MULTIPHASE

CORRELATION
TECHNICAL REPORT

Student Name: Garion Charles

Student Number: 816006491
Course/Course Code: Production Engineering/ PENG 6015
Date Submitted: 23/02/2021
Introduction

Multiphase flow is common in many fields, such as the nuclear, chemical and petroleum
industries. Within the petroleum industry multiphase flow commonly occurs throughout the
time of production and transportation of oil and gas. Multiphase flow also occurs in oil field
gathering systems, the well bore and flow lines which can be horizontal, vertical, and inclined
pipes. Since multiphase mixtures must be transported from the well to facilities for separation
and especially for offshore wells the lines use in transportation of these fluids can be of
significant length, the prediction of the pressure gradient is of great importance for design
decisions. Hence utilization of a proper multiphase correlation based upon the situation at hand
is of great importance.

Multiphase correlations are utilized mainly for the determination of the liquid holdup
which also allows for the determination of the mixture’s density and is important in the
determination of the pressure gradient. There are three types of Multiphase correlation models
Homogeneous Equilibrium Models (HEM), Separated Flow Models and Multiple Fluid
Models. Additionally, in the case of petroleum engineering applications the main
hydrodynamic features are: liquid holdup, flow pattern and the pressure drop.

Furthermore, the Flow patterns can be divided into horizontal and vertical in which
flow pattern maps exist which allows for the determination of the flow regime for multiphase
fluids based upon the individual properties of the liquid and gas phases, for horizontal flow the
most used flow maps are Baker (1954) and Taitel & Dukler (1976) and for vertical flow the
most used flow map is Hewitt and Roberts (1969).

1
Flow Correlations

Poettmann & Carpenter – 1952

Poettmann & Carpenter proposed the first attempt at a multiphase flow correlation. The
correlation was created using a database of 49 wells which consisted of 34 flowing and 15 gas
lifted. The correlation was created for relatively low flowrates (300 – 800) bbl/day which had
GLR ranging (100 -800) scf/bbl using 2 3/8 and 2 7/8-inch tubing.(Poettman and Carpenter
1952)

Advantages:

• Easy to solve for pressure gradient and liquid holdup.

Disadvantages:

• Application is limited since only low flowrates data was used.

• Does not account for water cut.
• Does not take flow regime into consideration.

Baxandall & Thomas -1961

As a result of Poettmann & Carpenter using relatively low flowrate data, it resulted in their
correlation being inapplicable to high flowrate data. To resolve this issue Baxandall & Thomas
utilizes the same “energy-loss’’ method from Poettmann & Carpenter but applies it to a series
of experiments up to a max rate of 5000 bbl/day, these experiments were conducted at Cia.
Shell de Venezuela’s La Paz field and was done using tubing strings which were fitted with
electronic surface-recording pressure elements.

The experiments result in a correlation between the energy-loss factor and mass flowrate. As a
result of their approach the correlation is typically can be used for a wide range of different
conduit sizes and types of crudes at high flow rates, typically applicable for flow rates higher
than 900 bbl/day with a tubing OD size of 2 7/8 inches. At lower flow rates the energy loss
factor is not a constant for any mass flowrate but instead changes with the amount of gas in
place and the mixtures velocity. Therefore, a practical compromise for the lower flowrates is

2
the inclusion of a modified curve that was made by Poettmann and Carpenter.(Baxendell and
Thomas 1961)

Duns & Ros – 1963

A method which was created because of an experiment in which pressure gradients and liquid
holdup were measured, approximately 4000 two-phase flow tests were done using a 185 ft
vertical flow loop, with diameters of the pipes ranging from 5.6 to 1.26 inches were used.
Additionally, it also had two annulus arrangement.

A lot of the test was conducted at conditions close to atmospheric, where air was used as the
gas phase and either water or liquid hydrocarbons was used as the liquid phase. Furthermore,
measuring of the liquid holdup was done by use of a radioactive-tracer method and a clear
section was installed for seeing of the various flow patterns. Three flow patterns were observed
for which correlations were made for slip velocity and friction factor, which then can then be
used for the calculation of liquid holdup.(Duns Jr. and Ros 1963)

Advantages:

• One of the best correlations for mist flow

• Applicable across a wide range of conditions

Hagedorn & Brown (1964)

The method was based data gathered from a 1500 ft vertical well experiment where the gas
phase was air while four liquids were utilized: crude oil, which had viscosities of 110, 30, and
10 cp and water. Tube sizes of 1.5, 1.25, 1 inch were also utilized. Additionally, for Heavy
Oil’s (13-25 o API) it over predicts the pressure loss and it also overpredicts for GLR’s greater
then 5000 and for lighter Oil’s (40 – 56 o API) it under predicts the pressure loss.(Hagedorn
and Brown 1965)

Advantages:

• Data used was extensive and covered a wide range of properties which extremely
beneficial for empirical correlations.

3
Disadvantages:

• Liquid holdup is not accurate but rather it is referred to as a pseudo liquid hold up and
requires a correction factor to obtain the true liquid hold up.

Orkiszewski (1967)

Orkiszewski methods was develop by using the best correlations for bubble and mist flow at
the time and utilizing a new correlation for slug flow which Orkiszewski proposed using the
Hagedorn and Brown data. Griffith and Walls method was selected for bubble flow and Duns
and Ros was selected for mist flow. The correlation is generally applicable for tubing sizes of
(1 – 2 in) further increase causes an over prediction in the pressure drop, for low oil gravities
(13 – 30 o API) the correlation gives a over prediction however as the API gravity increases
the prediction of the pressure loss is improved.(Orkiszewski 1967)

Advantages:

• Makes use of the most accurate correlations for different regimes therefore able to
apply accurate results for across large range of flow patterns.

Disadvantages:

• The method can lead to convergence issues in computing algorithms because equation
pairs do not actually meet at a particular mixture velocity, this therefore leads to
pressure discontinuities and requires additional work to eliminate this problem.

Beggs & Brill (1973)

The correlation was develop using a transparent acrylic 90 ft long pipe, the fluids that were
used in the experiment was air and water and for each pipe size the liquid and gas flow rates
were varied to allow for all flow patterns to be produced. The parameters that were studied for
the development of the correlations were as follow: Gas flowrate (0 - 300 Mscf/day), Liquid
Flowrate (0 - 30 gal/min), Average System Pressure (35 – 95 psia), Liquid holdup (0 – 0.87),
Pipe diameter (1 and 1.5 in) inclination angle ( -90o - +90o) and Pressure gradient (0 to 0.8
psi/ft)(Beggs and Brill 1973)

4
Advantages:

• Can be used for any inclination.

Disadvantages:

• Over predicts pressure drop and therefore needs a correction factor to be applied.

Ansari. Et al. (1990)

A mechanistic model composed of a flow-pattern prediction model and a set of independent

models used for the prediction of such flow characteristics as pressure drop and holdup in
annular, slug and bubble flow. The performance of the model was compared with six
commonly used empirical correlations and once compared was found to perform best when
compared to the field data consisting of data from 1712 wells.(Ansari et al. 1994)

Advantages:

• Typically performs better than most empirical correlations that are commonly used

Disadvantages:

• Can only be used for upward fluid flow.

Gray 1974

A method developed by Gray for the determination of the pressure gradient in a vertical gas
well which also produces condensate fluids or water. The empirical correlation was developed
with the use of 108 well-test-data sets, of which 88 were producing free liquids.

However, it should be noted that caution should be taken if the fluid velocity is over 50ft/s, if
condensate/liquid loadings are above 50 bbl/MMscf, if the nominal diameter is above 3.5 in
and if the water/liquid loading are above 5 bbl/MMscf.(Chu 1994)

Advantages:

• More Accurate for condensate gas wells then dry gas correlations

Disadvantages:

• Dependability is only known based upon the test range from which the correlation was
created from.

5
Petroleum Experts (1,2,3) (2010)

The Petroleum Expert 1 correlation utilizes the best feature of other correlations, for flow map
it utilizes Gould et al, while for slug flow it uses Hagedorn Brown correlation and Duns and
Ros for mist flow. Petroleum expert 2 improves upon this correlation with improved VLP
calculations of lower flow rates well stability, which allows for more accurate predictions of
minimum load – up rates. Finally, Petroleum Experts 3 has all the features of 2 but now includes
original data for foamy, viscous, and volatile oils.

Ashena et al. (2010)

A model that was created using Back Propagation neural network models with different
network structures and fed it data from two major Iranian oil fields. The trained model in terms
of performance was shown to perform better in determining the bottom-hole circulating
pressure than the mechanistic model that they had been using on both oil fields.(Ashena et al.
2010)

Advantages:

• Can be used for inclined annular two-phase flow.

• Provides better results compared to most empirical correlations.

Disadvantages:

• Data set is limited therefore small range of application.

Al-Shammari (2011)

The model was created by applying the Fuzzy Logic technique for the estimation of bottom-
hole pressure and was applied to create an Adaptive Neuron Fuzzy Inference System (ANFIS)
based on a data set of 596 wells.

The model was able to successfully predict the bottom-hole pressures for 199 well test sample
with an average absolute error of 4.9%.(Al-shammari and Aramco 2011)

Advantages:

• Covers most of the applicable range in the process industry.

6
Disadvantages:

• Use outside the data range has a negative impact on results.

Table 1: Table to show Summery of flow correlations.

Slip
Correlation Application Assumptions Flow Regime
Considered
Poettmann & Only useful for None
Homogenous
Carpenter low flowrates, No Considered
mixture
vertical flow
Baxandall & Useful for high None
Homogenous
Thomas flowrates, No Considered
mixture
vertical flow
Duns & Ros Useful for a Bubble, Slug,
wide range of Mist
flow conditions Non-
especially Homogenous Yes
useful for mist mixture
flow, vertical
flow
Hagedorn & Useful for a Non- None
Brown wide range of Homogenous Considered
Yes
flow conditions, mixture
vertical flow
Orkiszewski Vertical flow Non- Bubble, Slug,
Homogenous Yes Mist
mixture
Beggs & Brill Can be used for Stratified,
Non-
any inclination Wavy, Annular,
Homogenous Yes
Plug, Slug,
mixture
Bubble, Mist

7
 Ansari. Et al. Vertical flow Non- Bubble, Slug,
Homogenous Yes Churn, Annular
mixture
Gray Especially None
useful for Non- Considered
condensate gas Homogenous Yes
wells, vertical mixture
flow
Petroleum Can be applied Bubble, Slug,
Non-
Experts (1,2,3) to foamy, Mist
Homogenous Yes
viscous, and
mixture
volatile oils

References

Al-shammari, Ahmed, and Saudi Aramco. 2011. “Accurate Prediction of Pressure Drop in
Two-Phase Vertical Flow Systems Using Artificial Intelligence.” SPE/DGS Saudi
Arabia Section Technical Symposium and Exhibition. Society of Petroleum Engineers,
2011.

Ansari, A. M., N. D. Sylvester, C. Sarica, O. Shoham, and J. P. Brill. 1994. “Comprehensive

Mechanistic Model for Upward Two-Phase Flow in Wellbores.” SPE Production and
Facilities 9 (2): 143–52. https://doi.org/10.2118/20630-pa.

Ashena, Rahman, J. Moghadasi, A. Ghalambor, M. Bataee, Rahim Ashena, and A. Feghhi.

2010. “Neural Networks in BHCP Prediction Performed Much Better Than Mechanistic
Models.” Society of Petroleum Engineers - International Oil and Gas Conference and
Exhibition in China 2010, IOGCEC 1: 187–93. https://doi.org/10.2523/130095-ms.

Baxendell, P.B., and R. Thomas. 1961. “The Calculation of Pressure Gradients In High-Rate
Flowing Wells.” Journal of Petroleum Technology 13 (10): 1023–28.
https://doi.org/10.2118/2-pa.

Beggs, H. Dale, and James R. Brill. 1973. “Study of Two-Phase Flow in Inclined Pipes.”
JPT, Journal of Petroleum Technology 25: 607–17. https://doi.org/10.2118/4007-PA.

8
Chu, Lee. 1994. “Multiphase Flow Production Model Theory and User ’ S Manual.”

Duns Jr., H, and N C J Ros. 1963. “Vertical Flow of Gas and Liquid Mixtures in Wells,”
June.

Hagedorn, Alton R., and Kermit E. Brown. 1965. “Experimental Study of Pressure Gradients
Occurring During Continuous Two-Phase Flow in Small-Diameter Vertical Conduits.”
Journal of Petroleum Technology 17 (04): 475–84. https://doi.org/10.2118/940-pa.

Orkiszewski, J. 1967. “Predicting Two-Phase Pressure Drops in Vertical Pipe.” Journal of

Petroleum Technology 19 (06): 829–38. https://doi.org/10.2118/1546-pa.

Poettman, Fred H, and Paul G Carpenter. 1952. “The Multiphase Flow of Gas, Oil, and Water
Through Vertical Flow Strings with Application to the Design of Gas-Lift Installations.”