Comparison between experimental data and … · Objectives: • To compare experimental data against LedaFlow® in terms of: – Flow patterns – Void fraction – Pressure gradient

$Page 1: Comparison between experimental data and … · Objectives: • To compare experimental data against LedaFlow® in terms of: – Flow patterns – Void fraction – Pressure gradient$
Comparison between experimental data and LedaFlow® for two-phase flow in pipelines

Students:

Fabio Lopez

Hugo Pineda

Supervisors:

Nicolas Ratkovich

Ricardo Carvalho

July 15th, 2013, Campinas - Brazil1

Objectives:

• To compare experimental data against LedaFlow® in terms of:

– Flow patterns

– Void fraction

– Pressure gradient

• To compare empirical correlations against experimental data andLedaFlow® in terms of:

– Void fraction

– Pressure gradient

• To generate reliable synthetic data in LedaFlow® to train neuralnetworks.

2

Software tools used: LedaFlow®

1. Base case with the interface

3

2. Generation of .qs file and parametrization throught scripting

4

3. Several cases simulation with the command line

4. Export results of interest variables into .csv files

MATLAB®

– A program was developed to determine void fraction (67empirical correlations) and pressure drop (6).

– The program can group the data according to any of theparameters used.

– To select the correlation that best fits the data according toRMS error criteria.

– Calculation of the void fraction for operating conditions andfluids in the database by Pereyra 2012 and Choi 2012.

5

Data base:

• “An Efficient Drift-Flux Closure Relationship to Estimate LiquidHoldups of Gas-Liquid Two-Phase Flow in Pipes ”

– Choi et al. 2012

• 1136 data points used to compare experimental data againstLedaFlow® results in terms of void fraction and pressure drop.

6

Data base:

• “A methodology and database to quantify the confidence level ofmethods for gas–liquid two-phase flow pattern prediction”

– Pereyra et al. 2012

• 8850 data points used to generate synthetic void fraction andpressure drop in LedaFlow®

.Flow Pattern Flow Pattern # of points %

[Number] [Name]

1 Dispersed Bubble 816 9.2

2 Stratified Smooth 582 6.6

3 Stratified Wavy 1090 12.3

4 Annular 1631 18.4

5 Intermitent 4574 51.7

6 Bubble Flow 157 1.87

Manabe’s 2001 experiment summary (Oil-Natural Gas)

• Part of Pereyra et al. 2012 database: Initial comparison based on these data.

Flow Pattern Flow Pattern # of points %

[Number] [Name]

1 Dispersed Bubble 0 0

2 Stratified Smooth 0 0

3 Stratified Wavy 6 2.4

4 Annular 89 36

5 Intermitent 151 61.1

7 Bubble Flow 1 0.4

8

Manabe’s experiment summary (Oil-Natural Gas)

9

Flow pattern comparison (0º,460 psi)

10% Patterns predicted accurately 80.9

Experimental Ledaflow

Flow pattern comparison (90º,460 psi)

11% Patterns predicted accurately 92.7


Flow pattern comparison (90º, 211 psi)

12

% Patterns predicted accurately 88.1


13




14




Void fraction comparisonManabe (2001)Oil-Natural gas

15

LedaFlow® - Empirical correlations (0º, 211 psi)

16

LedaFlow® - Empirical correlations(0º, 460 psi)

17


18


19


20


21

Void fraction comparisonGokcal (2005, 2008)

Oil-AirHorizontal pipe

22

Experimental - LedaFlow® (0º)

23

Experimental - Empirical correlations (0º)

24

LedaFlow® - Empirical correlations (0º)

25

Void fraction comparisonFelizola (1992), Roumazeilles (1994)

Kerosene-AirInclined pipe

26

Experimental - LedaFlow® (-30º, -90º, 0.9 – 2 bar)

27

Experimental - Empirical correlations (-30º, -90º, 0.9 -2 bar)

28

LedaFlow® - Empirical correlations (-30º-90º, 0.9 - 2 bar)

29

Void fraction comparisonCarvalho (2013)

Oil-Air

30

31


32

Experimental - Empirical correlations (90º)

33

LedaFlow® - Empirical correlations (90º)

34

Reference

RMS

Ledaflow

-Exp

RMS

Ledaflow-

Corr

RMS

Exp-

Corr

Remarks

Manabe 0º, 460 psi - 7.7 -

Void fraction data

not available

Manabe 0º, 211 psi - 5.9 -

Manabe 1º, 465 psi - 6.0 -

Manabe 1º, 211 psi - 5.3 -

Manabe 90º, 465 psi - 6.4 -

Manabe 90º, 211 psi - 6.6 -

Gokcal 55.4 13.7 24.6 Incomplete data

Roumazeilles & Felizola 12.2 8.2 12.6

Carvalho 6.5 12.4 10.5

Void fraction results summary

35

Pressure drop comparisonGokcal (2005, 2008)

Oil-Air

36


37

Pressure drop comparisonFelizola(1992), Roumazeilles(1994)

Kerosene-Air

38

Experimental - LedaFlow® (-30º, -90º,0.9 – 2 bar)

Future Developments

39

Source Flow Pattern Void Fraction Pressure Drop

Pereyra X O O

Inoue’s X X O

Carvalho X X O

Choi O X X

Others O X X

O: Data to complete for neural network training

• To train neural networks for void fraction,

pressure drop and flow pattern determination.

Data base:

• Eduardo Inoue

.

40

Pattern # of data points %

Slug Flow 170 37.3

Vertical Annular 78 17.1

Intermitent 75 16.4

Horizontal Annular 57 12.5

Stratified 51 11.2

Vertical Dispersed Bubbles 25 5.5

Horizontal Dispersed Bubbles 0 0.0

Conclusions

41

• Ledaflow results are in good agreement with experimentalresults.

• Synthetic data generation with Ledaflow is a promising tool totrain neural networks provided the physical properties andoperating conditions are available.

• Ledaflow simulations can be adapted to match the requirementspresented by the neural network development.

Bibliography

• Choi, J; Pereyra, E; Sarica, C; Park, C; Kang, J. An Efficient Drift-Flux

Closure Relationship to Estimate Liquid Holdups of Gas-Liquid Two-

Phase Flow in Pipes. Energies. 2012, 5, 5294-5306.

• Woldesemayat, M; Ghajar, A. Comparison of Void Fraction Correlations

for Different Flow Patterns in Horizontal and Upward Inclined Pipes. Int.

J. Multiphase Flow. 2007, 6, 347-370.

• Pereyra, E; Torres, C; Mohan, R; Gomez, L; Kouba, G; Shoham, O. A

Methodology and Database to Quantify the Confidence Level of

Methods for Gas-Liquid Two-Phase Flow Pattern Prediction. Chem.

Eng. Research and Design. 2012, 90, 507-513.

• Godbole, P; Tang, C; Ghajar, A. Comparison of void fraction correlations

for different flow patterns in upward vertical two phase flow. Heat transfer Eng. 2011, 32, 843-860.

42

Comparison between experimental data and … · Objectives: • To compare experimental data against LedaFlow® in terms of: – Flow patterns – Void fraction – Pressure gradient

Documents