Multiphase CFD Applied to Steam Condensation Phenomena ...mdx2.plm.automation.siemens.com/sites/default/files/...Visualization by L. Araneo, POLIMI STAR Japanese Conference, Yokohama,

Multiphase CFD Applied to Steam Condensation Phenomena in the

Pressure Suppression Pool

Marco Pellegrini - IAEColin Josey, Emilio Baglietto - MIT

STAR Japanese ConferenceYokohama, Japan – June 2nd, 2015

N U P E C

BACKGROUND

-3 0 3 7 10 13 16 19 22 25 28 31

0.0

0.1

0.2

0.3

0.4

0.5

3/1112:00

3/1118:00

3/120:00

3/126:00

3/1212:00

3/1218:00

3/130:00

Time after scram [hour]

DW

Pre

ssur

e (M

Pa[

abs]

)

Time [date]

UNIT 3

UNIT 2

RCIC system DW Pressureearthquake

STAR Japanese Conference, Yokohama, Japan6/9/2015

2

RCIC MAIN DIFFERENCES

0.283 m

1.275 m

UNIT 2VERTICAL JET

UNIT 3HORIZONTAL JETS

steam flow

2.577 m

0.033 m

0.680 m

Sparger detail

steam flow

• 1F2 RCIC suspected to have worked in two-phase flow

• 1F2 torus suspected to have been flooded by the tsunami

• 1F3 RCIC worked at the same time with cycling SRVs

Bottom closed


3

EXPERIMENTAL ACTIVITIES AND COLLABORATIONS

TITech FacilityG. Gregu, M. Takahashi

pool scrubber

SIET Facility

3 m

0.5

mSTAR Japanese Conference, Yokohama, Japan6/9/2015

4

SPARGER STRATEGYVent pipe - RCIC 1F2 RCIC 1F3

Petrovich, Int, J. Heat and Mass Tr, 2007

Steam mass flux [kg/m2-s]Diameter [m]

Sub

cool

ing

[K]

T-quencher

0.02 m

0.2 m

D

D

0.1 m

D


5

Petrovich, Int, J. Heat and Mass Tr, 2007.

Steam mass flux [kg/m2-s]Diameter [m]

Sub

cool

ing

[K]

STAR Japanese Conference, Yokohama, Japan

CONDENSATION REGIME MAP

CHUGGING BUBBLING JETTING

Experiment at SIET labs, ItalyVisualization by Prof. L. Araneo, POLIMI

6/9/2015

6

TITech EXPERIMENT: CHUGGING PHENOMENOLOGY1000 fps

time [ms]

Mass flow rate: 3.9 g/sTpool: 23.7 °C100

0

-60

pres

sure

[kP

a]pr

essu

re [k

Pa]

-60

100

0

0.2 0.40 0.6 1 1.2 1.4 1.60.8

• 65ms: bubble formation at outlet• 170ms: bubble collapse• 258ms: condensation inside the pipe• 599ms: condensation inside the pipe• 997ms: condensation inside the pipe• 1550ms: bubble formation at outlet• 1679ms: bubble collapse

pressure signal – G. Gregu, POLIMI/TITech


7

UNIT 3 RCIC SPARGER

Visualization by L. Araneo, POLIMI


Steam flow

Tpool = 30 °C

Steam flow

6/9/2015

8

TWO-FLUID MODEL: MOMENTUM EQUATION

∙

∙

Standard Drag

12 4

Phase momentum equation

Interphase momentum transfer drag forcevirtual mass forcelift forceturbulent dispersion force

Schiller-NaumannTomiyamaBozzano-Dente

correction factor

,

Two-fluid model approach


9

TWO FLUID MODEL: ENERGY EQUATION

∙ ∙

∙ , ∙ ∙ ∙

Phase energy equation

∆

∆ ∆

Source term in the energy equation

interaction length scale area density

Differently from two-fluid for boiling applications, the interaction length scale is generally differently defined from the area density in condensation applications.

lt

General bubble surface


10


TWO FLUID MODEL: ENERGY EQUATION

Model Formulation / Reference

Large eddy / Fortesque and Pearson (1967)

Small eddy / Banerjee et al. (1968)

Surface divergence / / Banerjee (1990)

SD no shear0.3 2.83 /

2.14 / / Banerjee (1990)

Gas flow

T

Main historical heat transfer models

ltvt

∆ ∆

Surface renewal period

⁄⁄

⁄ ⁄

6/9/2015

11


TWO FLUID MODEL: INTERFACIAL AREA DENSITY

Sauter mean diameter

Magnitude of Volume Fraction Gradient

Example of volume fraction

1.000.750.500.250.00

volume fraction

3L

LMore proper in case of boiling applications

EULERIAN-EULERIAN TWO-FLUID APPROACH

6/9/2015

12

COMPRESSIBILITY EFFECT

1.000.750.500.250.00

volume fraction


13

WATER Constant or temperature dependent density

STEAM

incompressible compressible

P

ρPressure limit

TITech EXPERIMENT: MESH SENSITIVITY

0.5 m

0.5 m

200,000 cells 800,000 cells

COARSE FINE

TEST CONDITIONS

Pipe diameter = 2.7 cmMass flow rate = 5.58 g/sMass flux = 9.75 kg/m2-sPool bulk T = 19 ºCSteam T = 100 ºC (saturated)


14

0

5

10

15

20

25

30

0 5 10 15

Con

dens

atio

n m

ass

trans

fer [

g/s]

Time [ms]

CHUGGING AT LARGE SUBCOOLING AND MASS FLUX

1.000.750.500.250.00

volume fraction 1.000.750.500.250.00

volume fraction

COARSE FINE

fine mesh

coarse mesh

Inlet mass flow rate5.58 g/s

Tpool = 19 ºC Tpool = 19 ºC


15

NON ENCAPSULATING BUBBLE6.8 ms 6.9 ms 7.0 ms 7.2 ms6.0 ms

0510152025303540

05

1015202530

0 2.5 5 7.5 10 12.5

Mas

s tra

nsfe

r [g/

s]

Time [ms]

Tota

l are

a [c

m2 ]


16


CHUGGING: LOW SUBCOOLING AND MASS FLUX

4

2

0

-2Pre

ssur

e [k

Pa]

0 100 200 300 400

• Pressure starts decreasing below zero due to condensation greater than inlet mass flow rate of steam

• An implosion time is reached at the minimum pressure value

• Afterwards the interface flows in the pipe and the steams gets compressed

interface within the pipe

Marks and Andeed, 1979

implosion

SIET facility

Pipe diameter = 0.2 mMass flow rate = 0.1 kg/sMass flux = 3.18 kg/m2-sPool bulk T = 65 ºCSteam T = 100 ºC (saturated)

6/9/2015

17


CHUGGING: PHENOMENA INTERPRETATION

4

2

0

-2Pre

ssur

e [k

Pa]

0 100 200 300 400

• Pressure starts decreasing below zero due to condensation greater than inlet mass flow rate of steam

• An implosion time is reached at the minimum pressure value

• Afterwards the interface flows in the pipe and the steams compressed

interface within the pipe

Marks and Andeed, 1979

implosion

20 ms 40 ms

IMPLOSIONLOW PRESSUREINTERFACE

MOVING UPWARD

6/9/2015

18


RAYLEIGH-TAYLOR INSTABILITY

g A

LIGHT FLUID

HEAVY FLUID LIGHT FLUID

HEAVY FLUID

Gravitation field Accelerating flow field

separator

6/9/2015

19


RAYLEIGH-TAYLOR INSTABILITY

LIGHT FLUID

HEAVY FLUID

A

steam

water

Gravitation field Accelerating flow field

g Psteam

Pwater

6/9/2015

20


APPROACH FOR RTI IMPLEMENTATION

d ndt ( , , , , )n f g k A

Amplitude growth description

wave number

2 4 2n Agk k k

2

w s

kn Ag k

Duff et al. Physics of Fluid, 1962

Livescu, Physics of Fluid, 2004

Classic instability theory

n Agk

acceleration

wave number

Atwood number

viscosity

surface tension

Livescu

Duff

Classical theory

6/9/2015

21

C. Josey, E. Baglietto, 2013


IMPLEMENTATION OF THE RTI IN STAR-CCM+

max 3

w sAgk

w

w

Pg

2

1i ska

n tt t te

Wave number termAcceleration term

ν ν

Duff and Livescu combined model for RTI

Final terms for area growth

6/9/2015

22

POOLEX: LOW SUBCOOLING AND MASS FLUXPOOLEX facility detail

Experiment conditions at the POOLEX

Pipe diameter = 0.2 mT pool = 62 °CSteam Mass Flux = 8 kg/m2s

velocity inlet

pressure outlet

adiabaticwalls

Mesh elements: 405,067

steam inlet

Domain Discretization


23

RTI MODEL RESULTSRayleigh-Taylor Instability ModelMinimum Area Model


24

Tpool = 62 ºCTpool = 62 ºC

With RTI model the steam interface re-enters the pipe and a new cycle is started once decrease of interfacial area, turbulence, subcooling creates the proper conditions.

MASS TRANSFER COMPARISON

00.5

11.5

22.5

33.5

44.5

5

0 0.5 1 1.5 2 2.5

mas

s tra

nsfe

r [kg

/s]

time [s]

Rayleigh-Taylor Instability model

Minimum area model


25

Bubble at the pipe mouth

Inlet mass flow rate0.3 kg/s

Without a model that takes into account the growth of the area the bubble remains oscillating at the outletRayleigh-Taylor Instability adds an exponentially increasing surface area that reproduces the bubble collapse.

BUBBLE IMPLSOSION COMPARISON

120 ms 190 ms 210 ms 230 ms

EXP

no RTImodel

RTI model

120 ms 190 ms 210 ms 220 ms

Tanskanen, Ph.D. Thesis 2012


26

TEMPERATURE EVOLUTIONRayleigh-Taylor Instability ModelMinimum Area Model


27

Wrong prediction of interface movement will tend to overestimate the creation of stratification in the pool.

SIET FACILITY - STRATIFICATION

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Tem

pera

ture

[°C

]

Time [min]

TP1TP2TP3TP4TP5TP6TP7TP8TP9TP10TP11TP12

Chugging stops


28

Stra

tific

atio

n =

55 ºC

once chugging is occurring the temperature will be uniform “almost” independently on the location of the pipeonce chugging stops stratification starts proportional to the distance from the surface

• There is large potential to employ CFD in severe accident applications (DCC is one of them)

• Generally a SA code employs one large node to model the whole S/C

• CFD can be used as informative tool for SA code but…

• … it can be used also for industrial applications:– Design operation– Severe Accident Management Guidelines

REMARKS


29

Multiphase CFD Applied to Steam Condensation Phenomena ...mdx2.plm.automation.siemens.com/sites/default/files/...Visualization by L. Araneo, POLIMI STAR Japanese Conference, Yokohama,

Documents