszewski J. Predicting Two-Phase Pressure Drops in Vertical Pipe, Esso Production Research Co. Houston, 1967. SPE- 1546 Prats M. - The Effect of Heat Transfer Between Nearby Layers on the Volume of the Steam Zones, Michael Prats and Assocs. Inc., SPE Journal, 2001. SPE 71869 PipeSim 2000-Manual, Baker Jardine & Associates Petroleum Engineering and Software, London, 2000. Ragnarsson A. Geothermal Development in Iceland 1995 - 1999, Orkustofnun, Iceland, 2000 Roach G. J., Whitaker T. S. Long Term Use of A Multi Phase Flow Meter, Kwarner Oilfield Product Ltd. UK, 1999. Semenov N.I., Kostern S.I., -Results of Studies the Speed of Sound in moving gas-liquid systems, Teploenergetica, Thermal Engineering Journal 11(6), 1964 Stetfjeding E., Gudmundsson J. S., - Friction factor in high pressure natural gas pipelines from measurements, Department of Petroleum Engineering and Applied Geophysics, Norwegian University if Science and Technology, Trondeim, 1998.

9. References 172 __________________________________________________________________________

Steward D. Multiphase Metering Today, Offshore Engineer Journal, January 14, 2003 Stokes E.G., Mitchell M. H., Perry D.T., Halvorsen M., Application of The First Multiphase Flowmeter in the Gulf of Mexico, Conoco Inc., PetroTraces, Fluenta a.s., New Orleans, Louisiana 1998, SPE 49118 Theuveny B.C, Segueral G., Pinquet B., Multiphase Flow Meters in Well Testing Aplications, Schlumberger Oilfield Services, SPE Annual Technical Conference and Exhibition held in Louisiana, 2001. SPE 71475 Toral H., Cai S., Akartuna E., Stothard K. Neural Nets in Multiphase flow metering, PSL, Multiphase Metering Conference, Aberdeen 1999. Trusler J.P.M. Physical Acoustic and Metrology of Fluids, Adam Hilger, 1991 Uphady R.N., Hartz J.D. Tomkoria B.N. Comparison of Calculated and Observed Pressure Drops in Geothermal Wells Producing Steam-Water Mixtures SPE-AIME, Union of Oil Company, California 1977. SPE - 6766 Uphady R.N., Hartz J.D. Tomkoria B.N. Comparison of Calculated and Observed Pressure Drops in Geothermal Wells Producing Steam-Water Mixtures SPE-AIME, Union of Oil Company, California 1977. SPE 6766 Warren P.B., Hussain S., Ghamdi S., Background and Operational Experience of Multiphase Metering in the Safaniya Field-Offshore Saudi Arabia, Saudi Aramco, 2001, SPE 71534 Watters G.Z. Analysis and Control of Unsteady Flow in Pipelines, Butterworth Publishers, ISBN-0-250-40 492-3, California, 1978 William and McCain D. McCain, Jr., Analysis of Black Oil PVT Reports Revisited- SPE

9. References 173 __________________________________________________________________________

Annual Technical Conference and Exhibition held in San Antonio, 2002. SPE 69071 Williams J. Status of Multiphase Flow Measurements Research, Oryx Energy Co., SPE Annual Technical Conference and Exhibition held in New Orleans, 1994. SPE 28515 Wood A.B. A Textbook of Sound, Bell and Sons Ltd, London, 1944 Wylie E. B., Streeter V. L. Fluid Transients in Systems, Prentice Hall, New Jersey, 1993. Zabaras G. J. - Prediction of Slug Frequency for Gas/Liquid Flows, Shell E&P Technology Co., SPE Annual Technical Conference and Exhibition held in Houston, 1999. SPE - 65093

Appendix A
Multiphase Metering Projects

Appendix A Multiphase Metering Projects 175 __________________________________________________________________________

Following text contains a brief description of some MFMs projects that are now commercially available. The tables given below the text A.1, A.2, A.3, A.4, contain comparison of the methods with regard to the techniques that are used for measurement purposes.

Roxar AS MFI
Several elements are combined to measure the phase flow rates. The phase fraction sensor is mounted vertically with microwave transmitters and receivers protruding a few millimetres into the flow passage. In the fraction sensor, a resonant cavity is contained between two sets of microwave absorbers. Measurements of the frequency in the cavity are claimed to be proportional to water cut in the liquid phase. A single-energy gamma densitometer is used to measure the total mixture density. The gas fraction and the water cut are derived from this measurement. The mixture velocity is gained from cross-correlation between two microwave sensors located a known axial distance apart. (Stokes et al., 1998)

Roxar AS Fluenta
This meter uses several different sensors in combination. Capacitance and inductance sensors are used to measure bulk electrical properties of the flowing mixture in oil and water continuous flows respectively, with the water cut derived from these measurements. A single energy gamma densitometer measures the average mixture density by attenuation of gamma photons. The phase fractions can then be extracted from this information, as described earlier. Velocity measurement is by cross-correlation of capacitance signals for oil continuous flow, and by differential pressure across a venturi when the flow is water continuous. Velocity and phase fraction measurements are then combined to give phase flow rate information. (Warren, et al., 2001)

Appendix A Multiphase Metering Projects 176 __________________________________________________________________________

Kvearner-DUET
This flowmeter uses the attenuation of gamma rays at two different energies to derive the oil, water and gas phase fractions. The mass absorption coefficients of oil and water vary as a function of gamma photon energy and so the two different absorption rates and continuity relationship allow the phase fractions to be determined. To maximise the transmission of the lower energy gamma rays the sources and detectors are arranged around a GRP pipe section. Velocity measurement is by cross-correlation of two gamma densitometer signals, so it responds most accurately to distinct multiphase flow features like liquid slugs. (Roach, Whitaker, 1999)

Schlumberger-Framo
A mixer, which it is claimed gives both spatial and temporal mixing, is utilised to precondition the flow entering a venturi. The mixer consists of a large plenum chamber and piccolo tube. The piccolo tube penetrates the base of the plenum chamber and conducts the flow to the venturi, the aim being to draw the gas and liquid into the venturi at equal velocity. The differential pressure across the venturi is proportional to the total volume flow rate. A dual-energy gamma densitometer is mounted at the throat of the venturi and is used to derive phase fractions. (Hansen, 1997)

Jiskoot Mixmeter
The Mixmeter makes use of two separate radioactive sources to measure both the phase fractions of the multi phase mixture and the mixture velocity. An integral part of the flowmeter is a static mixer which conditions the mixture (spatially) so that an even distribution of the phases is maintained at the measurement cross-section. The phase fractions are determined by taking radiation attenuation measurements over a spectrum of energies. Passage of the gamma rays through the pipe is facilitated by the use of low absorption windows. Determination of the mixture velocity is by cross-correlation of photo detector signals received from two sources of equal energy, but mounted a known axial distance apart.

Appendix A Multiphase Metering Projects 177 __________________________________________________________________________

Differential pressure measurement across the static mixer is also used as secondary measure of flow velocity. (Dowty et al., 1991)

FlowSys
The FlowSys multiphase flowmeter uses two arrays of vertically mounted capacitance sensors, in what is effectively a very sophisticated liquid level sensing system. Two thin parallel plates are mounted axially in the flow passage, at a known distance apart. An array of capacitance sensors is mounted on the surface of each plate. The capacitance of the flow is thus measured at several vertical intervals in the flow passage. The gas/liquid interface is determined and the gas fraction calculated. The average water cut of the liquid is derived from the capacitance between the lower sensor pairs below the gas/liquid interface. The meter only operates in the slug flow regime, as this is a requirement for effective cross-correlation between the two sets of capacitance sensors. A slip correlation is then used to estimate the bulk gas and liquid velocities. The current version for oil continuous flow is being up-dated for operation in water continuous flows. (Al-Taweel, Barlow, 1999)

Agar
The flowmeter contains a rotary positive displacement flowmeter, modified for multiphase use, and two Venturis in series in a vertically upward flow. An algorithm in the control computer derives the gas and liquid volume flow rates from these outputs. The water content of the flow is derived from the power absorbed by the process fluid from an in-line microwave monitor, and the continuous liquid phase is detected by the phase shift between the transmitter and two differentially spaced aerials. The measurement of the liquid phase water cut can then be derived from the gas fraction and the microwave monitor output. Individual oil, water and gas flow rates are then computed from these parameters. (Agar, and Farchy, 2002)

Appendix A Multiphase Metering Projects 178 __________________________________________________________________________

ISA Scrollflow
Essentially this is a positive displacement flowmeter. It has two counter-rotating shafts which are machined to form a continuous and constant volume cavity, and the rotation created to the shafts by the fluid passing through the meter is claimed to be proportional to the total volume flow rate. A single-energy gamma densitometer located at the centre of the meter measures the overall mixture density. If the water cut of the liquid phase is known, then the phase flow rates can be determined from these measurements. At present the meter design does not include an integral water monitor to determine liquid water fraction. (Millington, 1999)

Haimo
The Haimo multiphase flowmeter, developed and used in China, uses an electromagnetic valve to periodically switch flow through a bypass trap where liquid collects. The velocity of this liquid passing through the meter is measured by cross-correlation between two singleenergy gamma densitometers. The gas velocity is derived using a slip relationship. (Busaidi et al., 2002)

Esmer
The Esmer multiphase flowmeter, currently under development, uses a neural network approach to interpret signals from capacitance/conductance sensors and pressure transmitters, combining measurements from 'training' of the meter to predict the expected values of phase velocities. (Toral, at al., 1999) Tables below contain comparison of the methods with regard to the techniques that are used for measurement purposes.

Appendix A Multiphase Metering Projects 179 __________________________________________________________________________

Table A.1 Methods utilizing Flow Conditioning

Flow Conditioning

Homogenisation

Leave-as-it-is

In-line separation

Jiskoot-Mixmeter SchlumbergerFramo

Schlumberger-VX Roxar AS Fluenta Roxar AS-MFI

Agar WellCamp Accuflow Kvearner-CCM Megra (for GVF > 25%) Haimo Jiskoot-Starcut

TEA-Lyra Kvearner-DUET ISA- Scrollflow Megra (for GVF < 25%) ISA-Solarton Esmer Yokogawa

Appendix A Multiphase Metering Projects 180 __________________________________________________________________________

Table A.2 Methods utilizing a Gamma Source

Gamma Source

Used

Not Used

Roxar AS Fluenta Schlumberger-Framo Schlumberger-VX TEA-Lyra ISA- Scrollflow Kvearner-DUET Megra Jiscoot-Mixmeter

Agar Esmer WellComp Jiskoot-Starcut Kvearner-CCM TEA-Vega ISA-Solarton TEA-Lyra (for WC < 25%) FowSys Yokogawa

Appendix A Multiphase Metering Projects 181 __________________________________________________________________________

Table A.3 Methods utilizing Intrusive methods

Intrusive

Yes

No(*)

Jiscoot-Mixmeter Agar Schlumberger-Framo Accuflow WellComp Kvearner-CCM ISA- Scrollflow ISA-Solarton TEA-Vega Jiskoot-Starcut Haimo

Schlumberger-VX Roxarn AS-Fluenta ROxar AS-MFI Kvearner-DUET Esmer Megra TEA-Lyra FlowSys Yokogawa

(*)

Venturis are not regarded as intrusive devices

Appendix A Multiphase Metering Projects 182 __________________________________________________________________________

Table A.4 Methods utilizing Cross Correlation

Cross Correlation

Yes Roxar AS-MFI Kvearner-DUET Roxar AS-Fluenta FlowSys Yokogawa Haimo

None Jiskoot-Mixmeter Agar Accuflow WellComp Jiskoot-Starcut Kvearner-CCM ISA-Scrollflow ISA-Solarton Esmer Megra Schlumberger-VX TEA-Lyra Schlumberger-Framo TEA-Vega

Appendix B
Duns and Ros, Orkiszewski - Multiphase Flow Correlations

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 184 __________________________________________________________________________ B.1 Duns and Ross correlation

Flow-Pattern Prediction
Figure B.1 shows the flow-pattern map developed by Duns and Ros. They identified four separate regions for computation purposes, Regions I through III and a transition region. Duns and Ros also identified the heading region as a fifth region, but this is now considered part of Region II. In this work I will refer to Regions I through III as bubble, slug, and mist flow, respectively.

Figure B.1 Duns and Ros Flow Patterns Map The flow-pattern transition boundaries are defined as functions of the dimensionless gas and liquid velocity numbers NGu and NLu. For these transition boundaries, Duns and Ros proposed these equations. Bubble/slug boundary:

N G u B / S = L1 + L2 N L u

(B.1)

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 185 __________________________________________________________________________

where L1 and L2 are functions of Nd dimensionless diameter number. The gas and liquid velocity numbers are given as:

N Lu = uL 4

L g L G g G

(B.2)

and
N G u = uG 4

(B.3)

Where u is the superficial velocity, g gravity acceleration, and density of the gas and liquid indicated with subscripts. Slug/transition boundary

N G u S / Tr = 50 + 36 N L u
Transition/mist boundary

(B.4)

N G uTr / M = 75 + 84 N L u 0,75

(B.5)

Liquid holdup prediction: Duns and Ros chose to develop empirical correlations for a dimensionless slip-velocity number - NKu, rather than for liquid holdup. NKu is defined in similar way to the gas and liquid velocity numbers.

N K u = uK 4

L g L

(B.6)

The slip velocity was defined as


u K = uG u L = uK G u KL 1 HL HL

(B.7)

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 186 __________________________________________________________________________

where HL is the liquid holdup. Pressure gradient contains three components: gravitational, frictional, and accelerational. dp dp dp = + K g + dz dz f dz acc

(B.8)

In order to calculate the gravitational pressure gradient following procedure was proposed - Calculate the dimensionless slip velocity NKu, using the appropriate correlation. The correlations for NKu are different for each flow pattern and are given later. - Solve the equations given above for slip velocity. - Calculate the liquid holdup from equation given above. - Calculate the slip density from equation (B.9)

K = L H L G (1 H L )

(B.9)

- Calculate the elevation component of the pressure gradient as given in equation (B.8) Flow pattern identification: - Bubble flow exists if NGu < NGuB / S For Bubble flow, the dimensionless slip-velocity number is given by
NGu N K u = F1 + F2 N L u + F3 ' 1+ N u L

(B.10)

(B.11)

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 187 __________________________________________________________________________

where F1 and F2 are function of the liquid velocity number NL. F3 can be obtained from
F4 Nd

F3 ' = F3

(B.12)

where F4 and Fd are also the functions of NL. The friction pressure-gradient component for bubble flow is given by f L uL uM dp = 2d dz f

(B.13)

From experimental data Duns and Ros developed the equation for friction factor:
f2 f3

f = f1

(B.14)

The friction factor is governed mainly by f1, which is obtained from a Moody diagram Fig. (B.2) as function of Reynolds number for the liquid phase.

Re =

L uL d L

(B.15)

The factor f2, is a correction for the in-situ gas/liquid ratio. The factor f3 is considered by Duns and Ros as second-order correction factor for both liquid viscosity and in-situ gas/liquid ratio. It becomes important for kinematics viscosities greater than approximately 50 cSt (0,744 Pas) and is given by

f3 = 1 +

uG f1 4 50 u L

(B.16)

Duns and Ros considered the acceleration component of the pressure gradient to be neglected for bubble flow.

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 188 __________________________________________________________________________

Slug flow exist if N G u B / S < N G u < N G u S / Tr (B.17)

For slug flow the dimensionless slip-velocity umber is N G u 0,982 + F6 ' (1 + F7 N L u )

N K u = (1 F5 )

(B.18)

where F5, F6 and F7 are function of the liquid viscosity number NL, and F6 ' = 0,029 N d + F6 (B.19)

Figure B.2 Moody diagram

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 189 __________________________________________________________________________

The friction pressure gradient component for slug flow is calculated exactly the same way as the bubble flow. Also, the accelerational component slug flow is considered negligible. Mist flow exist if N G u > N G uTr / M (B.20)

Duns and Ros assumed that, at high flow rates, the liquid is transported mainly as small droplets. The result is nearly a no slip condition between the phases. Thus NKu = 0 and uK = 0 and HL = (1-). The mixture density for use in elevation component of the pressure gradient then is calculated from

M = G + (1 ) L

(B.21)

Friction in the mist flow pattern originates from the shear stress between the gas and the pipe wall. Thus, the friction component of the pressure gradient is determined from
f G uG dp = 2d dz f
2

(B.22)

Because there is no slip, the friction faction is obtained from a Moody diagram Figure (B.2) as a function of a Reynolds number for gas phase

Re =

S uG d G

(B.23)

Duns and Ros noted that the wall roughness for the mist flow is the thickness of the liquid film that covers the pipe wall. Waves on the film cause an increased shear stress between the gas and the film that, in turn, can cause the greatest part of the pressure gradient. These waves result from the drag of the gas deforming the film in opposition to the surface tension. This process is affected by viscosity and also is governed by a form of the Weber number

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 190 __________________________________________________________________________

N We

G uG 2 = L

(B.24)

This influence was accounted for by making a new function of dimensionless number containing liquid viscosity.

N =

L2 L L

(B.25)

The value of roughness may be very small, but the relative roughness never becomes smaller than the value of the pipe itself. At the transition to slug flow, the waviness may become large, with the crests of opposite waves touching and forming liquid bridges. Then /d approaches 0.5. Between these limits, /d can be obtained from equations

NWe 0,005;

0,0749 L G uG 2 d

(B.26)

and
N We > 0,005;

0,0371 L

G uG d
2

(N We N )

0 , 302

(B.27)

where d is in [ft] unit, and uG is in [ft/s] G in pounds per cubic foot and in dynes per centimetre. Values of f for the mist flow-pattern can be found for /d > 0,005 from extrapolation of the Moody equation.
1, 73 1 f = 4 + 0,067 d 0,27 4 log d

(B.28)

As the wave height of the film on the pipe wall increases, the actual area variable for gas decreases because the diameter open to flow of gas in now d-. Duns and Ros suggested that

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 191 __________________________________________________________________________

the friction component of the pressure gradient could e refined by replacing d with d- and uG with uG
d 2

/( d-)2 throughout the calculations. This results a trial-and error procedure to

determine the . In mist flow, acceleration often can not be neglected as it was in bubble and slug flow. The accelerational component of the pressure gradient cab be approximated by
u M u G M dp dp = p dz dz acc

(B.29)

The derivation f this equation provided by Beggs and Brill and Mukherjeecan be found in Brill and Makherjee monograph (1999). Transition region exists if N G u S / Tr < N G u < N G uTr / M (B.30)

If this region is predicted, Duns and Ross suggested linear interpolation between the flowpattern boundaries, to obtain the pressure gradient. This will require a calculation of pressure gradient with both slug-flow and mist-flow correlations. The pressure gradient in the transition region then is calculated from dp dp dp = A + (1 A) dz mist dz slug dz f

(B.31)

Increased accuracy was claimed in the transition region if the gas density used in the mistflow pressure-gradient calculations was modified to be
NGu N G uTr / M

G ' = G

(B.32)

where G is the gas density calculated at the given conditions of pressure and temperature. This modification accounts for some of the liquid being entrained in the gas.

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 192 __________________________________________________________________________ B.2 Orkiszewski correlation

- Bubble/slug transition:

B / S = LB
2

(B.33)

u L B = 1,071 0,2218 M d

(B.34)

where uM is in [ft/s] and d in [ft]. LB is constrained algebraically to be 0.13. - Bubble flow exist if

LB
The liquid holdup for bubble flow is determined from
2 uM u 1 uM 4 G HL =1 1+ 1+ 2 uK uK uK

(B.35)

(B.36)

which is equivalent to equation for Duns and Ros correlation. Orkiszewski adopted the Griffith suggestion that 0.8 [ft/s] is a good approximation of the average uK that is function of the gas liquid densities and surface tensions (Brill, 1999). The liquid holdup determined from above equation is then used to calculate slip density with equation (B.9) which in turn is used to calculate the elevation component of pressure gradient. The friction pressure-gradient component for bubble flow is given by
uL f L H dp L = 2d dz f
2

(B.37)

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 193 __________________________________________________________________________

The friction factor is obtained from a Moody diagram Figure (B.4) as function of relative roughness and Reynolds number for the liquid phase
2

Re L =

uL L H L

(B.38)

The acceleration pressure-gradient component for bubble flow was considered negligible. - Slug flow Slug flow exist if inequalities are satisfied

>B/S
and N G u < N G u S / Tr The slip density is calculated from

(B.39)

(B.40)

K =

L (u L + u B ) + G u G
uM + uB

+ L

(B.41)

Orkiszewski developed above equation (B.41) by performing mass and volume balances on a typical slug unit consisting of a bubble and liquid slug. A similar Griffith and Wallis development neglected the presence of a liquid film around the bubbles ad the possibility of liquid droplets being entrained in the bubbles. Consequently Orkiszewski used the data of Hagedorn and Brown and proposed the last term in equations (B.41), to account for the distribution of liquid in this region. This modification was meant to extend the Griffith and Wallis work to include the high-velocity flow range. Griffith and Wallis correlated the bubble-rise velocity uB by the relationship
u B = C1 C 2 g d

(B.42)

Appendix B Duns and Ros, Orkiszewski Multiphase Flow Correlations 194 __________________________________________________________________________

where C1 and C2 are the functions of ReB and ReL. The precise procedure of uB calculations including iteration process for higher Reynolds number may be found in Brill and Mukherjeeand Makherjee monograph (1999).

Re B =

L uB d L L uM d L

(B.43)

and
Re L =

(B.44)

The friction pressure-gradient is given by


f L uM uL + uB dp = u +u + 2 g d dz f B M
2

(B.45)

The friction factor f is calculated from the Moody diagram using the following definition of Reynolds number
1444 L u M d

Re B =

(B.46)

Pressure drop due to acceleration is neglected in slug-flow pattern. For transition and mist flow regime Orkiszewski uses the Begs and Brill methods that are described above in previous section of the Appendix B.

Appendix C
Sound Wave Propagation Process in Steam Water Mixture

Appendix C Sound Wave Propagation Process in Steam Water Mixture 196 __________________________________________________________________________

The complex physical process that occurs during propagation of sound wave in water-vapor two-phase system was described by (Kieffer, 1977). Consider the temperature-entropy diagram of water shown above on Figure (5.7). A mixture of saturated water and steam is represented as point G where the chord ratio FG/FH is the mass fraction x of steam and the mixture. Isentropic pressure changes, such as the compression and rarefaction which occurs during propagation of sound wave are presented by movement up and down the constant entropy line CGK. If steam and water remain in thermal equilibrium on the saturation line, there must be mass transfer between the phases, since the fraction of steam in mixture changes (BC/BD FG/FH IK/IM). This requires that condensation or evaporation take place. An isentropic pressure increase from p to p + p corresponds to movement of the mixture from G to C in the temperature-entropy diagram. The pressure increase in the water phase corresponds to movement from F to A; as a result, the water phase becomes subcooled. The pressure increase in the steam phase corresponds to movement from H to E as the steam becomes superheated. The induced temperature difference between the steam and the water leads to heat transfer from the superheat steam to subcooled water. If the original composition of the mixture G lies to the right of the peak of the two-phase loop as shown in (Figure 5.7), some water will be vaporized and the mass fraction of steam in the mixture will increase during adiabatic compression (BC/BD > FG/FH). If the original composition lies to the left of the top of the two-phase loop (assumed to be symmetric), some steam will condense, and the mass fraction of steam in the mixture will decrease during adiabatic compression. Thus by heat and mass transfer both water and the steam are restored to the saturation line, the water by the path A to B, and the steam by the path E to D. An isentropic pressure reduction from p to p - p corresponds to the movement from G to K in the temperature - entropy diagram. The pressure decrease in the water phase corresponds to movement from F to R; as result, the water becomes superheated above its saturation temperature point I (IR is a continuation of the pressure line from the water region). The pressure decrease in the steam phase alone corresponds to movement from H to N on the diagram, and the steam becomes subcooled or supersaturated with respect to its saturation temperature, point M (MN is a continuation of a constant pressure line from the superheated region). If the original composition of the mixture G lies on the right side of the two-phase

Appendix C Sound Wave Propagation Process in Steam Water Mixture 197 __________________________________________________________________________

loop, some vapor will condense to form a mixture of saturated water and steam (point L), and thus the subcooled steam may move toward the stable state M. The mass fraction of steam in such a mixture will decrease (IK/IM < FG/FH). If the original composition lies to the left of the two-phase loop (assume to be symmetric), cavitation, and some vaporization of the subcooled liquid will be dominant process and the mass transfer of steam in the mixture will increase. Cavitation of the water creates a local mixture with the composition of point J and the superheated water at point R thus moves towards the stable state I, a mixture of water and steam.

Figure C.1 Temperature entropy diagram for H2O

Appendix D
PipeSim 2000-Multiphase Flow Simulator

Appendix D PipeSim 2000 - Multiphase Flow Simulator 199 __________________________________________________________________________ D.1 PipeSim Well Performance Analyses

PipeSim is the program developed by Baker Jardine & Associates, London. The program allows predicting the relationship between flowrates, pressure drop, and piping geometry (length, diameter, angle, etc.) for the fluids produced from reservoir. Individual PipeSim modules are used for a wide range of analyses including: well monitoring, nodal analysis, artificial lift optimalisation, pipeline and process facilities modelling, and field planning. A major future of PipeSim 2000 is the system integration and openness that allows constructing the total production model from reservoir to processing facilities. The model can be either operated hooked up to the reservoir simulator, such like Eclipse, or with a simplified material balance. The simulations can track multiphase issues such as slugging, fluid velocities, or thermal performance and also be used to monitor topside equipment such as compressor power requirements. Program allows deciding between the customary and SI units and enables the different input and output units depending on which are more convenient for calculations.
D.1.1. Fluid Properties Correlations

At the beginning the program requires the fluids properties to be specified. To predict pressure and temperature changes from the reservoir along the wellbore (or flow line tubular), it is necessary to predict fluid properties as a function of pressure and temperature. PipeSim has a PVT section which can generate fluid properties using standard correlations and allows them to be modified. In present work The Black Oil Glas correlation has be used to calculate the fluid properties. The Black oil correlations futures are described in chapter 5 of this work which deals with Multiphase flow in wells. Glas PVT correlation has been recommended for light North Sea crude by PipeSim Manual authors (PipeSim Manual) developed from analysis of crude from the following fields: Ekofisk, Statfiord, Forties, Valhall.

Appendix D PipeSim 2000 - Multiphase Flow Simulator 200 __________________________________________________________________________

Figure D.1 Black Oil PipeSim input data window The mathematical relationship between bubble point and solution gas oil ratio Rs given by Glas is
API 0.989 R s = G 0.172 p B T
1.2255

(D.1)

where T is in oF, API = specific gravity of oil

141.5

131.5 , G is the specific gravity of gas (air = 1), O is the

The bubble point pressure is given by the relation


0 .5 p B = 10 (2.8869 (14.18113.3093log ( p )) )

(D.2)

The other models available for Black Oil correlations in program are: Lasater, Standing, Vasques and Begg. The Standing correlations are widely used in oil industry. They are based primarily on California crude oils and these correlations do not correct for other oil types or

Appendix D PipeSim 2000 - Multiphase Flow Simulator 201 __________________________________________________________________________

hydrocarbon content. Glas modified the Standing correlation to make the independent of oil type (Glas, 1980). Glass oil formation volume factor correlation can be expressed mathematically as Bo = 1.0 + 10 A

(D.3)

where
A = 6.58511 + 2.91329 log B * ob 0.27683(log B *ob )
2

(D.4)

and B*ob is correlating number defined by


0.526

G B *ob = R s O

+ 0.968 T

(D.5)

Glas also presented the bubble point pressure. The proposed bubble point pressure

correlation can be expressed as (Glaso, 1980)


log p b = 1.7669 + 1.7447 log p *b 0.30218 (log p *b )
2

(D.6)

with p* b a correlating number defined by


0.816

Rs p *b = G

T 0.172

API 0.989

(D.7)

where Rs is in scf/STBO, T in oF, and G is the total specific gravity of the surface gases. Optionally the program allows defining the viscosity data for the required pressure and temperature ranges. The viscosity of crude oil with dissolved gas is an important parameter in pressure-loss calculations for flow in pipes. There are four steps to calculating the liquid viscosity as follows:

Appendix D PipeSim 2000 - Multiphase Flow Simulator 202 __________________________________________________________________________

Step 1: Calculate the dead oil viscosity at atmospheric pressure and the flowing fluid temperature. Step 2: Calculate the saturated live oil viscosity at the flowing fluid pressure and temperature assuming that the oil is saturated with dissolved gas. Step 3: Establish if the flowing pressure is above the bubble point pressure for the flowing fluid temperature. viscosity. Step 4: Determine the viscosity effects of water in the liquid phase. If not, continue to step 4, otherwise calculate the undersaturated oil

Figure D.2 Viscosity data input window The dead oil viscosity can be also calculated from Glas Correlation. He presented an empirical correlation based on North Sea data (Glaso, 1980). The correlation can be expressed as

od = (3.141 1010 ) T 3.444 (log API )10.313log T 36.447


where T is in oF.

(D.8)

Appendix D PipeSim 2000 - Multiphase Flow Simulator 203 __________________________________________________________________________

Saturated live oil viscosity depends on the solution-gas content. Oil viscosity decreases with rising pressure as the solution gas increases, up to the bubble point pressure. There are few correlations to determine the viscosity of saturated oil system, the Begs and Robinson correlation has been chosen in present work. The empirical form of this equation is (Brill, 1999)

O = 10.715 (R s + 100)0.515 od b
where
b = 5.44 (Rs + 150 )
0.338

(D.9)

(D.10)

Above the bubble point pressure, rising pressure increases the viscosity o foil because of its compressibility. The correlation proposed by Vasquez and Beggs is the one of available in PipeSim to calculate the undersaturated oil viscosity. p p b
a

O = Ob

(D.11)

where
a = 2.6 p 1.18710
b

(D.12)

and
b = (3.9 10 5 ) p 5
D.1.2 Advanced calibration data

(D.13)

In many cases, actual measured values for some properties show a slight variance when compared with the value calculated by the black oil model. In this situation it is useful to "calibrate" the property using the measured point. PipeSim can use the known data for the property to calculate a "calibration constant" Kc .

Appendix D PipeSim 2000 - Multiphase Flow Simulator 204 __________________________________________________________________________


KC = measured property ( P, T ) calculated property ( P, T )

This calibration constant is then used to modify all subsequent calculations of the properties. Properties which may be calibrated in this manner are; oil formation volume factor, gas compressibility, live oil viscosity, gas viscosity. Additionally PipeSim allows entering the fluid composition data
D.2 Profile model

I these work the vertical well performance analysis model has been selected for simulations. The other models available in the program are; injection well performance, and surface and facilities model. The program requests only the well hardware data required by the option which was selected. This future is convenient because for particular section calculations full model do not have to be built up, and I present work the surface equipment could be skipped. The main window displays the designed model Figure D.3 and allows for easy access to each of the section in order to model the required parameters.

Figure D.3 Designed model

Appendix D PipeSim 2000 - Multiphase Flow Simulator 205 __________________________________________________________________________

The profile model section of the program allows describing the production tubing, SSSV (Surface - Controlled Subsurface Safety Valve, Figure D.4), and restriction. Downhole equipment parameters may be specified in program using one of the two options available, the simple and detailed model. The displayed screed is different depending on the option selected. Using the detailed model more factors to be taken into calculations but obviously requires more data to be entered.

Figure D.4 Surface - Controlled Subsurface Safety Valve


D.2.1. Detailed model

The detailed profile contains following sections: - Deviation survey This section has been left empty in present work as the vertical well model has been assumed. If such deviation data are available it may be copied to the program from other spreadsheet applications.

Appendix D PipeSim 2000 - Multiphase Flow Simulator 206 __________________________________________________________________________

- Geothermal survey The geothermal survey and overall heat transfer coefficient need to be specified in the detailed model. The program request at least two values at the wellhead and at the bottom. The heat transfer coefficient describes the resistance to heat flow by all mechanisms (convection, radiation and conduction) from the well to its surroundings. The program proposes the typical default values of the coefficient that maybe used in calculations. Temperature prediction may be performed using enthalpy balance temperature model. This requires to define the well environment including all casing, string, cement tops, formation lithology etc. -Tubing configuration The data entered in this section determines extend of tubing modelled. The program takes the bottom of the last tubing as the fluid inflow point. Program offers the tubing tables containing standard values of tubing and casing diameters. Also the wall thickness and roughness may be specified using values given in the PipeSim tubing tables. Casing inner diameters are only required if the flow is set to be annular or both tubing and annulus. -Downhole equipment This option allows specifying the necessary data depending on the equipment was selected. Program allows to set up the following optional equipment properties: gas lift injection points, ESP (Electrical Submersible Pumps), Choke, separator G/L (Gas/Lift) valve system, SSSV. Perforations data may be also entered in this mode.
D.2.2. Simplified model

For present work simple model was utilized Figure D.5. This model requires the limited number of information compared to the detailed motel described above. Deviation survey is replaced by the one deviation angle from horizontal, where the kick off depth may be

Appendix D PipeSim 2000 - Multiphase Flow Simulator 207 __________________________________________________________________________

specified. Temperate data are limited to the surface ambient and reservoir (well bottom). Tubing limited up to the four sections, and if more are required detailed profile need to be used. The downhole equipment assumed in simple mode contains only SSSV, G/L vale and perforation data. The accuracy of the equipment description may by verified by making an equipment summary. The program draws and displays the simple sketch of designed downhole or surface equipment. Figure D.5 Simple profile mode

Figure D.5 Simple profile mode


D.3 IPR Data

In present work the one oil well the IPR curve was also calculated using PipeSim simulator. This requires additional data defining the reservoir inflow performance. The program allows for thirty inflow options including the oil and gas and gas condensate reservoirs, to model the flow of fluids from the reservoir, trough the formation, and into the well. The choice which of them should be used depends on the available information and the type of sensitivities that

Appendix D PipeSim 2000 - Multiphase Flow Simulator 208 __________________________________________________________________________

are supposed to be run. This is beyond the scope of this work to present them all and there is the bulk of the literature which describes those models. The average reservoir pressure and reservoir temperature must be entered for inflow performance models; however both the Multi-rate Fetkovich and Multi-rate Jones models can be used to calculate the reservoir pressure. Well skin can be either directly entered or calculated using Locke, Macleod or Karakas and Tariq methods for mechanical geometrical skin, and the Cino/Martia -Bronz or Wong - Clifford method for a deviation-partial penetration skin. The Elf - Skin Aide model is also available. The calculated IPR can be matched to measured data and used to calculate the IPR pressures for any rate and water cut value. Also relative permeability can be applied to all IPR model in PipeSim, and thus the total mobility of oil, gas and water may be determined. For calculations in present work the Darcy inflow equation has been selected above the bubble point and the Vogel solution below the bubble point. Required input is: reservoir permeability (total permeability at the prevailing watercut and Rs), reservoir thickness (thickness of the producing reservoir rock), - drainage area, - wellbore radius, Dietz shape factor (to account for the shape of the drainage area). The additional data for simulations has been taken from Kleppe Manuscript (1990), where author gives the characteristic of the most Norwegian North Sea fields. The other mentioned models available in the program take into considerations the deliverability change with time, non Darcy flow, multi-rate flow, hydraulically fractured well, dual porosity, horizontal and multi layer wells.

D.4. Matching option

The matching option allows user for data quality control and fine adjustment of model parameters to enable well model to produce observed data. Appropriately matched model is requested for accurately performance prediction. The matching option offers the following calculation options: Correlation comparison.

Appendix D PipeSim 2000 - Multiphase Flow Simulator 209 __________________________________________________________________________

This option allows pressure gradient plots to be generated with different correlations to be compared with the measured gradient survey data and each other. VLP and IPR matching

This option enables to tune the wellbore multiphase flow correlations to fit a range of measured downhole pressures and rates. Once the VLP is matched, the IPR can be adjusted to match the observed rates and pressures also. Gradient matching

Exits correlations can be modified using non-linear regression to best fit a gradient survey. Comparison of the fit parameters will identify which correlation requires the least adjustment to match the measured data. Surface pipe matching

The PipeSim program uses actual wellhead and manifold pressures together with temperature data points to match the surface pressure drop correlations. Separate screens allow the match parameters to be viewed and the best match selected. Tubing/pipeline correlations parameters

The match parameters can be inspected, reset or entered by and using this option. This capability may be useful for troubleshooting, or to input match parameters determined previously. Correlation thresholds

This is a capability of the program that allows specifying a threshold angle for both tubing and pipeline correlations at which the program will automatically change to another specified correlation. This option is particularly dedicated to the vertical risers in subsea completions to be modeled more accurately.

Appendix D PipeSim 2000 - Multiphase Flow Simulator 210 __________________________________________________________________________ D.5 VLP correlations and applications

It is no universal rule for selecting the best flow correlations for a given application. It is recommended that correlation comparison always be carried out (PipeSim Manual). By inspecting the predicted flow regimes and pressure results, the user can select correlation that best models the physical situation. In present work for oil wells calculations the Duns and Ros correlation was selected. This correlation is recommended by PipeSim authors (PipeSim Manual) for vertical well calculations. The following calculations are compared below PipeSim contains also the correlations developed by Shell EP - Technology and Application Research, Those correlations are described t combine the best futures of the existing correlations (Zabaras, 2000), and also contains additional futures for predicting low-rate VLPs well stability or for viscous, volatile, ad foamy oils. These correlations could not be used as requires additional extra license for its use. Duns and Ros

This correlation was precisely described in previous chapter 5. The description presents the basis of the original Duns and Ros published method. PipeSim contains the primary correlation that was enhanced and improved for use wit condensates. The Duns and Ros correlation performs well in mist flow cases and may be used in high Rs oil wells (Brill, 1999). It tends to overprotect VLP in oil wells. Despite this, the minimum stable rate indicated by the VLP curve is often a good estimate (PipeSim Manual). Fancher Brown

It is a no slip holdup correlation that is provided in use as a quality control. It gives the lowest possible value of VLP since it neglect the gas-liquid slip. It should always predict a pressure which is less than he measured value. Even if it gives good match to measured downhole pressure Fancher Brown should not be used on the quantitative work. Measured data falling

Appendix D PipeSim 2000 - Multiphase Flow Simulator 211 __________________________________________________________________________

into the left of the Fancher Brown on the correlation comparison plot indicates the problem with fluid density or field pressure data (PipeSim Manual). Orkiszewski

This correlation often gives a good match to measured data. However its formulation includes a discontinuity in the calculation method. The discontinuity can cause instability during the pressure matching process, thus is not recommended for calculations (PipeSim Manual). Begs and Brill

The correlation is primarily horizontal pipeline correlation (Brill, 1999), thus was not taken into account for wellbore simulations. It generally over predicts pressure drops in vertical and deviated wells.

- OLGA-S steady-state OLGA-S is based in larger part on data from the SINTEF two-phase flow laboratory near Trondheim, Norway. The test facilities were designed to operate at conditions that approximated field conditions. OLGA-S considers four flow regimes, stratified, annular, slug and dispersed bubble flow and uses unique minimum slip criteria to predict flow regime transitions. This correlation is available to all members of the SINTEF syndicate, and to nonmembers on payment of the appropriate fees. - Govier and Aziz The correlation is used for pressure loss, holdup, and flow regime. The Govier, Aziz & Fogarasi correlation was developed following a study of pressure drop in wells producing gas and condensate and is thus recommended for them. The new prediction method incorporates an empirical estimate of the distribution of the liquid phase between that flowing as a film on

Appendix D PipeSim 2000 - Multiphase Flow Simulator 212 __________________________________________________________________________

the wall and that entrained in the gas core. It employs separate momentum equations for the gas-liquid mixture in the core and for the total contents of the pipe. BJA correlation

Baker Jardine & Associates have developed a correlation for two phase flow in gascondensate pipelines with a no-slip liquid volume fraction of lower than 0.1. This model represents no major advance in theory, but rather a consolidation of various existing mechanistic models, combined with a modest amount of theoretical development and field data testing (PipeSim Manual). The model uses the Taitel Dukler flow regime map and a modified set of the Taitel Dukler momentum balance to predict liquid holdup. The pressure loss calculation procedure is similar in approach to that proposed by Oliemans, but accounts for the increased interfacial shear resulting from the liquid surface roughness. The BJA correlation is not recommended for systems having a non-slip liquid volume fraction greater than 0.1. The quite extensive testing of the correlation against operating data has been undertaken for horizontal and inclined flow, the test data for vertical flow is not so comprehensive. (PipeSim Manual). Ansari:

The Ansari model was developed as part of the Tulsa University Fluid Flow Projects (TUFFP) research program. A comprehensive model was formulated to predict flow patterns and the flow characteristics of the predicted flow patterns for upward two-phase flow. The comprehensive mechanistic model is composed of a model for flow pattern prediction and a set of independent models for predicting holdup and pressure drop in bubble, slug, and annular flows. The basic plot and data fro calculations can be written to an Excel file by specifying one of the Excel options for the output file format. This will then create an Excel data file called with the extension .plt

Appendix E
HOLA 3.1 - Multiphase Flow Simulator

Appendix E HOLA 3.1 - Multiphase Flow Simulator 214 __________________________________________________________________________ E.1. Introduction

The geothermal steam-water flow flowing temperature and pressure profiles has been simulated using HOLA 3.1 wellbore geothermal simulator. The simulator can also determine the relative contribution of each feedzone for a given discharge conditions. The flow within the well is assumed steady-state at all times, but time changing reservoir pressures are allowed. Te HOLA 3.1 simulator uses several files in its computations. The some of them are created by the program but others are read only and must be construed as a text file with DOS acceptable names and with program acceptable format. HOLA also creates the calculations and iterations output files. The first contains pressure and temperature profile including related parameters change across the wellbore. The second provides information on the iterations executed by HOLA in order to obtain the above output file. Plot of calculated and measured pressure and temperature can be drawn by the program. This option enables to view the results of the simulations.

The simulator can handle both single and two-phase flows in vertical pipes and calculates the flowing temperature and pressure profiles in a well. It solves numerically the differential equations that describe the steady-state energy, mass and momentum flow in a vertical pipe. The code allows for multiple feedzones, variable grid spacing and radius. The code was developed in the Fortran programming language by Lawrence Berkley Laboratory, University of California. A detailed description of the formulation used in the simulator is given in a separate report by Bjornsson (1987). A reference to this publication is made frequently in the following text.

Appendix E HOLA 3.1 - Multiphase Flow Simulator 215 __________________________________________________________________________

The simulator is consistent with SI units in all calculations except for pressure (bar instead of
Pa) and enthalpy (kJ/kg instead of J/kg).

Before downhole calculations are performed with the simulator HOLA 3.1, a definition of all input parameters is required. The following text gives the program description, and also shows the approaches that were used in order to solve the program deficiency problems and calculate required parameters.

E.2 Governing Equations

The flow of fluid in a geothermal well can be represented mathematically by two sets of equations. Between the feedzones, the flow is represented by one-dimensional steady state momentum, energy and mass flux balances. When a feedzone is encountered, mass and energy balances between the fluid in the well and the feedzone are performed. The solution of these equations require fully defined flow conditions at one end of the system (inlet conditions) and fully defined boundaries (wellbore geometry, lateral mass and heat flow). The governing equations are then solved in small, finite steps along the pipe. Whenever a feedzone is encountered, the mass and energy of inflow (or outflow) are known and mass and energy balances performed, allowing for continuation of the calculations. The governing steady-state differential equations for mass, momentum and energy flux in a vertical well have already been given in chapter 5, I will bring them again here in simple form:

Appendix E HOLA 3.1 - Multiphase Flow Simulator 216 __________________________________________________________________________

dm =0 dz dp dp dp dp = + + dz dz f dz elev dz acc dEt = Q dz

(E.1)

(E.2)

(E.3)

where m is the total mass flow, p is pressure, Et is the total energy flux in the well and z is the depth coordinate. Q denotes the ambient heat loss over a unit distance. The plus and minus signs indicate down flow and up flow respectively. The pressure gradient is composed of three terms: wall friction, acceleration of fluid and change in gravitational load over dz. The above terms were described in detail in chapter 5 which deals with multiphase flow in wells. The governing equation of flow between the well and the reservoir is:
k rL L k rG G m feed = PI + G L (Pr p wf

(E.4)

where mfeed is the feedzone flowrate, PI is the productivity index of the feedzone, kr is the relative permeability of the phases (subscripts L for liquid and G for steam), is the dynamic viscosity, is density, pr is the reservoir pressure and pwf is the flowing pressure in the well. The relative permeabilities are here calculated by linear relationships (krG = S and krL = 1-S where S is the volumemetric steam saturation of the reservoir). A flow into the well is positive and flow from the well into the formation takes a negative sign (Bjornsson, 1987).

Appendix E HOLA 3.1 - Multiphase Flow Simulator 217 __________________________________________________________________________ E.3 The computational models of HOLA 3.1

The simulator HOLA offers six modes of calculating downhole conditions in geothermal wells. These are:

1. Outlet conditions known at the wellhead: The simulator calculates pressure, temperature and saturation profiles from given wellhead conditions and given flowrates and enthalpies at each feedzone except the bottom one.

2. Required wellhead pressure and multiple feedzones: The simulator finds the downhole conditions that fulfill a required wellhead pressure. Also given are the productivity indices, reservoir pressure and enthalpies at each feedzone. The feedzones must have a positive flowrate.

3. Required wellhead pressure and two feedzones: This mode is similar to mode 2, except that only two feedzones are allowed and each can either accept or discharge fluid.

4. Required wellhead flowrate and two feedzones: In this mode, the simulator finds downhole conditions that fulfill a required wellhead flowrate. Only positive flow is allowed from the feedzones.

Appendix E HOLA 3.1 - Multiphase Flow Simulator 218 __________________________________________________________________________

5. Required wellhead injection rate and two feedzones: The simulator iterates for the downhole conditions that provide the required wellhead injection rate. Only two feedzones are allowed and both must accept fluid. Thus has negative flowrates in the program.

6. Variations in wellhead pressure and enthalpy for a constant flowrate and given reservoir pressure history at two feedzones: This mode is similar to mode 4, except that now a history is specified for the reservoir pressure. Only two feedzones are allowed and both must discharge to the well.

E.4 Heat loss parameters

The simulator handles heat transfer between the well and the reservoir by formulation given by Bjornsson (1987). The heat loss parameters necessary are:

rock thermal conductivity rock densities rock heat capacity time passed from initial discharge

These parameters specify the thermal conductance to the ambient rocks. The parameters used in present work have been taken from Prats SPE paper (2001). His work treats about steam injected to the reservoir, but the heat transferred from steam to the reservoir goes on the same rules like for geothermal wells, and the typical rock properties presented in his work given in

Appendix E HOLA 3.1 - Multiphase Flow Simulator 219 __________________________________________________________________________

the appendix (7.1.) are assumed to be appropriate for these simulations. 2 W/m/C, 2800
kg/m3, are used for the rock thermal conductivity and density respectively. The heat loses in

this option may also be neglected when the thermal conductivity value is set to be zero in this option.

In addition a lateral temperature gradient between the well and the ambient may be established defining a reservoir temperature curve from the surface to the feedzone level. The simulator requires in this option the temperature value at given depth and then interpolates linearly between these data points in order to evaluate the formation temperature gradient.

E.5 Wellbore geometry

The wellbore geometry need to be specified. The program allows for two different tubing diameter segments. The inflow is interpreted to be at the bottom of the deeper section, thus in the cases where the last feedzone does not occur at the well bottom the apparent depth need to be set equal to the depth of the lowest feedzone. The well inner radius and wall roughness need to be specified. The well roughness in present work was assumed to be 0.024 mm, and was taken from the PipeSim Table E.1.

The downhole computations proceed along the well in finite-difference steps of z. After each step the flowing conditions of the well are calculated at the subsequent nodal points defined in the program. The nodal points divide well into depth segments where parameters of the well like diameter or feedzones may change. The distance between the nodal points are

Appendix E HOLA 3.1 - Multiphase Flow Simulator 220 __________________________________________________________________________

defined by the user. More than 25 m distance is suggested by the program authors due to computer speed, however current PC handle the calculations for 1m nodal distance with few seconds.

E.6 Feedzone properties

The feedzone properties that must be specified in the program depends on the computational models selected. In mode 1 only the number of the feedzones, flowrate and the enthalpy of the each feedzone except the bottom one must be specified. The properties of the bottom feedzone, by conservation principles, are simply the residuals of the wellhead energy and mass flow minus energy and mass flow of all the other feedzones. Using the other program modes additionally productivity index, enthalpy and pressure at the feedzone need to be entered A special care must be taken in specifying the feedzone enthalpies or one may end up with negative temperatures in the well.

In present work only one feedzone was assumed for calculation, and some problem has occurred due to the mode 4 requires two or more feedzones to be specified. The problem was solved by placing both feedzones with a short distance of 20 m and assuming the same conditions for both feedzones. Such system may be assumed to work as the one feedzone.

Appendix E HOLA 3.1 - Multiphase Flow Simulator 221 __________________________________________________________________________ E.7. Velocities of individual phases

The simulator HOLA 3.1 offers four different methods of calculating the average cross sectional velocities of steam and water in a flowing well. Four different methods of calculating the average cross sectional are available in the program. All of the methods have their limitations and it is difficult to know which of them is best for a certain well. Experience in using wellbore simulators and access to downhole data in the well under consideration will often provide an appropriate selection of the phase velocity method.

In present work the Orkiszewski correlation was used. The description and the recommendation for this choice are given in chapter 5, which deals with the multiphase flow. The other correlations available are Armand, Chisholm and Bjornsson (for the 9 5/8 wells) The Armand relations are semi empirical and based on limited experimental data, collected in small diameter pipes at low flowrates (Bjornsson, 1987). The Chisholm in his correlation proposed the relation for slip in the form

1 + 0.4
k = 0.4 + 0.6

1 x G L L x 1 x G 1 + 0.4 x

(E.5)

This formula was modified in program for low pressure and low mass saturation of steam. Existing downhole pressure profiles were used to estimate the slip ratio and a correction

Appendix E HOLA 3.1 - Multiphase Flow Simulator 222 __________________________________________________________________________

factor to the Chisholm formula defined. A care should be taken in the use of this relation since it still needs validation (Bjornsson, 1987).

E.8 Productivity Index estimation

The IPR curves were calculated for the wells examples under considerations in this work in order to asses the performance and fin the well operating point for two tubing diameter sizes. This requires productivity index to be known for particular well that is the mathematical means of expressing the ability of a reservoir to deliver the fluids to the wellbore. The performance relationships are presented in Chapter 3. and here only the technical approach used in order to determine the productivity index is described.

HOLA 3.1 simulator has not a detailed mode to calculate the productivity index, and the productivity index was estimated from number of repeated simulations made in mode 2. The wellhead parameters including temperature and flowing pressure and reservoir average pressure was known. The enthalpy in this mode may be defined only at the feedzone thus in the first step the enthalpy was adjusted from several simulation to arrive at the appropriate wellhead temperature. Then initial guess for the productivity index was entered and the obtained wellhead mass flowrate was compared with required. Increasing or decreasing the productivity index results in estimation of the productivity index. Such approach was necessary in present work, because there was no specific data about the reservoir parameters available.

The vertical lift performance (VLP) curves for geothermal wells, was also calculated from HOLA 3.1. The downhole flowing pressure pwf was calculated several times for the same

Appendix E HOLA 3.1 - Multiphase Flow Simulator 223 __________________________________________________________________________

wellhead pressure but different flowrates with the 2 kg/s step. The curves revealed to have the unstable shape probably due to a problem with the simulator, e.g. because of flow regimes and/or fluid properties. The curve fitting application in MATHLAB 6.5 program allowed to get rid of the curves fluctuations and draw the plots.

Appendix E HOLA 3.1 - Multiphase Flow Simulator 224 __________________________________________________________________________

Table E.1 Thermal Conductivities


Material Density (kg/m3) Anhydrite Carbon Steel Concrete Weight Coat Corrosion Coat (Bitumen) Corrosion Coat (Epoxy) Corrosion Coat (Polyurathane) Dolomite Gypsum Halite Ice Lignite Limstone Line pipe Mild Steel tubing Mud Neoprene Rubber Plastic coated pipe Plastic coated tubing Polyurathane Foam (dry) Polyurathane Foam (wet) PVC Foam (dry) 30 - 100 100 - 340 1500 900 7900 2000 - 3000 Thermal Conductivity Btu/hr/ft/F 0.75 28.9 0.81 - 1.15 0.19 0.17 0.12 1.0 0.75 2.8 1.27 2.0 0.54 27 26 0.75 - 1.5 0.17 20 20 0.011 - 0.023 0.023 - 0.034 0.023 - 0.025 46.7 45 1.3 - 2.6 0.3 34.6 34.6 0.02 - 0.04 0.4 - 0.6 0.040 - 0.044 2.2 50 1.4 - 2.0 0.33 0.30 0.20 Thermal Conductivity (W/m/K)

Appendix E HOLA 3.1 - Multiphase Flow Simulator 225 __________________________________________________________________________


Sandstone Shale Stainless Steel Stainless steel (13%) Stainless steel (15%) Syntactic Foam (dry) Syntactic foam (wet) Volcanics Wet Sand 1600 500 1.06 0.7 8.67 18 15 0.052 .017 1.6 1.04 - 1.44 1.8 - 2.5 15 31.14 26 0.09 0.3

Table 7.2 Roughness


Material Drawn tubing(brass, lead, glass, and the like) Commercial steel or wrought iron Asphalted cast iron Galvanized iron Cast iron Wood stave Concrete Riveted steel ft. 0.000005 0.00015 0.0004 0.0005 0.00085 0.0006-0.003 0.001-0.01 0.003-0.03 in 0.00006 0.0018 0.0048 0.006 0.010 0.0072-0.036 0.012-0.12 0.036-0.36

Appendix F
Simulations results in oil wells