0% found this document useful (0 votes)
25 views16 pages

Choking Flow Model _yoon revankar IJMHT 2006

This document presents a mechanistic model for choking flow in two-phase two-component systems, focusing on mechanical non-equilibrium characterized by a slip ratio. The model evaluates choking mass flow rates based on pressure, quality, and slip ratio, and compares predictions with experimental data, showing reasonable accuracy in high-quality regions. The study highlights the importance of understanding choking flow for safety assessments in nuclear power plants, particularly during loss of coolant accidents.

Uploaded by

raknaver8461
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
25 views16 pages

Choking Flow Model _yoon revankar IJMHT 2006

This document presents a mechanistic model for choking flow in two-phase two-component systems, focusing on mechanical non-equilibrium characterized by a slip ratio. The model evaluates choking mass flow rates based on pressure, quality, and slip ratio, and compares predictions with experimental data, showing reasonable accuracy in high-quality regions. The study highlights the importance of understanding choking flow for safety assessments in nuclear power plants, particularly during loss of coolant accidents.

Uploaded by

raknaver8461
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 16

Nuclear Engineering and Design 236 (2006) 1886–1901

Choking flow modeling with mechanical non-equilibrium


for two-phase two-component flow
H.J. Yoon ∗ , M. Ishii, S.T. Revankar
School of Nuclear Engineering, Purdue University, West Lafayette, IN 47906, United States
Received 6 May 2005; received in revised form 6 February 2006; accepted 8 February 2006

Abstract
A mechanistic model which considers the mechanical non-equilibrium is described for two-phase choking flow. The choking mass flux is
obtained from the momentum equation with the definition of choking. The key parameter for the mechanical non-equilibrium is a slip ratio. The
dependent parameters for the slip ratio are identified. In this research, the slip ratio which is defined in the drift flux model is used to identify
the impact parameters on the slip ratio. Because the slip ratio in the drift flux model is related to the distribution parameter and drift velocity, the
adequate correlations depending on the flow regime are introduced in this study. In this mechanistic modeling approach, the choking mass flow rate
is expressed by the function of pressure, quality and slip ratio. The developed model is evaluated by comparing with the air–water experimental
data to eliminate the thermal effect. The comparison of predicted choking model for mechanical non-equilibrium with other experimental data in
high quality region (up to 80%) is quite reasonable with a small error.
© 2006 Elsevier B.V. All rights reserved.

1. Introduction type of models is called the homogeneous equilibrium model. In


this model, the two phases are at equilibrium with equal veloci-
The choking flow is an important phenomenon that occurs ties and temperatures. The thermodynamic properties which are
in wide range of industrial system. This phenomenon is very used in this model are attainable from the steam table, and the
important in the safety assessment of a water cooled nuclear equation of state can be used. However, this assumption is valid
power plant. In the event of loss of coolant accident (LOCA), only for some ideal conditions such as for a long pipe where
the choking flow determines the water inventory of the reac- there is sufficient time for equilibrium to be achieved and when
tor vessel, and the integrity of core eventually depends upon the flow pattern gives sufficient interphase forces to suppress
the choking flow (Schrock and Amos, 1984). Therefore, ana- significant relative motion.
lytical description and prediction of choking flow rate plays an
important role in the design of the engineered safeguards and 2. Review of choking flow models
operation of the nuclear power plant. The serious study of two-
phase flow was started as early as Sauvage (1892). Rateau (1905) It is also important to understand the existing choking flow
showed the existence of choking flow in the boiling water flow models for choking model development. In this section, the
through a nozzle, where the choking condition was obtained choking flow models are classified as homogeneous equilibrium,
by reducing the back pressure. Rateau also presented quality homogeneous non-equilibrium, non-homogeneous equilibrium,
calculation method for the discharged saturated water through and non-homogeneous non-equilibrium model.
nozzle based on isentropic expansion. In the early mechanical
and thermodynamic equilibrium between gas and liquid phases 2.1. Homogeneous equilibrium model
were commonly assumed in the early choking flow models. This
According to Starkman’s homogeneous equilibrium model
(Starkman et al., 1964), it is based on the assumptions of no
∗ Corresponding author. Tel.: +1 765 496 3902; fax: +1 765 494 9570. slip, thermal equilibrium between phases, isentropic expansion
E-mail address: [email protected] (H.J. Yoon). and equation of state in the steam table. Thermal equilibrium

0029-5493/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.nucengdes.2006.02.007
H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901 1887

Nomenclature t total
up upstream
A area w wall
c specific heat x-s cross section
cp specific heat
Co distribution parameter
D diameter means that the pressure and temperature of liquid and vapor are
Dsm Sauter mean diameter equal and linked by the saturation curve.
f friction factor The flow rate predicted from these assumptions is
F function
{2[ho − (1 − xE )hfE − xE hgE ]}1/2
G mass flux (kg/m2 s) Gc = (1)
1−xE
h enthalpy ρf + 1
ρg
j superficial velocity
Conceptually the choking mass flux of this prediction can be
L length
reached by decreasing the downstream pressure until the flow
MF frictional pressure gradient in multi-bubble sys-
rate reaches the maximum value. Beyond this, any back pressure
tem
reduction will not change the discharge flow rate because of the
MF∞ Frictional pressure gradient in single-bubble sys-
pressure singularity at the choking flow point. As can be noticed
tem
in Eq. (1), the choking mass flux in the homogeneous equilibrium
ṁ mass flow rate
model depends not on the break geometry or pipe length but on
N experimental parameter
the upstream thermodynamic condition.
n polytropic exponent
A feature of the homogeneous equilibrium model is adher-
P pressure
ence of the fluid properties along the saturation condition. In
Re Reynolds number
reality, the liquid can be superheated in order to nucleate near
s entropy
the wall during the depressurization process. If the single-phase
S slip ratio
liquid velocity is high enough, the choking flow may occur at
T temperature
the outlet of the pipe with the nucleation of the first bubble. If
v specific volume
the coolant temperature is low and nucleation is suppressed by
V velocity
non-equilibrium phenomena, it might be required to consider
Vgj  weighted drift Velocity
the transition from single-phase to two-phase choking flow dur-
V̄gj mean drift velocity
ing the early stage of a blowdown in a LOCA analysis (Wallis,
x quality
1980). According to the review of Abdollahian et al. (1982),
z axial distance
the homogeneous equilibrium model showed good agreement
Greek letters with the Marviken critical flow data for subcooled stagnation
α void fraction condition and long pipe condition (L/D > 1.5). However, in
γ isentropic exponent general, the homogeneous equilibrium model is the simplest
Γ mass change rate approach and not highly accurate for the subcooled liquid stag-
η pressure ratio nation state (Schrock and Amos, 1984). Recently, Fthenakis et
ρ density al. (2003) showed the different opinion for the homogeneous
τ shear stress, relaxation time equilibrium model. The homogeneous equilibrium model can
ω correlating parameter overestimate the choking flow rate in case that the break is large,
and that makes the rapid depressurization. However, accord-
Superscripts ing to results of Darby (1998), the homogeneous equilibrium
* nomalized value model underestimated the choking flow rate in short pipe. In
that reference, the homogeneous equilibrium model is appli-
Subscripts cable for the long pipe because this model is based on the no
c choking slip velocity, whereas, there is no sufficient time to establish the
down downstream mechanical equilibrium in the short pipe. However, the homo-
E equilibrium geneous equilibrium model overestimates the mixture density
f liquid in case of large break, and this results in more liquid being
g vapor or gas phase discharged through the break. Therefore, the homogeneous
i interface equilibrium model overestimates the choking flow rate at the
m mixture break.
o stagnation As the other approach of the homogeneous equilibrium
s saturation model, there is Omega method. The analytical solution using
this method was first proposed by Epstein et al. (1983) for the
steam–water mixture, and generalized by Leung (1986) for any
1888 H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901

flashing two-phase mixture. They introduced the approximate good agreement using the “relaxation length” by comparing
equation of state which is given by with other data. This “relaxation time” or “relaxation length”
  corresponds to the super heat of 2–3 ◦ C.
v Ps
=ω −1 +1 (2) Henry and Fauske (1971) tried to account for the thermal
vfo P non-equilibrium effect by introducing an empirical parameter,
where ω yields which represented the deviation from the homogeneous equi-
 2  2 librium model. This model assumed the phase separation at the
xo vfgo cfo To Po vfgo vfgo entrance and neglected the friction loss. It is directly proportional
ω= + ≈ αo + ρo cfo To Po
vo vo hfgo hfgo to the difference between the exit qualities. In their method, the
(3) simplified choking mass flux expression is given as
⎡ ⎧
Based on these parameters, the expression for the normalized ⎨ (1 − x )N ds
x v
G2c = ⎣
o g o fE
mass flux from the momentum equation is + (vg − vfo )
nP ⎩ sgE − sfE dP
{−2[ω ln η1 + (1 − η1 )(ω − 1)]}1/2
G∗ =   (4)   ⎫⎤−1
ω η11 − 1 + 1 xo cpg ⎬
1
− 1
n γ
− ⎦ (6)
where η1 is the pressure ratio at the inlet. The choking equation P(sgo − sfo ) ⎭
t
based on Eq. (4) is given by (Leung and Grolme, 1986):
  where n is the polytropic exponent and N is an experimental
Po ρfo 1/2
Gc = ηc (5) parameter.
ωηs
If N is unity, the prediction of above equation converges close
where subscripts c and s means choking and saturation, respec- to that of the homogeneous equilibrium model, and if N is zero
tively. it gives the homogeneous frozen model where no phase change
The author also pointed out that the homogeneous equilib- takes place. Therefore, the quantity N describes the degree of the
rium model was physically unrealistic. However, the opinion of non-equilibrium phase change occurring at the throat between
author was that this model was valuable for understanding qual- two limiting conditions. However, Abdollahian et al. (1982)
itative features of two-phase choking, using in scaling law, and pointed out that the Henry–Fauske model in the subcooled flow
estimating quantitative prediction (Leung and Epstein, 1990). regime overpredicted the Marviken data.
Schrock et al. (1977) developed a two-step model. Initially
2.2. Homogeneous non-equilibrium model “frozen” flow (constant quality) is assumed until the pressure
drops by a certain amount below the saturation pressure. At that
Thermal non-equilibrium implies that there is a temperature point sudden equilibrium is achieved, and downstream from
difference between the phases. This is believed to be one of the this point, again the frozen flow occurs. It was stated by the
major causes of the discrepancy between model predictions and authors that the model did not represent the real physical behav-
experimental results, especially if subcooled fluid enters the test ior, however they claimed that there was no satisfactory means to
section or the test section is very short. According to the review predict the number and size of microbubble in the liquid trigger-
presented by Abdollahian et al. (1982) and Saha (1978), one ing the nucleation process. Schrock and Amos (1984) pointed
particular type of choking flow has escaped full understanding out that the choking model which considered the thermal non-
to date, even though a number of theoretical and experimental equilibrium should include the physics of the nucleation process
studies of choking flow have been reported. It occurs with a and the bubble growth.
subcooled upstream condition that often leads to significant non- Levy and Abdollahian (1982) developed the simplified homo-
equilibrium thermodynamic conditions at the point of choking geneous non-equilibrium flashing model. When there is a rapid
flow. At that point the liquid is superheated beyond the saturation pressure decrease along the flow direction, this model may be
temperature. This non-equilibrium effect is generated because applicable. This model assumed that water should be super-
the two-phase fluid depressurization speed may be faster than heated at a given local pressure before liquid flashed into steam
the thermal exchange speed between two phases (NEA, 1982). and that in downstream enough vapor would be generated to
This phenomenon occurs in the short L/D configuration. This reduce the water superheat. In addition, the flow was assumed
metastable condition has been difficult to predict and is sensitive to be homogeneous. Finally, an isentropic process was used to
to the specific geometric characteristics of the break. This is calculate the non-equilibrium quality and choking flow rate.
because the local nucleation depends on the liquid superheat This model was compared with the data of Reocreux (1974)
and the bubble nucleation affects the relative motion between and Zimmer et al. (1979). For those two data sets, this model’s
phases. prediction was reasonable. A good agreement was also obtained
Actually, thermal non-equilibrium is related to the bubble with the large scale Marvinken tests and many other small scale
nucleation. This was proven by several researchers using a delay experiments. However, the model may not properly account for
time for nucleation time of order of 1 ms (Khajehnajafi and the impact of depressurization rate upon non-equilibrium con-
Shinde, 1995; Schrock et al., 1977). Fauske (1985) showed a ditions. In addition, it underpredicts small scale tests at high
H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901 1889

pressures when the contraction zone is not followed by a con- superficial velocity in choking condition is much high, therefore,
stant cross-section length. it is reasonable that the slip ratio does not affected by the quality.
Finally, there is the frozen model as it is referred previ- Based on the non-homogeneous equilibrium assumption, Fauske
ously. This model is the limiting case of the homogeneous derived the choking relationship as follows:
 1/2
S
Gc = − dv
(7)
{(1 − x + Sx)x} dPg + {vg (1 + 2Sx − 2x) + vf (2xS − 2S − 2xS 2 + S 2 )} dP
dx
+ {S[1 + x(S − 2) − x2 (S − 1)]} dv
dP
f

However, the weak point of these models is that they are


non-equilibrium model and it is based on the assumption of
derived based on the ideal flow patterns, annular or separated
no heat and mass transfer between phases. By using the isen-
flow. Furthermore, they are based on the energy or momentum
tropic expansion for the vapor phase, the choking flow rate
maximization or minimization assumption which has no the-
can be calculated from the gas dynamic principle. Lenzing et
oretical bases. In addition the slip ratio is not independent of
al. (1998) used the homogeneous frozen flow model by intro-
the quality. Cruver and Moulton (1967) pointed out the problem
ducing the two-phase discharge coefficient for the non-flashing
in the assumptions of Fauske’s theorem. In the Fauske’s the-
two-component flow and Henry and Fauske model by using the
sis, the author assumed that the pressure gradient at the choking
two-phase discharge coefficient for the flashing one-component
point is a finite maximum. However, according to Cruver et
flow.
al.’s opinion, ∂vm /∂S is a minimum, not a maximum because
∂2 vm /∂S 2 = 2x(1 − x)vg /S 3 is always positive. Therefore, they
2.3. Non-homogeneous equilibrium model insist that the finite maximum for dP/dz does not exist.

The early model of the non-homogeneous equilibrium model 2.4. Non-homogeneous non-equilibrium model
derived by Moody (1965) is an extension of the homogeneous
equilibrium model, by allowing different vapor and liquid veloc- In the two-fluid model, it is possible to consider the mechan-
ities. A slip ratio, S, defined as the velocity ratio between the ical non-equilibrium and the thermal non-equilibrium simul-
vapor and liquid, is treated as a variable which is determined by taneously through proper constitutive relations for interfacial
the condition of maximum kinetic energy flux at the exit. This transfer rates.
category of models is called “slip flow model”. The slip between In this approach, a separate set of conservation equations
two phases allows the gas phase to be discharged with the higher is used for each phase as shown by Bouré (1974) and Ishii
velocity than liquid phase and this is more realistic approach (1977). Therefore, these equations have to contain terms describ-
than the homogeneous flow assumption (Fthenakis et al., 2003). ing mass, momentum and heat transfer between the phases.
In Moody model, the important assumptions that characterize These constitutive relations are not very accurate in the present
his model are as follows. Both phases are at the same static state-of-the-art, particularly under the choking flow conditions
pressure and in local equilibrium at entrance and the exit. The where very large convective acceleration and depressurization
two-phase flow pattern at exit is annular flow without entrain- take place. Thus simplification of the above equations and often
ment. The exit velocities of each phase are uniform. The exit quite arbitrary assumptions are introduced to overcome the lack
slip ratio is an independent variable and given by a correlation. of knowledge. The separate flow model is easy to visualize
Moody model gives the local slip ratio of S = Vg /Vf = (ρf /ρg )1/3 to the mechanical non-equilibrium and thermal non-equilibrium
get a maximum two-phase kinetic energy flow, whereas Fauske between phases, however, the actual geometry of the interface
(1963) obtains the slip ratio as S = Vg /Vf = (ρf /ρg )1/2 by mini- can be quite different from this flow due to wall nucleation and
mizing the momentum flow rate. The detailed description was flashing.
shown in ANL report based on his thesis (Fauske, 1962). Fauske The typical existing interfacial exchange models were
emphasized the importance of the mechanical non-equilibrium described by Trapp and Ransom (1982). They discussed the
in terms of the slip. According to his thesis, the slip is gener- calculation procedure and the numerical implementation of the
ated due to the density ratio between two phases and, therefore, choked flow criterion. Possible weakness of their model involves
light phase is easily accelerated by means of pressure differ- the modeling of non-equilibrium mass and heat transfer terms,
ence between the upstream and choking section. The author thermal resistance to heat flux, virtual mass coefficient, and the
also insisted that the friction between the phases and droplet use of derivatives of the equations of state in thermal equilib-
caused by the liquid entrainment reduces the relative velocity rium. This implies that the interfacial momentum and energy
between two phases. For the saturated or subcooled water flow, in highly non-equilibrium flow may not be accurately modeled
the accelerative force makes the pressure gradient and the vapor by the conventional models in the code, particularly due to the
is generated. This results in the existence of different veloci- significant nucleation at the wall which is difficult to predict.
ties between two phases. In this situation, the vapor is always Richter (1983) developed the separated flow model to calcu-
faster than the liquid. However, Fauske insisted that the slip late the choking flow rate for the steam–water mixture. Gener-
is independent of the quality using the results of Lottes et al. ally, the separate flow model requires several correlations for
(1959). According to their experimental results, the effect of slip the unknown parameters. However, it deals both mechanical
ratio decreased with pressure increase. According to Fauske, the and thermal non-equilibrium in one approach. In this research,
1890 H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901

the author also tried to find the flow regime dependency in Gas phase
the choking flow phenomena. This model was considered with ∂ ∂[αAx−s ] ∂Ax−s
the assumptions of steady state, one dimensional, and non- − [PαAx−s ] z − Pi z + Pw z
∂z ∂z ∂z
homogeneous flow. Additionally, it was assumed that heat trans-
fer between phases limited mass transfer. In this research, fric- ∂
−gρg αAx−s z sin θ + τi pi z = [ρg αAx−s Vg z]
tional pressure drop was not neglected for the long pipe case. ∂t
According to Richter’s results, the slip ratio in the bubbly ∂
flow regime was about unity and it was reasonable to assume + [ρg αAx−s Vg2 ] z − Ax−s ΓVi z (9)
∂z
this as many authors assumed. However, the author insisted
where τ is the shear stress. The subscripts w and i denote the
that if the evaporation in expanding bubbly flow was limited
wall and interface.
by heat transfer, the small velocity difference played an impor-
For steady state flow, additional assumptions are considered
tant role in the heat transfer mechanism. That implies that the
such that (i) the radial pressure gradient is assumed to be zero and
mechanical non-equilibrium affect the thermal non-equilibrium.
(ii) the interfacial pressure, Pi , and wall pressure, Pw , are equal
For the churn-turbulent flow regime, it was concluded that
to the static pressure, P. Accounting for the above assumptions,
the thermal non-equilibrium was negligible and the mechani-
Eqs. (8) and (9) are reduced respectively to
cal no-equilibrium was dominant in this flow regime. Although
Richter’s result was very interesting, the results depended on ∂P τ w pe τi P 1 ∂
−(1 − α) − + = [ρf (1 − α)Ax−s Vf2 ]
the chosen correlation when this model was solved, and it was ∂z Ax−s Ax−s Ax−s ∂z
questionable whether it might be applied to the low pressure. (10)
This model calculation was compared with the experimental
results of high pressure in this research by the author. Accord- ∂P τi P 1 ∂
−α − = [ρg αAx−s Vg2 ] (11)
ing to the review of Abdollahian et al. (1982), this model was the ∂z Ax−s Ax−s ∂z
most successful in predicting the entire region of the Marviken Hence, by combining Eqs. (10) and (11), the mixture momentum
data. equation is obtained as follows:

3. Modeling description ∂P τw pe 1 ∂
− − = [ρf (1 − α)Ax−s Vf2 + ρg αAx−s Vg2 ]
∂z Ax−s Ax−s ∂z
The main objective of research is to develop the mech- (12)
anistic choking model which considers the mechanical non- In the choking condition, the effect of the wall friction on
equilibrium. In the view of the liquid and gas velocity, it is the pressure drop is negligible compared to that attributed to the
homogeneous flow if both velocities are same. As the non- convective term. Hence, the pressure drop due to the wall friction
homogeneous flow, there are two types of non-homogeneous is neglected in developing the choking model. Therefore, after
flow, mechanical equilibrium flow and mechanical non- rearranging the convective term, the final momentum equation
equilibrium flow. Even though both velocities are different, if for the choking condition in a separated two-phase flow is given
the relative velocity (Vg − Vf ) is steady state, it is the mechanical by
equilibrium flow. Mechanical non-equilibrium flow is important
in the inertia dominant flow. In the nozzle geometry, the pres- ∂P 1 ∂ ∂
− = [ṁf Vf + ṁg Vg ] = G [(1 − x)Vf + xVg ]
sure drop is mainly caused by the acceleration of each phase. ∂z Ax−s ∂z ∂z
In addition, near the choking point, acceleration rates of each ∂
phase, DVf /Dt and DVg /Dt, are different. Therefore, the model- =G {[(1 − x) + xS]Vf } (13)
∂z
ing for mechanical non-equilibrium in the choking flow is very
important. where ṁ, G and S denote mass flow rate, mass flux and slip ratio,
In writing the momentum equation of the separated flow, sev- respectively. The slip ratio in Eq. (13) is defined by
eral assumptions are made in advance, namely: (i) the surface Vg
S= (14)
tension, Reynolds stress, and virtual mass effects are assumed Vf
to be small enough to be neglected and (ii) the pressure of each
and the mass flux given by
phase is assumed to be equal. The momentum equations for the
liquid and gas phase are as follows: G = [αρg Vg + (1 − α)ρf Vf ] = [αρg S + (1 − α)ρf ]Vf (15)
Liquid phase
In the adiabatic condition, the flow quality is not changed
∂ ∂[αAx−s ] ∂Ax−s even though the void fraction varies due to geometry changes.
− [P(1 − α)Ax−s ] z − Pi z + Pw z
∂z ∂z ∂z Therefore, it is more convenient to replace the void fraction
−gρf (1 − α)Ax−s z sin θ − τw pe z + τi pi z with the flow quality. The relationship can be obtained from the
continuity equation as
∂ ∂
= [ρf (1 − α)Ax−s Vf z] + [ρf (1 − α)Ax−s Vf2 ] z
∂t ∂z x ρρgf
α= (16)
+Ax−s ΓVi z (8) S(1 − x) + x ρρgf
H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901 1891

The liquid velocity can be expressed by combining Eqs. (15) formed into the following equation:
and (16) as 
dP ṁ2 ∂F ∂S ∂P ∂F ∂S ∂x ∂F ∂x ∂F ∂ρf ∂P
G − = + +
Vf =   (17) dz Ax−s ∂S ∂P ∂z ∂S ∂x ∂z ∂x ∂z ∂ρf ∂P ∂z
Sρf 
S(1−x)+x(ρf /ρg ) ∂F ∂ρg ∂P ∂F ∂Ax−s
+ + (22)
The modified momentum equation is obtained after inserting ∂ρg ∂P ∂z ∂Ax−s ∂z
Eq. (17) into Eq. (13) as
⎧  ⎫ The second and third terms do not seem to affect the choking
⎨ ρf ⎬
dP ∂ G S(1 − x) + x ρg
flow rate because it is not coupled with the pressure drop. How-
− =G (xS + 1 − x) ever, it actually contributes to the choking flow rate when there
dz ∂z ⎩ Sρf ⎭
is phase change. Terms inside the parenthesis of the right-hand
  side of Eq. (22), ∂F/∂S, ∂F/∂x, ∂F/∂ρf , ∂F/∂ρg , and ∂F/∂Ax−s ,
ṁ2 ∂ 1 S(1 − x) + x ρρgf
= (xS + 1 − x) (18) can be obtained from Eq. (18) by partial differentiation:
Ax−s ∂z Ax−s Sρf  
∂F 1 ρf Φ1
= −x S − S + Sx + x(x − 1) =
2 2
The term inside the bracket of the second equation in Eq. (18) ∂S Ax−s S 2 ρf ρg Ax−s
is a function of area, flow quality, gas density, liquid density and (23)
slip ratio, i.e.
    
F = F (Ax−s , x, ρf , ρg , S) (19) ∂F 1 ρf ρf
= (1 − 2x) S − S +
2
− S 1 − 2x
∂x Ax−s Sρf ρg ρg
In general, the flow quality, liquid density and gas density are
functions of pressure and entropy. The slip ratio is a function of Φ2
= (24)
the quality and pressure. Therefore, differentiation of the term Ax−s
inside bracket with respect to z is expanded by the chain rule as
∂F (xS + 1 − x)(1 − x) Φ3
∂F ∂F ∂S ∂F ∂x ∂F ∂ρf ∂F ∂ρg ∂F ∂Ax−s =− = (25)
= + + + + ∂ρf Ax−s ρf2 Ax−s
∂z ∂S ∂z ∂x ∂z ∂ρf ∂z ∂ρg ∂z ∂Ax−s ∂z
   
∂F ∂S ∂P ∂S ∂x ∂F ∂x ∂P ∂x ∂s ∂F (xS + 1 − x)x Φ4
= + + + =− = (26)
∂S ∂P ∂z ∂x ∂z ∂x ∂P ∂z ∂s ∂z ∂ρg Ax−s Sρg2 Ax−s
   
∂F ∂ρf ∂P ∂ρf ∂s ∂F ∂ρg ∂P ∂ρg ∂s ρf
+ + + + ∂F 1 S(1 − x) + x ρg Φ5
∂ρf ∂P ∂z ∂s ∂z ∂ρg ∂P ∂z ∂s ∂z =− (xS + 1 − x) = (27)
∂Ax−s Ax−s Sρf Ax−s
∂F ∂Ax−s
+
∂Ax−s ∂z In Eq. (22), all variables, except ∂S/∂P, ∂S/∂x and ∂x/∂z,
   are known. Therefore, the choking relation can be obtained
∂F ∂S ∂P ∂S ∂x ∂P ∂x ∂s
= + + from the momentum equation if the constitutive relations for
∂S ∂P ∂z ∂x ∂P ∂z ∂s ∂z
    ∂S/∂P, ∂S/∂x and ∂x/∂z are established. Indeed, these parame-
∂F ∂x ∂P ∂x ∂s ∂F ∂ρf ∂P ∂ρf ∂s ters are the keys in solving the critical flow of the two-phase
+ + + +
∂x ∂P ∂z ∂s ∂z ∂ρf ∂P ∂z ∂s ∂z flow because they reflect the degree of the mechanical and
  thermal non-equilibrium. While ∂S/∂P is related to the mechan-
∂F ∂ρg ∂P ∂ρg ∂s ∂F ∂Ax−s
+ + + (20) ical non-equilibrium and ∂x/∂z is related to the thermal non-
∂ρg ∂P ∂z ∂s ∂z ∂Ax−s ∂z equilibrium, respectively, ∂S/∂x is the parameter that account
for both mechanical and the thermal non-equilibrium.
In this research, the flow is assumed to be isentropic, which In the present analysis, the drift flux model is introduced to
implies that the entropy does not change along the flow direction. provide constitutive relations for ∂S/∂P such that the slip ratio
Based on this assumption, Eq. (20) is expressed as is written as
   
∂F ∂F ∂S ∂P ∂S ∂x ∂F ∂x ∂F ∂ρf ∂P x(Co − 1)ρf ρf  Vgj 
= + + + S = Co + + (28)
∂z ∂S ∂P ∂z ∂x ∂z ∂x ∂z ∂ρf ∂P ∂z ρg (1 − x) G(1 − x)
 
∂F ∂ρg ∂P ∂F ∂Ax−s
+ + (21) The distribution parameter, Co , indicates the effect of the
∂ρg ∂P ∂z ∂Ax−s ∂z radial void fraction and volumetric flux distribution. The drift
velocity, Vgj  is a measure of the local slip. The distribution
Therefore, rewriting Eq. (21) in terms of pressure drop and parameter and drift velocity are depending on the flow regime
comparing it with Eq. (18), the momentum equation is trans- and flow component.
1892 H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901

In the air–water condition, it is an adiabatic or no phase ∂S x(Co − 1)ρf


= 2 = Ω3 (38)
change flow, i.e. ∂ρg ρg (1 − x)
∂x
=0 (29) Using Eqs. (36)–(38), Eq. (35) is expressed as
∂z
∂S ∂Co ∂ρf ∂ρg
Therefore, Eq. (22) is simplified as = Ω1 + Ω2 + Ω3 (39)
∂P ∂P ∂P ∂P

dP ṁ2 ∂F ∂S ∂P ∂F ∂ρf ∂P ∂F ∂ρg ∂P It is required to obtain Co and Vgj  based on the known
− = + +
dz Ax−s ∂S ∂P ∂z ∂ρf ∂P ∂z ∂ρg ∂P ∂z parameters. Hibiki and Ishii (2003) developed the correlations
 for both parameters depending on the flow regime. The devel-
∂F ∂Ax−s
+ (30) oped correlations for Co depending on the flow regime for the
∂Ax−s ∂z
horizontal flows are given by
Bubbly flow
For the axial pressure drop, Eq. (30) is transformed as   
  −22Dsm 
∂F ∂S ∂F ∂ρf ∂F ∂ρg dP Co = 2.0 exp(−0.000584 Re) + 1.2 1 − exp
− 1+Φ +Φ +Φ D
∂S ∂P ∂ρf ∂P ∂ρg ∂P dz 
∂F ∂Ax−s ×{1 − exp(−0.000584 Re)} − 2.0 exp(−0.000584 Re)
=Φ (31)
∂Ax−s ∂z   
−22Dsm 
where Φ is the ṁ2 /A
x−s . In the parenthesis of the left-hand
+1.2 1 − exp
D
side, ∂F/∂S, ∂F/∂ρf , ∂F/∂ρg , and ∂F/∂Ax−s , are obtained from 
Eqs. (23), (25)–(27). Therefore, Eq. (31) is expressed as ρg
×{1 − exp(−0.000584 Re)} − 1 (40)
   ρf
∂S ∂ρf ∂ρg dP ∂Ax−s
− 1 + G Φ12
+ Φ3 + Φ4 = G2 Φ5
∂P ∂P ∂P dz ∂z Slug flow
(32) 
ρg
Co = 1.2 − 0.2 (41)
The choking condition occurs when the bracket on the left- ρf
hand side of Eq. (32) equals zero, i.e.
Churn flow
−1 
G2c = ∂ρ
(33) ρg
∂S
Φ1 ∂P + Φ3 ∂ρ
∂P + Φ4 ∂P
f g
Co = 1.2 − 0.2 (42)
ρf
The partial differentiation of S with respect to P is subdivided
Annular flow
with the related variables:
∂S ∂S ∂Co ∂S ∂x ∂S ∂ρf ∂S ∂ρg ∂S ∂G 1−α
= + + + + Co ≈  +1 (43)
ρ
∂P ∂Co ∂P ∂x ∂P ∂ρf ∂P ∂ρg ∂P ∂G ∂P α + 4 ρgf
∂S ∂  Vgj 
+ (34) The developed correlations of the drift velocity, Vgj , for
∂  Vgj  ∂P
the different flow regime are as follows:
Bubbly flow
Currently, it is assumed that the flow is adiabatic. On choking
 1/4  1/2
condition, mass flux is independent of pressure. In addition, it is
√ MF∞ σ 18.67(1 − α)5/2 MMF∞ F
assumed that the change of the drift velocity is negligible with  Vgj = 2 ×  3/7
respect to the pressure. Based on the above-mentioned assump- ρf2 1 + 17.67(1 − α)6/7 MF
tions, Eq. (34) is simplified as follows: MF∞

(44)
∂S ∂S ∂Co ∂S ∂ρf ∂S ∂ρg
= + + (35)
∂P ∂Co ∂P ∂ρf ∂P ∂ρg ∂P Slug flow
The partial differentiation of the slip ratio with respect to three  1/2
MF D
variables in Eq. (35) is obtainable from Eq. (28) as follows:  Vgj = 0.37 (45)
ρf (1 − α)
∂S xρf
=1+ = Ω1 (36) Churn flow
∂Co ρg (1 − x)
 1/4  1/4
∂S x(Co − 1)  Vgj  √ MF∞ σ MF
= + = Ω2 (37)  Vgj = 2 × (46)
∂ρf ρg (1 − x) G(1 − x) ρf2 MF∞
H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901 1893

Annular flow eter is the gas sonic velocity. Therefore, in this condition, Eq.
(54) is simplified as
1−α
V̄gj ≈  j (47)
ρ 1
α + 4 ρgf G2 = − (55)
∂ρ
(Φ1 Ω1 Θ1 + Φ1 Ω3 + Φ4 ) ∂Pg
V̄gj for annular two-phase flow was developed by Ishii (1977).
In the correlation for Vgj , MF∞ is the frictional pressure The choking mass flux based on Eq. (55) is obtained through
gradient in a single-bubble system which is given by several calculation steps. Fig. 1 is the flowchart for calculating
the choking mass flux for the adiabatic air–water condition as the
f particular case. Because Eq. (55) is independent of the upstream
MF∞ = ρf Vf 2 (48)
2D condition, it is required to input the choking pressure, tempera-
where f and D are the wall friction factor and pipe diameter, ture and flow quality to use it at the choking plane. The choking
respectively. MF is the frictional pressure gradient in multi- pressure is assumed to be 0.56 of the upstream pressure. To cal-
bubble system. This is defined as culate the slip ratio based on the drift flux model, void fraction,
  liquid velocity, drift velocity, and phase distribution parameter
4τw dP are required. Because all parameters are related to each other,
MF ≡ = − (49)
D dz F iteration is necessary. Therefore, the mass flux and slip ratio
are assumed before calculating the previous parameters. Based
The pressure drop per unit length is calculated using the on the assumed mass flux and slip ratio, the new slip ratio is
Lockhart–Martinelli correlation (1949). Using this correlation obtained. The iteration for slip ratio is continued for fixed mass
for Co , it is possible to calculate ∂Co /∂P. For example, ∂Co /∂P flux until the error criterion is satisfied. The error criterion for
for slug and churn flow is given by the slip ratio is
∂Co ∂Co ∂ρg ∂Co ∂ρf  
= + (50)  S1 − S2 
∂P ∂ρg ∂P ∂ρf ∂P   < 0.001 (56)
 S 
1
As the example, from Eqs. (41) and (42), ∂Co /∂ρg and
Once obtaining a reasonable slip ratio, the choking mass flux
∂Co /∂ρf are obtained as follows:
is calculated. The error criterion for the choking mass flux is
 1/2  
∂Co 1  G1 − G 2 
= −0.1 = Θ1 (51)   < 0.001
∂ρg ρg ρf  G  (57)
1
 
∂Co 0.1 ρg 1/2 If the error criterion is not satisfied by the assumed mass flux,
= = Θ2 (52)
∂ρf ρf ρf the calculation is returned to the fourth step in the flowchart with
a new mass flux assumption.
The variable Θ1 and Θ2 vary depending on the flow regime.
Among the number of two-phase choking model, the Fauske
By inserting Eqs. (51) and (52) into Eq. (39), it can be seen that
model (1962) is applicable for the air–water flow which is cor-
∂S/∂P is affected by the density change of the liquid and the gas
responding to Eq. (7). Originally, this model was developed for
with respect to the pressure change as follows:
the one-component two-phase flow. However, Fauske mentioned
 
∂S ∂ρg ∂ρf ∂ρf ∂ρg that this model is applicable to the air–water flow with no phase
= Ω1 Θ1 + Θ2 + Ω2 + Ω3 change assumption, dx/dP = 0. In addition, dvf /dP is negligible
∂P ∂P ∂P ∂P ∂P
compared to the dvg /dP, therefore, specific volume change with
∂ρg ∂ρf
= (Ω1 Θ1 + Ω3 ) + (Ω1 Θ2 + Ω2 ) (53) respect to the pressure change is neglected. With these assump-
∂P ∂P tions, the Fauske model is simplified as
Thus, the two-component choking flow equation may be writ-  1/2
ten as S
Gcrit = − dv
(58)
1 {(1 − x + Sx)x} dPg
G2c = − (54)
(Φ1 Ω1 Θ2 + Φ1 Ω2 + Φ3 ) ∂ρ
∂P
f

∂ρ The developed choking flow model, Eq. (55), is compared


+(Φ1 Ω1 Θ1 + Φ1 Ω3 + Φ4 ) ∂Pg with the experimental data for the nozzle geometry (Yoon et al.,
2002) and the existing model, Fauske model. The choking flow
experiments were performed using PUMA (Purdue University
4. Results Multidimensional integral test Assembly) facility of PURDUE
University. The PUMA test facility is a scaled integral model of
4.1. For given choking pressure the General Electric Nuclear Energy, simplified Boiling Water
Reactor (SBWR) (GE Nuclear Energy, 1992; Ishii et al., 1996).
Air–water condition is the adiabatic flow. Therefore, there is Experimental upstream pressure conditions were 0.207, 0.345,
no flashing and no quality change. Usually, the dominant param- 0.517, 0.689, and 1.034 MPa.
1894 H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901

Fig. 1. Calculation procedure.

The reactor pressure vessel (RPV) and drywell were used as section was 25.4 mm except at the choking section (nozzle and
the pressure boundary in this experiment. The RPV was used as orifice). The throat size was 5.4 mm in both geometries.
the upstream stagnation reservoir and the drywell was used as the In the Fauske model, the choking mass flux is calculated
downstream receiver with the horizontal test section connected based on the predetermined choking pressure which is same as
between the RPV and the drywell. The inner diameter of the test the developed model calculation. For the developed model, the
H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901 1895

choking mass flux is calculated with some simplification. In this the flow regime for every flow condition. According to results
calculation, the correlations for slug flow regime were used for (Fig. 2(a)–(e)), the Fauske model overpredicts the choking mass
Vgj  and Co because main discrepancy between the existing flux at low quality region in all pressure conditions. However,
models and the experimental data generally occurs at the low the mechanical non-equilibrium model (MNEM) developed pre-
quality region. For detailed calculation, it is required to identify dicts the choking mass flux of experimental data through whole

Fig. 2. Developed model comparison with existing model and experimental data for given choking pressure.
1896 H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901

experimental flow quality region (up to 10%) and pressure con- prediction of MNEM will be improved more. The discrepancy
ditions. The small discrepancy between the experimental data between the Fauske model and the MNEM decreases as the flow
and MNEM may be caused by one flow regime consideration. If quality increases. This indicates that the constant slip ratio based
the correct correlation is used depending on the flow regime, the on density ratio is applicable only to the high quality region.

Fig. 3. Choking mass flux comparison between developed model and Purdue experimental data by solving momentum equation.
H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901 1897

4.2. Solving momentum equation choking mass flow rate is obtained based on the inlet quality
and the numerically obtained exit pressure. The slip ratio at the
As the next application of the developed model, the momen- inlet and outlet is calculated based on Eq. (28) to calculate the
tum equation (18) is solved numerically through a nozzle. In this choking mass flux. The choking mass flux is obtained at the exit
calculation, the pressure profile can be obtained based on the of the nozzle from Eq. (54). In this approach, the effect of the
inlet condition which is the same with the experiment. Finally, liquid sound velocity is not neglected.

Fig. 4. Slip ratio comparison between developed model and Purdue experimental data by solving momentum equation.
1898 H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901

Fig. 5. Void fraction comparison between developed model and Purdue experimental data by solving momentum equation.
H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901 1899

Fig. 3(a)–(e) shows the choking mass flux comparison


between the experimental data [33] and the developed model.
The model’s prediction has a good agreement with the exper-
imental data. This indicates that the momentum equation is
solved correctly with the iteration method.
In this approach, it is possible to calculate the inlet slip ratio
using the drift flux model. In the experiment, the slip ratio was
measured at the upstream of the choking point. According to
Fig. 4(a)–(e), the calculated slip ratios smaller than the measure
slip ratio in the churn and annular flow regime. Fig. 5(a)–(e)
are the void fraction comparison. The void fraction predicted
is quite accurate compared to the experimental data. Because
of the underprediction of the slip ratio in the churn and annu-
lar flow regime, the calculated void fraction is slightly over-
predicitng the measure void fraction. In this approach, it is
confirmed that the choking mass flux, slip ratio, void fraction,
and quality are connected each other, therefore, the choking
mass flux is the function of the slip ratio, void fraction and
quality.

4.3. Application to FOSSEGRIMEN experiment

The developed model is applied to the other experiment for


the model assessment. The chosen experiment is performed
by Lemonnier and Selmer-Olsen (1992) and called FOSSEG-
RIMEN experiment. This loop was built in order to provide
the experimental data for choking flow condition using the
air–water flow. The geometry of the test section is the con-
verging and diverging nozzle. The inlet diameter is 15 mm and
the throat diameter is 5 mm. The length in the converging noz-
zle section is 23.5 mm. Because the choking usually occurs
at the smallest flow area, the momentum equation including
the choking model developed is solved in the converging noz-
zle section. One of the characteristics of this experiment is to
use the special liquid injection devices in order to investigate
the effect of the liquid distribution. In the one liquid injec-
tion device, the liquid is injected centrally and the other device
injects the liquid in the vicinity of the wall as a thin liquid
film.
In this experiment, the inlet flow quality is controlled from
0.0009 to 0.965 for 0.2, 0.4 and 0.6 MPa. Fig. 6(a)–(c) shows
the comparison of the predicted choking mass flux with the
experimental data for three different upstream pressure. Between
two injection methods, the chosen data are for the annular
injection, which indicates that the liquid is injected as the
annular film flow. According to the comparison, the model
predicts the experimental data reasonably. As the upstream
pressure increase, the accuracy is improved. At the low qual-
ity region for the low upstream pressure, the discrepancy
may occur due to the different flow regime. In this experi-
ment, the liquid was injected as the liquid annular film even
in the low quality region. However, the model considers the
flow regime corresponding to low quality as the bubbly or Fig. 6. Choking mass flux comparison between developed model and FOSSEG-
slug flow regime. Therefore, the accurate information of the RIMEN experimental data by solving momentum equation.
flow regime and the adequate constitutive relation are very
important in the slip ratio calculation and the choking mass
flux.
1900 H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901

5. Conclusions Henry, R.E., Fauske, H.K., 1971. The two-phase critical flow of one compo-
sition mixtures in nozzle, orifices, and short tubes. J. Heat Transfer 93,
179–187.
In this research a mechanistic model which considers the
Hibiki, T., Ishii, M., 2003. One-dimensional drift-flux model and constitutive
mechanical non-equilibrium for the two-phase choking flow equations for relative motion between phases in various two-phase flow
is developed. The slip ratio which is defined in the drift flux regimes. Int. J. Heat Mass Transfer 46 (25), 4935–4948.
model is used to identify the impact parameters on the slip Ishii, M., 1977. One-dimensional drift-flux model and constitutive equations
ratio. Because the slip ratio in the drift flux model is related to for relative motion between phases in various two-phase flow regimes,
ANL-77-47.
the distribution parameter and drift velocity, the adequate cor-
Ishii, M., Revankar, S.T., Leonardi, T., Dowlati, R., Bertodano, M.L., Babelli,
relations depending on the flow regime are introduced in this I., Wang, W., Pokharna, H., Ransom, V.H., Viskanta, R., Han, J.T.,
study. In this mechanistic modeling approach, the choking mass 1996. Scientific Design of Purdue University Multi-Dimensional Inte-
flow rate is expressed by the function of pressure, quality and gral Test Assembly (PUMA) for GE SBWR, Purdue University Report
slip ratio. For the mechanical non-equilibrium modeling, the PU-NE-94/1, U.S. Nuclear Regulatory Commission Report NUREG/CR-
6309.
thermal non-equilibrium is neglected by assuming equilibrium
Khajehnajafi, S., Shinde, A., 1995. Prediction of discharge rate from pres-
flow. surized vessel blowdown through sheared pipe. Process Safety Progr. 14
This model is compared with the air–water experimental (1), 22–25.
data as the two-phase two-component flow. The mechanical Leung, J.C., 1986. A generalized correlation for one-component homo-
non-equilibrium model (MNEM) is applicable to the two-phase geneous equilibrium flashing choked flow. AIChE J. 32, 1743–
1746.
two-component flow because there is no phase change. The
Leung, J.C., Grolme, M.A., 1986. A generalized correlation for flash-
MNEM developed considers all factors which may affect the ing choked flow of initially subcooled liquid. AIChE J. 34 (4), 688–
slip ratio including the phase profile by introducing the drift 691.
flux model and reasonable correlation. Fauske model overpre- Leung, J.C., Epstein, M., 1990. A generalized correlation for two-phase
dicts the choking mass flux at low quality region in all pressure nonflashing homogeneous choked flow. J. Heat Transfer 112, 528–
530.
conditions. However, the MNEM developed predicts the chok-
Lemonnier, H., Selmer-Olsen, S., 1992. Experimental investigation and physi-
ing mass flux of experimental data through whole experimental cal modeling of two-phase two-component flow in a converging-diverging
flow quality region (up to 10%) and pressure conditions. The nozzle. Int. J. Multiphase Flow 18 (1), 1–20.
comparison of predicted choking model with other experimen- Lenzing, T., Friedel, L., Cremers, J., Alhusein, M., 1998. Prediction of the
tal data in high quality region (up to 80%) is also quite reasonable maximum full lift safety valve two-phase flow capacity. J. Loss Prevent.
Process Ind. 11, 307–321.
with a small error.
Levy, S., Abdollahian, D., 1982. Homogeneous non-equilibrium critical flow
model. Int. J. Heat Mass Transfer 25 (6), 759–770.
References Lockhart, R.W., Martinelli, R.C., 1949. Proposed correlation of data for
isothermal two-phase, two-component flow in pipes. Chem. Eng. Prog. 5,
A Group of Expert of the NEA Committee on the Safety of Nuclear Instal- 39–48.
lations, 1982. Critical Flow Modeling in Nuclear Safety, Nuclear Energy Lottes, P.A., Petrick, M., Marchaterre, J.F., 1959. Lecture notes on
Agency, June. heat extraction from boiling water power reactors, Presented at the
Abdollahian, D., Healzer, J., Janssen, E., Amos, C., 1982. Critical Flow Advanced Summer Institute at Kjeller, Norway, August 17–29, ANL-
Data Review and Analysis, Final Report, S. Levy, Inc. EPRI Report NP- 6063.
2192. Moody, F.J., 1965. Maximum flow rate of a single component, two-phase
Bouré, J.A., 1974. Two-Phase Flows with Application to Nuclear Reactor mixture. Trans. ASME J. Heat Transfer 86, 134–142.
Design Problem, Lecture Series. von Karman Institute for Fluid Dynam- Rateau, A., 1905. Experimental Researches on the Flow of Steam through
ics, Grenoble, France. Nozzles and Orifice, to which is added a Note on the Flow of Hot Water.
Cruver, J.E., Moulton, R.W., 1967. Critical flow of liquid vapor mixture. A. Constable and Co., Ltd., London.
AIChE J. 13 (1), 52–60. Reocreux, M., 1974. Contribution à létude des debits critiques en enoulement
Darby, R., 1998. Perspectives on relief valves sizing for two-phase flow. In: diphasique eau-vapeur, PhD thesis, Lúniversité Scientifique et médicale
Proceedings of the International Symposium on Runaway Reactions, Pres- de Grenoble.
sure Relief Design, and Effluent Handling, March 11–13, New Orleans, Richter, H.J., 1983. Separated two-phase flow model: application to critical
Louisiana, pp. 365–397. two-phase flow. Int. J. Multiphase Flow 9 (5), 511–530.
Epstein, M., Henry, R.E., Midvidy, W., Pauls, R., 1983. One-dimensional Saha, P., 1978. A review of two-phase steam–water critical flow models with
modeling of two-phase jet expansion and impingement, thermal- emphasis on thermal non-equilibrium, BNL-NUREG-50907.
hydraulics of nuclear reactors II. In: Proceedings of the Second Inter- Sauvage, 1892. Ecoulement de L’eau des Chaudieres. Annales des Mines 9
national Topical Meeting on Nuclear Reactor Thermal-Hydraulics, Santa (II).
Barbara, CA. Schrock, V.E., Starkman, E.S., Brown, R.A., 1977. Flashing flow of initially
Fauske, H.K., 1962. Contribution to the theory of two-phase, one-component subcooled water in convergent-divergent nozzle. Trans. ASME J. Heat
critical flow, PhD Thesis. Transfer 99, 263–268.
Fauske, H.K., 1963. Two-phase critical flow with application to liquid metal Schrock, V.E., Amos, C.A., 1984. Two-phase flow and heat transfer,
system, ANL-6633. China–U.S. Seminar on Two-Phase Flows and Heat Transfer, Sian, China,
Fauske, H.K., 1985. Flashing Flows or: some practical guideline for emer- pp. 115–138.
gency releases. Plant/Operations Progr. 4 (3), 132–134. Starkman, E.S., Schrock, V.E., Neusen, K.F., Maneely, D.J., 1964. Expansion
Fthenakis, V.M., Rohatgi, U.S., Chung, B.D., 2003. A simple model for of a very low quality two-phase fluid through a convergent-divergent
predicting the release of a liquid–vapor mixture from a large break in a nozzle. J. Basic Eng. TRANS. ASME Ser. D 86 (2), 247–256.
pressurized container. J. Loss Prevent. Process Ind. 16, 61–72. Trapp, J.A., Ransom, V.H., 1982. A choked-flow calculation criterion for non-
GE nuclear Energy, 1992. SBWR Standard Safety Analysis Report, Report homogeneous, nonequilibrium, two-phase flows. Int. J. Multiphase Flow
No. 25A5113 Rev. A. 8 (6), 669–681.
H.J. Yoon et al. / Nuclear Engineering and Design 236 (2006) 1886–1901 1901

Wallis, G.B., 1980. Critical two-phase flow. Int. J. Multiphase Flow 6, Zimmer, G.A., Wu, B.J.C., Leonhardt, W.J., Abuaf, N., Jones, O.C.,
97–112. 1979. Pressure and void distributions in a conversing-diverging noz-
Yoon, H. J., Ishii, M., Revankar, S. T., Wang, W., 2002. Mechanical non- zle with non-equilibrium water vapor generation, BNL-NUREG-
equilibrium effect on choking flow at low pressure in air–water experi- 26003.
ment, ICONE-10, April 14–18, Virginia.

You might also like