CSNI Report No. 49 ARCHIVES NUCLEAR SAFETY DIVISION TWO-PHASE CRITICAL FLOW MODELS A technical addendum to the CSNI state of the art report on critical flow modelling F. D'AURIA - P. VIGNI Université degli Studi di Pisa Istituto di Impianti Nucleâri May 1980 Work sponsored by: COMITATO NAZIONALE PER L'ENERGIA NUCLEARE Roma NUCLEAR ENERGY AGENCY ORGANIZATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT COMMITTEE ON THE SAFETY OF NUCLEAR INSTALLATIONS
277
Embed
archives - International Nuclear Information System (INIS)
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
CSNI Report No. 49
ARCHIVES
NUCLEAR SAFETY DIVISION
TWO-PHASE CRITICAL FLOW MODELS
A technical addendum to the CSNI state of the art report on critical flow modelling
F. D'AURIA - P. VIGNI
Un ivers i té degli Studi di Pisa Is t i tu to di Impianti Nucleâr i
May 1980
Work sponsored by:
COMITATO NAZIONALE PER L'ENERGIA NUCLEARE
Roma
NUCLEAR ENERGY AGENCY ORGANIZATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT
COMMITTEE ON THE SAFETY OF NUCLEAR INSTALLATIONS
CSNI Report No. 49
TWO-PHASE CRITICAL FLOW MODELS
A technical addendum to the CSNI state of the art report
on critical flow modelling
F. D'AURIA - P. VIGNI
U n i v e r s i t é d e g l i S t u d i di P i s a
I s t i t u t o d i I m p i a n t i N u c l e a r i
May 1980
Work sponsored by:
COMITATO NAZIONALE PER L'ENERGIA NUCLEARE
Roma
NUCLEAR ENERGY AGENCY ORGANIZATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT
COMMITTEE ON THE SAFETY OF NUCLEAR INSTALLATIONS
Riprodotto in offset presso il Laboratorio Tecnografico délia Direzione Centrale Relazioni Esterne del CNEN - Vialc Rcgina Margherita 125, Roma
The Organisation for Economic Co-operation and Development (OECD) was set up under a Convention signed in Paris on 14th December, 1960, which provides that the OECD shall promote policies designed:
— to achieve the highest sustainable economic growth and employment and a rising standard of living in Member countries, while maintaining financial stability, and thus to contribute to the development of the world economy ;
— to contribute to sound economic expansion in Member as well as non-member countries in the process of economic development;
— to contribute to the expansion of world trade on a multilateral, non-discriminatory basis in accordance with international obligations.
The Members of OECD are Australia, Austria, Belgium, Canada, Denmark, Finland, France, the Federal Republic of Germany, Greece, Iceland, Ireland, Italy, Japan, Luxembourg, the Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States.
The OECD Nuclear Energy Agency (NEA) was established on 20th April 1972, replacing OECD's European Nuclear Energy Agency (ENEA) on the adhesion of Japan as a full Member.
NEA now groups all the European Member countries of OECD and Australia, Canada, Japan, and the United States. The Commission of the European Communities takes part in the work of the Agency.
The primary objectives of NEA are to promote co-operation between its Member governments on the safety and regulatory aspects of nuclear development, and on assessing the future role of nuclear energy as a contributor to economic progress.
This is achieved by: — encouraging harmonisation of governments' regulatory policies and practices in
the nuclear field, with particular reference to the safety of nuclear installations, protection of man against ionising radiation and preservation of the environment, radioactive waste management, and nuclear third party liability and insurance;
— keeping under review the technical and economic characteristics of nuclear power growth and of the nuclear fuel cycle, and assessing demand and supply for the different phases of the nuclear fuel cycle and the potential future contribution of nuclear power to overall energy demand;
— developing exchanges of scientific and technical information on nuclear energy, particularly through participation in common services;
— setting up international research and development programmes and undertakings jointly organised and operated by OECD countries.
In these and related tasks, NEA works in close collaboration with the International Atomic Energy Agency in Vienna, with which it has concluded a Co-operation Agreement, as well as with other international organisations in the nuclear field.
LEGAL NOTICE
The opinion expressed and arguments employed in this publication are the responsibility of the Authors and do not necessarily represent those of the OECD.
Copyright OECD, 19^0
Queries concerning permissions or translation rights should be addressed to: Director of Information, OECD
2, rue André-Pascal, 75775 PARIS CEDEX 16. France.
Riprodocto in offset presso il Laboratorio Tecnografico della Direzione Cen-trale Relazioni Esterne del CNEN Viale Regina Margherita 125, Roma
- Ill -
ACKHOWLEVGEMENT
The. authors would like, to ac.knowle.dge. M. Mazzlnl o£ Visa Uni
ve.i<y and G. SantafioA&a o& CNEhl &on the.lt. hzlp dating the. de.-
ve.lopme.nt o£ the. wonk, and all the. pzK&on& who have contributed
f friction coefficient f„ Fanning friction coefficient F
F force
Fr Froude number
g gravity constant (exceptionally used in this work)
G flowrate = TA
h specific enthalpy
H total enthalpy
k slip ratio
kp slip ratio by Fauske
kM slip ratio by Moody
L pipe lenght or work
m mass
M total mass
n polytropic coefficient
p pressure
P wetted flowing duct perimeter
- XII -
Pr Prandtl number
q,Q heat exchanged by two phase mixture
R energy loss (Par. 3.3.3.10)
R bubble radius
Re Reynold number
s specific entropy
S thermal exchange surface
t time
t dimensional time
T temperature
u phase velocity when more than one dimension is considered
U vectorial velocity
v specific volume
V perturbation velocity
x quality or cartesian coordinate when more than one dimension is considered
w phase velocity
W specific volumetric flowrate
We Weber number
z cartesian coordinate
a void fraction
Y c /c 1 p' v
r specific flowrate
A(E) determinantal equation used when energy equation is adopted
A(s) determinantal equation used when entropy equation is adopted
e circumference
ç pressure loss n Pc/P0 X water conductivity y dynamic viscosity
- XIII -
v
ç
p
a
T
X
SI
kinematic viscosity
loss coefficient
mixture density
surface tension
wall shear
two phase flow multiplier
bubble wall interfacial heat flux
evaporation rate
volume
moreover: D/DN total derivative with respect to N
3/8N partial derivative with respect to N
<N> cross section average value
N vectorial quantity
(N is the generical variable)
SUBSCRIPTS INDEX
a
A
c
con
down
e
ex
E
f
fg
F
ambient
assigned
critical value
condensation
downstream
exit
exchanged
equilibrium
liquid phase
such that N-. =N -N-fg g f
frictional
- XIV -
g vapor phase
G vapor phase (only in Par. 3.3.4.?.9)
GSP gas single phase
h constant enthalpy
H homogeneous
HE homogeneous equilibrium
HF homogeneous frozen
i inlet
irr irreversible
L liquid phase (only in Par. 3.3.4.19)
LSP liquid single phase
m mixture
M momentum averaged
n phase index (may be f or g)
ns non isentropic
o initial value or reservoir value
p constant pressure
rev reversible
s constant entropy
sat saturation
st stationary
t throat
T total
up upstream
vise viscous
w wall
- XV -
SUPERSCRIPTS INDEX
x critical value
static value
derivative with respect to time
NOTE: When in the text we write "Mpody" we intend "Moody 1965'
•
3.1 INTRODUCTION
In the previous chapters the phenomenological aspects of the
Loss of Coolant Accident (LOCA) have been emphasized from a ther
mo-fluiddynamic point of view. In particular the concept of two
phase "maximum" flow has been analyzed, outlining the main para
meters it depends upon, and consequently the difficulties in
the mathematical models.
As far as the theoretical study is concerned, the difficul
ties are mainly due to "quantitative" evaluation of the follo
wing points :
- multidimensional effects and flow pattern (bubble flow, slug
flow, annular flow, etc.);
- phase change, generally under strong non-equilibrium conditions, 2
which take place when the high pressure fluid (70-150 kg/cm )
meets the ambient pressure (problems related to mass, momen
tum and energy exchange between liquid and vapour phases and
between each phase and the exterior);
- influence of the geometrical and thermodynamical situation on
critical conditions (internal flow paths, rupture position
with respect to liquid level, heat source, etc.).
In particular the last point makes troublesome the formula
tion of one theory valid for all plant situations.
In this chapter models presented so far are taken into ac
count, in the context of the SOAR concerning both the applicabi^
lity to any real situation and the future development of the
theoretical problem.
The following section deals with the criteria for the classji
fication of different models.
9
- 3 -
3.2 CLASSIFICATION OF THE MODELS
3.2.1 - Criteria
In the last 40 years, several tens of models have been publi
shed about calculation of the critical flow rate of a two phase
mixture.
In some models there is no theoretical support and they are
rather semi-empirical formulas, linking critical flow rate to
thermodynamical variables, generally representing the fluid sta
te in the pressure vessel. These models adopt adimensional empi
rical coefficients in order to fit the esperimental data.
Other models, instead, derive from the solution of a set of
two or more (up to 6) equilibrium equations, describing the con
servation of mass, momentum and energy, for each phase separate
ly or for homogeneous mixtures.
From a mathematical point of view, it ought to be possible to
draw, with suitable simplifying assumptions from complex theo
ries (e.g. based upon all of the six balance equations) models
with a smaller number of equations; this is hardly ever achieva
ble as in most models the result is obtained introducing evolu
tion laws, which are more or less particular or arbitrary.
Moreover, many theories become acceptable for the engineering
calculations, if these are related to a given experimental a£
paratus and/or to conditions assigned beforehand. On the contra
ry, if boundary conditions (that are usually simplifying assump
tions) are varied, the results sometimes diverge or become unusa
ble.
From this brief outline, one can foresee quite different for
mulations of the models and difficulty in group classification.
- 4 -
The following subdivision in four groups has usually been /10, 54/
adopted
1) Theories which assume thermodynamic equilibrium throughout the expansion
2) Non-equilibrium theories
1.a - homogeneous theories (k=l)
l.b - non-homogeneous theories (k ? 1)
2.a - "frozen"theories (k/ 1) but having given value (*)
2.b - non-homogeneous theories (k f 1)
This subdivision is maintained throughout this work.
Nevertheless, other characteristics allowing a deeper compari
son and giving an idea of the model applicability are referred
to; such characteristics are listed as following:
A) Model formulation
A.l) number of conservation equations, e.g. equations taking
account of mass, momentum and energy conservation of the
flowing system;
A.2) number of state and/or transformation equations: the sta
te equations are those describing the system state
through some variables (e.g. p, T, h, s, etc.); the tran
sformation equations outline the state change of the system,
according to given criteria (**)
(x) By "frozen" we intend that the composition at the inlet of the flowing pipe is the same as at the outlet.
(xx) The two types of equation have been matched in order to avoid sophi sticated definitions to distinguish them. Considering, e.g., the e-quation ds=0, at the same time it is a state equation and it indiyi duates an isoentropical transformation.
- 5 -
A.3) number of constitutive equations: they are usually empi
rical equations, derived either from adimensional analy
sis or from assumptions regarding the system behaviour ;
A.4) number of analytical conditions, added to the model, so
metimes without any physical meaning;
A.5) Necessity of semi-empirical parameters for problem solu
tion.
B) Assumptions (phenomenological aspects and parametersconside-
red)
B.l) transient phenomena;
B.2) multidimensional-effects;
B.3) non-homogeneity in pressure vessel;
B.4) heat exchange with exterior;
B.5) pipe length (essentially for friction);
B.6) orifice (I), nozzle (II), constant pipe area (III), any
kind of pipe (IV).
C) Output
C.l) diagrams or relations linking p0, ho and/or x0 to T;
C.2) diagrams or relations linking p and/or x to T;
C.3) diagrams giving exit thermodynamical variables vs reser
voir ones;
C.4) model applicability
In the following paragraph an answer to the aforesaid que
stion is given for each model taken into account. In the Par.3.3
a brief description of the considered models is reported.
(x) For example, T=f (h£,hg,p,v); %x~^2 (hf»hg»P»v) a r e intended as constitutive relationships,
(xx) For the definition of model applicability see paragraph 3.3.1
- 6 -
3.2.2 - Classification
The examined models are divided into the four main groups
mentioned above. Each model is located in Table 3.1 according
to the first author's name and to presentation date, with ref£
rence to literature available.
The models about maximum flowrate of a perfect fluid, both
throughout a De Laval nozzle and a cylindrical duct, are exa
mined in detail also owing to their use as reference point for
two phase fluid. Finally in Table 3.II for each model we give
the critical flowrate expression when it is written in expli
cit form by the authors.
We will note that Table 3.1 doesn't give an exact vision of
each model. Infact much more time would have been necessary to
analyze each theory in order to minimize the number of adopted
equations, to verify the consistency of the assumptions and to
understand its power. Rather the table (together with the Ta
ble 3.II) may be helpful to characterize each model and to
show the great differences among their analytical formulations.
[
1 1
1
!
f
1 i
!
!
i
1 1
1 f
1 I
5 i
MAM or TOI V I R S T AtrntoR
PERFECT CAS ( C l a e a i e a l Theory)
n V K T CAS (AaaiaaptLon i ) )
PEItPECT CAS (Aaauaetlon U »
INCOMPRESSIBLE LIQUID
•OMOClrtOUS BJOILIMItM HODCL
1ARST a t a l . 1977
SABITSKTT M 73
HOODT 19*5
HDOPT 19W
P A U U 1964
LETT 1965
a t o m « t a l . 1967
OGUAHAJtA 1969
OCASAVAJU 19*9
ABACRI 1975
ADACHI 1974
ADACEII 197*
HOODT 1975
CASTICLIA a t a l . 1979
mmmt at ai. »?s
«ALUS a t a l . 1 9 »
RAHSOH a t a l . 1979
ÉMMËt 1947
EAUXJDU 1943
STANDtt* a t a l . 19M
MOOT 1 H 9
KURT 1971
P'AICT 1971
ARDKM a t a l . 1976
RANSOM « t a l . 197S
RtJHR a t a l . 1970
HURT ( L / 0 >12) 1970
MEURT <UB >12) 1970
KUMXUIEL * t a l . 1971
ELÏRCEI1IEL a t a l . 1971
HALRES 197S
H i m 1973
RITA» a t a l . 197S
KMEGU 1976
MURE « t a l . 1976
ATOUT a t a l . 1977
TENTNER a t a l . I97S
HOUIRCU 1971
WIHTEM at a l . 1979
KATTO a t a l . 1974
«SOT a t a l . 1968
MEUNIER a t a l . 19M
CIOT a t a t . 1972
FLIRTA at a l . 197}
BADER a t a l . 1976
t T A M O 1977
m ^ m . 1 , , ,
HDOSL FOWOWÏlW
Al
i 1 1 i i
z
1
2
I
2
3
2
3
2
3
2
4
3
3
3
3
3
2
3
2
4
A2
j 1 j ! o
1 o
a
2
2
2
2
4
3
2
4
4
2
2
3
2
2
3
3
3
3
4
3
3
2
/ / 2
2
2
4
2
6
3
3
3
3
2
4
3
» 4
6
4
3
S
2
! 3
4
4
3
f
1
2
5
S
3
5
4
3
4
3
2
2
2
4
6
5
3
5
* 5
3
A3
a
I i
ï 1 / / / / / i
/ / i
/ / / / / / / / / / 3
/ CI
Î
1
/ / 1
1
1
<« / 1
1
/ / 3
1
4
« ' I
2
HR
3
A4
j 1
/
,
«
S
r
t
i
t
i
i
i
i
i
TES
TES
/. 1
TES
7
TES
TES
TES
1
1
1
1
TES
ASsïnrTICm (H ina.i.:le«.eal
Bl
|
j S
RO
NO
NO
HO
RO
NO
NO
NO
RO
RO
m
NO
HO
RO
NO
NO
RO
NO
NO
ru
RO
TES
TES
TES
/ 1
T U
/ /
TES
TES
TES
TES
TES
TES
TES
TES
TES
TES
t
TES
TES
TES
TES
RO
NO
m
RO
NO
RO
NO
TES
RO
NO
NO
NO
NO
TES
NO
TES
TES
TES
NO
TES
TES
RO
BI
S
1 ] 3
i NO
NO
RO
RO
RO
NO
NO
NO
RO
RO
RO
RO
RO
NO
RO
NO
NO
NO
RO
RO
rts
RO
RO
NO
NO
NO
RO
RO
RO
RO
RO
RO
HO
NO
RO
RO
RO
TE!
RO
NO
NO
NO
TM
NO
R3
i i
U II
II
II
II
II
T U
M
NO
T E S »
T E S »
RO
HO
HO
TES
T E S »
r t s " '
T E S »
T E S »
T E . "
T E . » '
RO
T E S " '
RO
II
II
TES
RO
TES
HO
T E S » )
NO
RO
T E S » '
T E S » '
HO
NO
HO
T U < "
TES
TES
TES
„,m HO
TES
TES
S4
3
•s
i
1 s RO
TES
HO
II
RO
NO
RO
RO
RO
RO
RO
RO
TES
II
RO
RO
II
NO
RO
TES
RO
TES
BJ
|
!
NO
NO
TES
HO
NO
TES
NO
RO
TES
TES
HO
RO
TES
// TES
TES
// HO
NO
TES
HO
TES
II
II
RO
RO
RO
NO
NO
TES
NO
NO
RO
HO
RO
TES
TES
TES
TES
TES
HO
TES
TES
NO
// // NO
RO
RO
RO
HO
TES
RO
HO
RO
HO
RO
TES
TES
TES
TES
TES
TES
TES
TES
NO
B6
! !
lli u
m
m
IT
IT
IT
IT
11,1?
I l l
m
IT
I I I
I »
I
I I
I I I
I
IT
IV
I I I
w
IV
I
I
I I
I I I
I I I
I I I .
I l l
IV
IV
IV
IV
IV
I I I
IV
IT
I I I
onrm
a I
| 3
l a
ù TES
' NO
RO
HO
TES
RO
NO
TES
TES
NO
NO
RO
HO
TES
TES
TES
TE»
RO
TES
NO
TES
NE
C2 es
I
3 a i
\ 1! 1! 3Z
8 s V.
ji 1| 1! II TES
TES
TBS
T U
TES
TES
HO
TES
TES
TES
TES
TES
TES
TES
TES
TES
II
TES
TES
TES
TES
RR
TES
HO
M
NO
HO
RO
NO
TES
TES
TES
RO
HO
tes
T U
T U
T U
RO
TES
RO
HO
T U
TES
C4
"J?
III
III
III
III
m
RO
HO
RO
TES
HO
RO
RO
RO
III
III
T U
III
NO
HO
RO
RO
III
NO
RO
TES
NO
TES
RO
TES
RR
HO
TES
TES
RO
RO
TES
TES
NR
NR
HR
RR
NR
TES
RR
TE*
TES
RO
TES
TES
TES
TES
RR
TES
TES
TES
TES
TES
TES
TES
RR
RR
RR
RR
TES
TES
RR
/// IN
RO
TES
T U
T U
RO
T U
HO
RO
HO
TES
HO
T U
HO
HO
HO
HO
RO
NO
T U
RR
III
III
III
HO
/// RO
HO
/// HO
HO
HO
HO
HO
T U
T U
T U
T U
TES
HO
RO
TES
HO
It « s
w S I
ïîl III
III
III
III
m
T U
RO
T U
T U
T U
RO
HO
T U
III
III
T U
III
T U
T U
T U
T U
l i t
"
1 a m
mi
IKI
mi
Ml
ni
mi
IUI
nv mi
IM
nu i-ni
tm
mi
/»/ /•/ ni
ni
im
nv mi
mi
l/l
VI
III
T U
l i t
T U
T U
III
T U
T U
T U
T U
T U
T U
T U
T U
T U
ru
T U
T U
T U
T U
EQUILIBRIUM ADIABATIC LIQUID-VAPOR FLOW
SOLVED IT THE CHARACTERISTIC DETERMINANT /
/
ARALTSIS OF ROCLEATtOH PROCESS /
RON EQUILURIim QUAUTT /
1MEVEHIBUE TRRIttODYKAMCS AHALTSIS f
•«COOLED LtQOlO TfflWOOl ORIFICU /
m lui 1191
mi
liai
iml
mi
nu mi
nu mi
mi
mi
lui
/Ml
IV»
ira
mi
mi
noi
nu
1X1
nu 100/
îos;
nu
1 M /
î a i
10»/
110/
mu " t J " U
3
1 i
i a f!
h 3.3.2.1
3.3.2.2
3.3.2.2
3.3.2.3
S.3.3.1
3.3.3.2
3.3.3.3
3.3.3.3
3.3.3.*
3.3.3.7
3.3.3.1
3.3.3.9
3.3.3.U
3.3.3.11
3.3.3.13
3.3.3.14
3.3.3.15
3.3.3.11
3.3.3.17
3.3.3.11
3.3.3.19
3.3.3.M
3.3.4.2
3.3.4.3
3.3.4.*
3.3.4.3
3.3.4.6
Î.J.4.7
3.3.4.»
3.3-4.9
3.3.4.11
3.3.4.12
3.3.4.12
9.3.4.13
3.3.4.1*
3.3.4.13
3.3.4.11
3.3.4.17
3.3.4.11
3.3.4.19
3.3.4.2C
3.3.4.22
3.1.4.23
3.3.4.21
3.3.3.1
3.3.5.2
3.3.5.3
3.3.3.3
3.3.5.4
3.3.3.5
3.3.5.»
3.3.5.7
/ - «a intéad " •» oaa"
/ / - t a l a t a * * %aat«caaaaer'1
/ / / - Ra intend "net intaraat lag l a th la analTata"
RR - «a lataad "act reported l a avai lable bibliography"
(1) » AeeLicabUtt* l a d e t l M d In ear. 3 .3 .1
'2> I t l a eeceeearr to e>>e ea ineut to tha calculat ion* the eathalpy evaluated a t tea height ot the piae-veaeal eoaaeetloe
Tafelt 3.1 - Suaaurr et the analysa*1 sod* I t .
9 -
NAME OF THE FIRST AUTHOR AND PUBLICATION DATE
MAIN MODEL CHARACTERISTIC
MAXIMUM FLOWRATE EXPRESSION (HHEN ACHIEVABLE)
PERFECT GAS CLASSICAL THEORY
- 1 I * • [ " . ' . ( ^ ) M ] PERFECT GAS (ASSUMPTION i))
max(p) p(pQ - p)
[»ho-h>] PERFECT GAS (ASSUMPTION ii))
m a x ( p ) p | 2 ( h Q - h )
INCOMPRESSIBLE LIQUID 2 P (P0 - P e)
IEMXENEOUS EQUJ_ LIBRIUM MODEL HEM T « max(p) ._
S o ( p o ) " S f ( p )
Vf •'
so{po)-sf(p)
— ^ 5 5 a w
LAHEY et al. 1977
FLUID-DYNAMIC APPROACH
-1
/ d v f v f g d \ [dp s f g dp
dv. v - d s . " f s . _l£ _ i i dp s f g dp^
)
J
f (p e )
BABITSKIY 1975
EQUILIBRIUM SCHEME
NOT EXPLICIT
MOODY 1965
SLIP EQUILIBRIUM MODEL (ENERGY MODEL)
r « max(p.k) <
nfg h - h , - (s - s , )
O f Sfg *" O f
k(s -s ) s - s , g o o f — s v,+ v s f g f s f f c fi
12 s - s r s -s„ o f + g o s f t k s f&
MOODY 1966
FIRST ANALYSIS OF THE KHOLE DEPRESSURIZATION PHENOMENON
DEFINITION AND ANALYSIS: Area average spe cific volume, momentum average specific volume, kinetic energy ave rage specific volume, velocity weighted spe cific volume
]!
U["<Ht*l3l-T''°MSl-HS)J j. _ A. w _ r. A. i 1/3 ds - O; k (VV
Table 3.II ./• continue
- 10 -
. / . Table 3 . II coi
OGASAKARA 1969
OGASAKARA 1969
ADACHI 1973
ADACHI 1974
ADACHI 1974
MOODY 1975 CASTIGLIA e t a l . 1979
TENTNER e t a l . 1978
KALLIS e t a l . 1978
RANSCM e t a l . 1978
BURNELL 1947
ZALOUDEK 1963
STARKMAN e t a l . 1964
MOODY 1969
HENRY e t a l . 1971
D'ARCY 1971
i i t inued
EIGENVALUE METHOD APPROACH TO CRITICAL TWO PHASE FLOW
AS ABOVE
TWO INDEPENDENT ENERGY EQUATIONS METHODS
AS ABOVE (FLOW FROM CYLINDRICAL DUCT)
AS ABOVE (FLOW FROM ORIFICES)
CONSISTENT SLIP MODEL
MAXIMUM ENTROPIC FLOW
METHOD OF CHARACTERISTICS TO SOLVE THE PROBLEM
ISENTROPIC STREAM TUBE MODEL
CHARACTERISTICS METHOD TO SOLVE THE PROBLEM (Four e q u a t i o n s )
SD1IB«»IRICAL CORRELATION
AS ABOVE
FROZEN COMPOSITION
PRESSURE PULSE MODEL
RELATED TO LOW TERMA NENCE TIME OF THE MIXTURE IN THE EXIT DUCT
PERTURBATION METHOD
NOT' EXPLICIT
AS ABOVE
NOT EXPLICIT
AS ABOVE
AS ABOVE
NOT EXPLICIT
AS MOODY 1 9 6 5 WITH CONDITIONS —- - 0 , I 5 - - 0 «fa a p
r •
r
r
r
r2
NOT EXPLICIT
• m a x ( p ) U yi . Y" ] i " l p . w. p , w i^n î .n fn n _
- 1
NOT EXPLICIT
- V2 - < 2 p f [ P u p - ( l - C ) p s a t J >
- C l { £ 2 > f ( P « p - P s a t > ï } / 2
i [ -, _ _y_( i * v • ( l - x ) v " ° *° o * ' 1
o g v o go L 1 0
fa ' ( l -o) ' 1
* «feal,kgtol ; x = x ° ;
VgUp ]s v f I 3 p j s
NOT EXPLICIT
NOT EXPLICIT
Table 3.II . / . continue
1* J
- 11 -
./. Table 3.II continued
ARDRON et al. 1976
RANSOM et al. 1978
HENRY et al. 1970
HENRY 1970
HENRY 1970
•
KLINGELBIEL et al. 1971
KLINGELBIEL et al. 1971
MALNES 197S
PORTER 197S
RIVARD et al. 1975
KROEGER 1976
BOURE et al. 1976
AVDEEV et al. 1977
TENTNER et al. 1978
MOESINGER 1978
WINTERS et al. 1979
UPPER BOUND FLOW
METHOD OF CHARACTERI STICS TO SOLVE THE PROBLEM (six equations)
LOW QUALITIES (x^ <0.02) MODEL
HIGH L/D RATIOS (L/D =12)
VERY HIGH L/D RATIOS (L/D >12)
ENTRAINED SEPARATED FLOW (ESF)
SEMI EMPIRICAL APPROACH
RELEASE OF DISSOLVED GASES
ANALYSIS PERFORMED OVER ALL MOLLIER DIAGRAM (from sub-cooled liquid to superheated vapor) TWO FIELD TWO PHASE MODEL
DRIFT FLUX APPROXIMATION
COMPLETE ONE DIMENSIO NAL TWO PHASE FLOW ANALYSIS ORIGINAL METHOD IN DETERMINING VAPOR FORMATION RATE CHARACTERISTICS ME THOD TO SOLVE THE PROBLEM DRIFT FLUX APPROXIMA TION ASSESSMENT LUMPED NON-EQUILI BRIUM MODEL; BUBBLE GROUT! 1 ANALYSIS
,»P l2fPcï 1 l x J P f
c |pgo goj IpJ X • K^ff^ff NOT EXPLICIT
r - FcHE
C ^
r 2 « -c
r 2 -c
lV Vfoj dp
x v ( 1
-f- - lv*>l '
- i
c
« ^ 1 ' dp
-1
c
NOT EXPLICIT
I" - LIKE MOODY 1965 KITH "k" OBTAINED FROM EXPERIMENT
NOT EXPLICIT
NOT EXPLICIT
NOT EXPLICIT
NOT EXPLICIT
NOT EXPLICIT
NOT EXPLICIT
NOT EXPLICIT
NOT EXPLICIT
NOT EXPLICIT
Table 3.II - Analysis of models with reference to: a) maximum flow rate analytical expression (when explicitly given);b) main model characteristic. All the models are referenced in preceding table.
- 13 -
3.3 SCHEMATIC DESCRIPTION OF EXAMINED MODELS
3.3.1 - Generality
In formulating any theory, a certain physical pattern, which
more or less reflects reality, is referred to.
In the present case , we take the situation shown in Fig.
3.1 as reference. In order to obtain flowrate it needs to con
sider at least:
a) initial thermodynamical conditions in the vessel: enthalpy
(H ) , mass (M ) and pressure (p );
b) geometrical data: length (L) and diameter (D) of the broken
pipe; position of the connection between broken pipe and pre£
sure vessel, with respect to initial liquid level (dimension
'b' in Fig. 1) .
Moreover, in the most general case, the theory should consi
der the following aspects:
- possible difference in pressure and temperature between liquid
and vapor phases while flowing in the duct, and consequently,
thermodynamical non-equilibrium;
- friction and heat exchange between each phase and the exte
rior;
- initial acceleration of the mixture as a function of breaking
time, initial thermodynamical conditions and rupture size;
- link between critical flowrate and pressure history in the vessel;
QO In this work we only examine models studying the flow of a two phase mix ture from a pressure vessel without internals. (The complex thermonuclear effects and heat exchange phenomena in a real vessel are not consi dered").
- 14 -
- mass, momentum and energy exchange between the two phases;
- multidimensional effects, expecially near geometrical discon
tinuities.
In the following we don't emphasize if the model matches or
not such aspects.
The flowrate and the exit thermodynamical variables have to
be the output of the calculation: they shall be expressed as a
function of one or more of the parameters given in the above
points a) and b) .
The description of the models will outline four aspects:
- model purpose;
- basic assumptions;
- essential equations;
- model applicability (we call a theory 'applicable' if it con
siders, at least, all the variables given in the above points
a) and b) as input, and if it shows, as result,a correlation
of the specific maximum mass flowrate (T) as a function of i-
(in which the term dvf/dp has been neglected) and then
* ., dx , r = f(x , — j — , p )
^ c dp rc (9)
dx The values of x and of —7—
c dp "I are made functions of p through an assumption (second assumption) with respect to the
mixture thermodynamic transformation:
lo = hf ( pc ) + x hfg ( pc) = c o n s t
from which
x =
h -h^ o f hf S
(10)
(11)
and
dx dp
xfC o dp
u
r dh. <*V hf g dp" " hf dp (12)
By substituting these last two equations in eq. 8 we obtain
r*=f(Pc).
The same author suggests the values to be used for the ratio
p /p as a function of L/D (see Fig.3.21) . In Fig.3.22 r vs h0 is
reported, pQ being the parameter.
It may be observed that the same expression for r given in
(8) can be obtained once again starting from an equation similar
to eq. (9) and following a different procedure
3.3.3.8 - LEVY 1965 /12/
This is a "lumped model": each phase is represented by a sin
gle mean velocity.
Besides assuming thermodynamical equilibrium, the basic as
sumptions are:
34 -
- no frictional or head losses;
- static pressure drop is the same for the two phases;
- the expansion is supposed to be isentropic.
Balance equations:
- mass conservation:
A. w- A w = D = —*- o —*- =
A pf 1-x A Mg x const (1)
momentum conservation equations:
- liquid phase
dp*p f w £ dw f - - ^ LTP
dz
vapor phase
1 2 wf dp+ — d(Ag pg wg) + j±- d(A£ p£ w£) •dz
(2 )
dz (3) CTP
From equa t ions (2) and ( 3 ) , one o b t a i n s :
r = -dp
dv m
- 1
-1 s dv m dpj s
where 2 (^ ^
x (1-x) V = V + V - - ^ -m g a f 1 - o
By s u b t r a c t i n g (2) from (3) i t r e s u l t s :
(4)
(5)
a(l-2cQ +a/(l-2a)2+ar(2pf/pg)(l-a)2-Hq(l-2al
X ~ 2 pf/pg (l-a)2 + d(l-2a) (6)
This is like imposing a constraint to the slip. For the calcu
lation of critical flowrate eq. (4) is directly employed, diffe
rentiating vm with respect to p. By taking into account the as
- 35 -
sumption:
i t r e s u l t s
ds = ds 3PJ X
+ 3s 3x
dx = 0 (7)
dvm
dp
3v, m 3p
9vm dsg dsf
3xJ |_ dp v dp_
In this expression the derivatives 3v m and
/Sfg
3v m_
3x
(8)
are obtained 3Pjx WA Jp
by differentiating eq. (5) e (6) respectively. After calculating
r it is possible to obtain the reservoir enthalpy from the eq.
(2) of section 3.3.3.5.
It should be noted that if we don't want the flow to be isen-
tropic, for instance an isenthalpic process, all the above equa
tions apply except for the replacement of entropy "s" with entalphy
"h" in equations (7) and (8).
Once neglected the frictional and head losses of the present
model, no further assumptions about slip are needed.
The functional link between "x" and "a" is obtained from a
momentum balance equation; it doesn't depend upon other adjoined
conditions that are necessary in the preceding theories (a pecu
liarity of the Levy model) .
In Fig.3.23 the maximum flowrate is given versus the exit qua
lity, the local static pressure is the parameter. In this figure
the dotted lines represent critical flowrate when assuming isen
thalpic process, the unbroken lines represent isentropic proces
ses. In Fig.5.24 the pressure is reported versus the critical
flowrate. With regard to these two figures, one may observe:
- the very little difference between isentropic and isenthalpic
processes ;
- 36 -
- that the solution exhibits a maximum in terms of flow quality (this
maximum in the pressure range reported stands in the range
0.01 <x < 0. 2) .
The author also gives a well known formula in order to calculate the
reservoir enthalpy when the exit quantities are found.
3.3.3.9 - tRUVER-MOULTON 1967
As the authors say, this is a unified theory of one dimensio
nal, isentropic, equilibrium, separated two phase flow.
In the initial part of the work, four conservation equations
are reported: they are related to mass, momentum, "mechanical"
energy and energy respectively. Only two of these equations are
used in the critical flowrate evaluation.
In the assumptions of constant flow area and steady state, n£
glecting gravity effects, the mass and mechanical energy conser
vation equations become:
- mass continuity
r = const. (1)
- mechanical energy conservation
r2 - - 2vH 3P
^ K E ^ (2)
-J s
The quantities v u and v„„ will be defined later. H Kb
Moreover the transformation equation is
ds = 0 (3)
at critical conditions. By solving equation (2), taking into ac
count equations (3) and (1), it results:
- 37 -
r= < -v.
2
H 3 v ^ - i [-HtriSl/iv^ia— S
(4)
From assumption (3) it still easily follows:
dx dp
1 Sfg
dsf dsfg' +x (5)
dp dp
The slip ratio value still has to be established in order to
obtain r.
Therefore the authors have written the following expressions
respectively for:
- velocity weighted specific volume:
1 f v = H TA I «
/A dA = (1-x) v,. + x v
g (6)
area average specific volume:
VA = X f P A JA
-1 (l-x)v£k + xVg
(1-x) k + x
- momentum average specific volume
VM = Jl f 2
pw dA = r A -A
XVg
TT t f l-x ) vf l+x(k-l)
kinetic energy average specific volume:
2 VkE =
1 f 3 AK - = - | p w dA I A } J IL
I <
- \ -£* + (1-x) vf l+x(k -1)
_ )
11 >
( 7 )
( 8 )
(9)
when k=l it follows that v. =v__ - vw = v. _. From the above espres A H M kE —
sions it may be clearly seen that the further degree of freedom
(k) appears when considering non-homogeneous flow instead of ho
mogeneous flow.
- 38 -
In order to find an appropriate value for k the authors prcj
ceed as follows: by differentiating the mechanical energy conse_r
vation equation (where eq. (3) has not yet been employed) with
respect to k, they obtain:
?AS r2 3(A vkE)2
cas i Kt _ ç 1 0 ) 3k 2T 3k avg
2 (x) It is easily shown that the function 3(v,p) /3k is negati
ve in the region l<k<(v /v-J™, it equals zero for k=(vg/vf)"^5 g 1/ & T
and it is greater than zero for k>(Vg/vf)'3 . it follows that the
entropy of the system increases with slip ratio, reaches a maxi
mum at k=(Vg/vf)/3 and decreases for k>(vg/v£) . Therefore as
sumption (3) implies:
f v
k = g Vf
%
/13/ /ll/
like Moody and Zivi
Therefore equations (4), (5), and (11) solve the problem.
In Fig. 3.25 r vs h is shown; another interesting result is
shown in Fig.3.26, where v., v , v , v are plotted as a function A H M kt
of k, when quality and pressure are assigned. It is shown that v., doesn't depend from k (obvious), v. continuously decreases M r Aj . with k, v has a minimum for k=(vg/v£)
2 and v, has a minimum 1/5 for k=(vg/v£)
2 2 (x) The authors show that in the above assumptions 3/3k(Av,p) = 3/3k(v )
- 3*1
5.3.5.10 - OGASAWARA 1969 772/
Another theoretical approach is used by Ogasawara in order to
determine critical flowrate when thermodynamical equilibrium be;t
ween phases is present.
In this model the criticality condition is obtained by em
ploying separate momentum equilibrium in order to adequately des_
cribe the momentum between the two phases.
With the assumption of axial steady flow where radial varia
tion of the thermodynamical quantities is neglected, the below
equations result:
- mass conservation:
_d_ dz
pa w + Pf (1 - a) wf =0 (1)
- momentum conservation in vapor phase:
dz p aw + ap g g
] = -= -F. -F fg g
(2)
- momentum conservation in liquid phase
à_ dz
p£(l - a) wf + (1 -a)p = F -F fg f
(3)
energy conservation:
d_ dz pgawg(hg + i) + pf(l-a) W T M = "R C4
The author points out that the momentum equations are written
separately so that the force F acting between the phases may
not be cancelled as internal force, as in other models.
Indeed the same separation is not necessary for energy equation
because of the assumption of saturated flow.
- 40 -
Rearranging the set (1) t- (4) appropriately, it may be written
as :
A(E)
dw.
dz
dp_ dz
dk dz
=
R
-F + F fg £
F . - F fg g
(5)
Since the right side member includes only finite terms and d w £ dp dk
the left side terms f—;—, -r> T~) are assumed as divergent in a dz dz dz
critical situation, it follows that the determinant A(E) must be
zero, that is the criticality condition is
A(E) = 0 (6)
This leads to a cubic equation in w , the roots of which indi
viduate the critical velocities. Since there is more than one
root, the author takes the root that better approximates the ex
perimental results as the real root. The slip ratio "k" is itéra
tively calculated when p and x are given. Therefore these two ' c c
quantities are necessary as input to the calculations. By inter
polating the theoretical results the author also gives a mathema_ % — — V tical expression of k , vs the ratio I p_(p )/p (p ) I 2,
•— r c g c —'
The results are given in Fig. 3.27 in which r vs x is repor-,
ted, p is a parameter; in Fig.3.28 r vs p , adopting the parame
ter x , is shown. In the same work Ogasawara presents another
theory obtained by substituting the energy equation (1) with an
entropy conservation equation. He concludes that critical condi
tions with energy conservation give a more accurate solution
than the ones with entropy conservation equation. In the latter case (en
- 41 -
tropy conservation) only few values of T are presented in a ta
ble. The values of r so obtained are slightly inferior to the
ones calculated when total energy is conserved (obviously p and
x are the same in the two cases). c
3.3.3.11 OGASAWARA 1969 774/
This other model formulated by Ogasawara is valid for flow
through orifices in which the flowing fluid friction and heat ex
change may obviously be neglected; moreover the fluid velocity
upstream the orifice is assumed as zero. The other explicit
assumption consists in neglected kinetic energy in comparison
with enthalpy at the exit. The equations are as follows:
- mass equation:
dz _ xpf+k(l-x)p J (1)
momentum equation:
-^ I r(l-x+kx) wf + p = F (2)
- energy equation
dz
v 2 2 2 — K W W
x C - y ^ + h ^ + C l - x H y + h£) = R (3 )
By integrating these equations one obtains f h. h. -h.
, fgo. fo fc x = x (r-*-) + r r— c oh, h -h
fgc go gc p = p - k p p_(l-x+kx)w_ /M *c o c gc fc c c c fc
(4)
(5)
where
42 -
M = xpr + k(l - x) p £ g
These are two equations containing four unkowns (i.e. p , x ,
Two further equations are then necessary to achieve solution.
One is obtained by the "eigenvalue method" (see also the prece
ding paragraph) setting the determinant of equations (1), (2),
(3) as equal to zero (following the same process of the prece
ding paragraph). The other comes from the FAUSKE's result, i.e. V2
k = (v /v ) g (6)
Substituting determinantal equation (determinantal equation is
the equation (6) of the preceding paragraph) and equation (6) in equa
tions (4) and (5) the problem is resolved. Unlike the preceding
model, in this theory Ogasawara gives r vs p (see Fig. 3.29);
with x as parameter. In Fig. 3.30 p /p vs x is shown; p is the o c " *c *o o o
parameter. The following result as:
- the smaller the fluid compressibility the larger the pressure
ratio becomes (Fig. 3.30); /123 13/
- since it has been experimentally confirmed ' that p /p
becomes smaller in the case of small reservoir quality, owing
to thermodynamical metastability, this model isn't worth using
at small qualities;
- the results differ from Moody's for about 201.
In the same report the author makes a very interesting analy_
sis (comparing theoretical and experimental results) regarding
two phase flow through orifices;in particular , he observes a drastic r£
duction in specific flowrate when orifice diameter increases
(all other variables being constant).
43 -
/HT,/ 3.3.3.12 MALNKS 1977
This theory is taken into consideration only to point out the
quantitative and qualitative differences arising in the two pha
se mixture maximum velocity calculation when the following two
assumptions are separately applied (the basic equations are the
same in the two cases):
a) homogeneous frozen flow (k=l, ip=0) ;
b) homogeneous equilibrium flow (k-1, sufficient to mantain
thermal equilibrium or equal temperatures of the two phases).
The steady state flow equations are:
- vapor-mass conservation:
T - (ap„ w ) - ty = 0 9z g g (1)
liquid-mass conservation:
9 9z
(1-a) pf wf i|> = 0 (2)
mixture momentum balance:
3p + 3
9z 9z 2 2
a p w + (1-a) p. w_ g g f f
= 0 (3)
In assumption a) we obtain:
w HF a
< 9p ^
9p 1-a
9p£
~3p~ ) s
In assumption b) we obtain:
a 2 1 w = - < HE p p
f 9 p * l 9p
t " J
f l - a '
T P f
' 3 P f
9p +
T
a
LV
f*«l 9T
I J
+ l - a ' 3 p f '
9T P.
f ï 9T 9p
I )
( ^2 3T
3P apg CPg+(1"°° pf CPf
(4)
(5)
- 44 -
The large differences between these two extremes may be seen
in Fig. 3.31.
The above equations are part of a more general non equilibrium
theory presented by the same author and to be considered later
in this work.
IM 3.3.3.13 - ADACHI 1973
The formulation of this theory is strongly different from th£
se shown so far.
The basic assumptions are the usual ones (apart from the im
plicit assumptions we have referred to in Par. 3.3.1). We shall
repeat them only for convenience:
a) one dimensional two-phase flow (separated flow);
b) steady flow;
c) adiabatic flow;
d) state change follows the saturation curve.
The author takes into consideration two energy balance equa
tions: the one related to the "flowing fluid", the other related
to the "existing fluid"; in the case of non homogeneous mixture
he shows that the two espressions are independent.
Let us consider the fluid flowing in a duct between sec
tions (1) (upstream) and (2) (downstream). The vapor and liquid
(if slip exists) which have simultaneously passed through control
surface (1) cannot reach section (2) at the same time. Therefore
there must be mass and energy exchange between the two phases;
as long as the "existing fluid" is considered the flow between
sections (1) and (2) is not adiabatic.
Viceversa, for the above assumption c) the "flowing fluid" e-
nergy equation is of adiabatic type.
45 -
Equations adopted:
mass conservation:
TA = T A = A c c
ap w + (1-a) p. w. g g £ £
= const. (1)
- energy of the flowing fluid conservation:
2 2 2
x w + (1-x) w.
L g i
+ d xh + (1-x) h^ g J f
= 0 (2)
- energy of the existing fluid conservation:
2 2 6w + (1-8) w£
+ (1-Ç ) JL+idL dp = 0 (3) w p p- x v 7
g £
where:
a p 0 = g
ap + (1-a) p g f x + (1-x) k
(4)
From the state equations (section 3.3.3.1) and from the assump
tions :
ds = 0
dh = 0
we obtain respectively:
d x ] s - •
and
dx
where Ç is defined by:
x ds + (1-x) ds,. g £
fg (1-ÇT)
x dh + (1-x) dh^ g . L r h. 4 fg
dp = (1-Ç_) dp « dp s ^T v Krev
(5)
(6)
(7)
(8)
(9)
- 46
and dp, = Krr, dp = dp. *h T r i r r
By s e t t i n g dx=dxj +dx_| and so lv ing in the unkn
(10)
awn x , i t
r e s u l t s :
x = x -o
p _ _ p [_x ds +(l-x)ds f J (1-&J.) r
g
J, ' fg J,
x dh + (l-x) dh^ g f
fg
( * )
(11)
The following procedure i s shown in the fol lowing flow diagram:
input po'VÇT'Çw
rp\ppi
| x from (11) 1
k for the new state ( ini t ia l guess)
w„ and W£ from (2)
| kj from (3) and (4)| (iainteraction index, in this case)
-GE£Q NO
r from (1) and a from (1) of s e c . 3 .3 .3 .4
output p »x
e»wge»wf e , a e» ' ' c ^ e
YES •<£^C>" NO TËNDI
S«-S o->f . (x) Note that i f ^=0 i t follows x= i . e . isentropic flow Sfg
47 -
Note :
- the iterative calculation of k and its dependence from £ and r w
- the absence of criticality assumptions in the calculation of
the exit thermodynamic quantities.
In Fig. 3.32 the variables r, w , w_ are shown as function of g f
fluid instantaneous pressure. The left side of the curve (with
respect to their maximum) can be achieved only in the case of
expansion occurring in a constant area duct (this means that "supe£
critical" flow cannot be observed in a constant area duct).
In Fig. 3.33 we report r vs p: x is the parameter for the a£ 2
signed initial value p = 60 kg/cm .
In Fig. 3.34, finally, k vs p is shown (the initial value is 2
still p = 60 kg/cm ).
It is still to be noted that no idea is given about the calcu
lation of two quantities K and Ç (we remind that they repre-W X
sent the losses).
On the basis of this formulation the authors present two other
theories applicable to cylindrical ducts and orifices respectively.
3.3.3.14 - ADACHI 1974 (cylindrical duct) /6/
The equations and the solution method are those reported in
the preceding paragraph. The author admits that in the case of
a cylindrical duct the friction losses have the major weight in
the total loss evaluation, whence it results:
T w
Moreover this hypothesis requires specific flowrate to be
constant.
- 48 -
The integral pressure drop and the integral frictional pressu
re drop respectively can be defined as follows:
fP C - -I dp = (p -p) (1)
i O
'Pp çF - - j
P dpF - (po-p)ç (2) *Po
The solution of the equations of the preceding paragraph is
shown in the following figures;
- Fig. 3.35 represents ç„ vs ç with r as parameter;
- Fig. 3.36 represents r vs p with parameter E, •
In both diagrams the pressure and the vapour quality of the
reservoir are supposed to be constant.
We observe that:
-. the maximum in Fig. 3.35 represents the critical flow: in fact
assuming one is not in a maximum point (point 2 of Fig..3.37)
at an exit pressure equal to p~, lowering this value the flow
rate increases and the representative point shifts to a curve
with higher flowrate. The transformation line on this diagram
will have a positive slope as it has to increase with ç . r
The right-hand points of the maximum values are not real ones
since the frictional pressure drop would decrease along the
flow;
-. the straight line 9 of Fig. 3.35 represents a limit case (r=0)
since the losses are only due to friction and not to fluid a£
celeration;
-. the dotted line in Fig. 3.36 corresponds to the dotted line
in Fig. 3.35: it represents the relationship between the cri
tical pressure and the critical weight velocity for given re
servoir quality and pressure;
49 -
-. the change of the flow in the constant area channel (r=const.)
yields a straight line parallel to the horizontal axis (in
Fig. 3.36, i.e., the line a-c-d represents the flow behaviour).
The interaction of lines c-d with the lines having constant Ç
means that Ç varies along the channel axis;
-. ç must be deducted separately if we want to obtain the criti
cal flowrate; from value of ç we may obtain both ç=p -p and
r from Fig. 3.35.
In the same work ADACHI also presents other curves showing the
relationship between the channel pressure and other parameters
(ct,w ,w ,k,x).
A semiempirical relationship is finally drawn between the pi
pe current adscissa and the pressure. The comparison with the ex
perimental results is also performed.
in 3.3.3.15 - ADACHI 1974 (orifice)
In this work information about the values to be given to the
efflux coefficient C area provided when the efflux from orifi-IM ces is studied and when the theory presented in (Par.3.3.2.13)
may be applied.
The author observes that:
- the coefficient is considered so as to include the effects of contrac
tion, friction, compressibility, and thermal non equilibrium of
the two phase discharge flow through the break. Fig. 3.38 (C
vs x ) and Fig. 3.39 (C vs p ) result from an interpolation of
experimental values.
However, when the orifice is attached to the pressure vessel,
the functional relationships can be written as (for a given
break geometry):
50
S = £(D>VV (1)
It may be observed that:
-. the relationship between the discharge coefficient C_ and the
high pressure reservoir tank quality x can be separated into
two regions:
a) region of x j*0 where x has little effect and o o
bl region of x =0 where x has a significant effect;
-. in the region of x 9*0, C is almost independent from p and
x and it can be determined from the orifice diameter. Cn is o D
larger for smaller diameter values; for 25<D<70 mm C is in
the range of approximately 0.8-0.52. The reason may be due to
the fact that the critical flowrate decreases because of the
contraction immediately downstream from the orifice and, in
addition, the extreme supercooling (50°C) phenomenon of the
vapor phase which occurs at the orifice; -. in the region x =0, C„ sharply increases as x decreases and & o D o
it reaches 1.5 from the 25 mm orifice. However for the same
value of x ,C becomes larger for smaller D. For higher p ,
C slightly decreases.
The reason for C exceeding 1 may be as follows: because of
the lack of nuclei, vaporization cannot follow the fast depressu
rization at the orifice and thus the discharge flow behaves more
like single phase fluid. The theoretical maximum value of C for
x =0 is given in Fig. 3.39 which is obtained from o
1 2Pf CP0-Pe) CD - °-« f(po,0) <2'
f(p ,0) being equation (1).
- 51 -
3.3.3.16 MOODY 1975 79/
The author recognizes that Fauske,Levy and his own preceding
models are based on either momentum or energy conservation but not
upon both and therefore they are not consistent. However he says
that the assumptions of the above models do not necessarily des
cribe the physical behaviour. On the contrary this study uses all
the conservation laws.
Besides the thermodynamic equation assumption, the author im
poses:
- monodimensional flow;
- any dissipative loss equal to zero;
- constant pressure in any section.
The equations are:
- continuity equation
d(TA) + d
- momentum equation
AV(- a v g
vf J CD
r2 Av m
+ d vrA + Adp = 0 (2)
- energy equation
fA h + d f A£ Ag
V — * W + — - h o L [ Vf £ Va 8-U g
- AVdp = 0 (3)
Second thermodynamic law:
rAs + d [ A £ Ag
V —i- s,-+ —s- s _lv f f vg 8J
50 (4)
in which
m x v + (1-x) kv,.
g t x +
1-x (5)
- 52 -
The critical conditions are expressed as follows:
V = 0 (6)
dV dp
dr dp
= 0 (7)
= 0 (8)
(i.e. perturbation velocity and its derivative with respect to
pressure are supposed as equal to zero) .
By substituting equations (6) and (7) in the set (1), (2),
(3), (4), we obtain a new system of three equations in the four
unknowns A, x, r, k if p is considered an independent variable. /7 ?/ 128/
The author says Ogasawara and Giot showed that mo
mentum or energy conservation for either phase is sufficient for
one more independent equation; however an approach which invol
ves entropy production is equivalent and it is employed in this
study.
The following state equations for either phase are used:
- Gibbs equation
T ds = dh - v dp (9)
- Clapeyron equation
T sfg ' hfg (10>
- state equation
h = hf + x h f p (11)
(x) For the author these conditions are representative of the physical fact consisting in wave blockage at the critical section.
- 53 -
Equation (9), (10) and (11) together with equations (1), (2)
and (3) produce a new relationship among dependent variables,
namely:
ds (p,x) = — mr + m„ _ fg gf
w - w _ _g L
2
; i
dp = £(p,x,k,r) (12)
where the terms m and mpf represent the vaporization and con
densation flowrates the expression of which is given in App. 3.3
Equations (1), (2), (3) with conditions (6), (7) and (8)
constitute a closed system. In order to obtain a solution it is
further necessary to find consistent values for k.
The author assumes
k * 1 (13)
and that k is maximum consistently with the condition ds=0.
The results are shown in Figs 3.40 and 3.41.
One may observe that:
- these results imply the thermodynamic quantities at the exit
knowledge ;
- the flowrate variation is too great with respect to k which
hasn't been fixed clearly and moreover it is difficult to ob
tain from blowdown experiments.
3.3.3.17 CASTIGLIA-OLIVERI-VELLA 1979 775/
This model is based upon the assumptions of a onedimensional
stationary motion and thermodynamic equilibrium.
The expansion is supposed to take place in a duct with fric
tion and adiabatic wall.
The balance equations are written as following:
- 54
- mass continuity
Ar = A
- momentum continuity
aw wf
_g_ + (i_a) — g f
= cons t (1)
dp _ r2 d dz dz
2 n ,2 V + — V-
a g (1-a) f + F (2)
- energy continuity
h =h +x h + — o g fg 2
3 n ^ 3
x ..(l-x) —;r v +- — v_ 2 g f1 .2 £
a (1-a) (3)
In order to evaluate ?, the assumption of maximum entropie
flowrate is used; according to such assumption , when reservoir
enthalpy and flowrate are assigned, the effect of irreversible
phenomena is such that entropie flowrate is maximum in each point
of the expansion and this maximum is obtained with respect to
void fraction, which is assumed to be the only restrained quant^
ty.
Analytically we have:
3s 3a
= 0 (4)
Taking the classical expression for "a" into account we ob
tain:
r k =
v ïV3
_g_ (5)
which is nothing more than the result obtained with different
assumptions by Moody and Zivi
It is to be observed that in this case equation (5) is valid
throughout the transformation, whereas in Moody's cases it was
- 55
obtained as a criticality condition and therefore valid only at
the critical section.
By substituting (5) in (3) and solving for x we obtain:
vf
x(p)=-
I , 2 Z.nys , 2 3.f]l/3
2/5 2/5 vg " vf
(6)
where q and q are functions of static pressure and of flow
rate (see App. 3.4).
By substituting equation (6) in the entropy espression we ob
tain mixture entropy through all the transformation as a func
tion of "p" only.
In Fig. 3.42 we report, in a p/s plane, the transformation
trends related to different r values. The line below point c) is
not admissible, since the fluid system is isolated (the entropy
cannot decrease); it follows that the other criticality condition
results: *
d S = 0 (7) dp
Moreover, in order to obtain the critical point only it is a.L
so necessary that:
d 2 s < 0 (8) dp2
From eq. (3) and taking into account conditions (4) and (7),
the diagrams of figures 3.43 and 3.44 are obtained.
It can be observed that in the region of very low qualities s£
me curves (having h as parameter) show a maximum: this means
that when reservoir enthalpy and flowrate are fixed, two points
of validity of equation (7) are present. The authors say that on
ly the point with greater quality satisfies both conditions (7)
- 56 -
and (8). Therefore in Fig. 3.43 only the region on the right of
the line connecting the maximum points is the real region. /13/
The analogy of this model with Moody's is evident, regar
ding numerical results too; in particular it refers to the pro-/9/
cedure later adopted by Moody , which consists in assuming
equation (5) as valid throughout the transformation.
At the conclusion of their work, the authors observe that
this model too allows to connect inlet thermodynamic condi
tions and flowrate to the pipe physical and geometrical charac
teristics: it is sufficient to integrate equation (2).
3.3.3.18 - TENTNER-WEISMANN 1978/10/
This model distinguishes itself from those presented above
especially for the solution method adopted, that is the charac
teristic method. Therefore the variable time appears in its e-
quations;it follows that the critical flowrate should be deter
mined, as a function of time, until it reaches the critical va
lue.
The prediction of critical flow is based upon the magnitude
of the characteristic slopes. Critical flow is predicted when
the smallest characteristic slope becomes equal to zero. This
corresponds to the physical fact that a pressure pulse cannot
propagate upstream.
Balance equations:
- mass conservation:
^[apg+(l-oOp£ >^-[apgkw£+(l-a)p£wf J =0 (1)
- momentum conservation:
- 57 -
3 r~ , ,-. •> ~î 3 r i 2 2 ., . 2-7 8p Tw . . . -Lap kwf+(l-a)pfw,J+-Lapkwf + Q - a)pfw J + - f = - - (2) 9 t L _ - g — £ ' f f - l 3z L.-"g" "f
- energy conservation:
L«P_B • (l-a)PfBf J + ^ L « P w£kE + (1-oOp f J -£-£ (3)
9 r ïïLVg
where i 2 2
K W f E = h + — -g g 2
Wf Ef " h£ + T
These are three equations in the three unknowns (a, p, w ) ,
since the other quantity (k) is found separately.
The authors note that a term accounting for form losses may
be added to the right side of equation (2) and that this term
will not change the nature of the characteristic directions be
cause it doesn't contain derivatives of w and a. For the same
reason terms T and q too don't influence the characteristic w nw slopes.
Besides the assumption of thermodynamical equilibrium and
the physically justifiable hypothesis:
3p£
~3P" = 0 (4)
the authors assume that"the slip ratio variation is low e-
nough to allow terms containing partial derivatives of k with
respect to p, w and a to be neglected"; i.e.
58 -
As we have already said, in order to solve the system it is
necessary to give the functional form of the slip ratio as input.
For void fraction below %0.7 the following Hughmark correla
t i o n i s used:
l f p f ï ( l - Z ">
T-Hifjhr-J (6)
where Z is a non linear function of an assigned parameter w
(see App. 3.5).
For void fraction greater than 0.7, values of k are obtain
ed using the classical expression:
k = H l-o ^ pf
l 1-x (7)
g ft ?
At mass velocities above ^2«10 lb /hr ft homogeneous flow m
is assumed (k=l).
Moreover, in their calculation, the authors use the Moody ex-
k=(vg/vff3
pression of the slip too .With regard to this last
assumption they conclude that Moody slip values are too high on
two counts :
- first,they are above experimental results;
- second,they cause a mathematically ill-posed problem.
In Fig. 3.45 r vs x is reported in two cases: e
a) i n i t i a l model-
b) k from Moody's c o r r e l a t i o n .
In order to so lve the problem complete ly terms T and q (*) w w
in equa t ions (2) and (3) need to be eva lua ted . (x) For this reason we have written in Table 3.1 that three constitutive re
lationships are necessary to solve the model: 1) related to k (given); (2) related to T (not given); 3) related to q (not given).
- 59 -
In the second part of the same report a simple approach to
non equilibrium problems is presented: it will be described in
the following paragraphs.
/76/ 3.3.3.19 - WALLIS-RICHTER 1978
The authors conclude that this model doesn't represent the de_
tails of choking realistically (or better, more realistically
than other models), but it may be considered as a certain limit;
its purpose may rather be to help in providing standards for com
parison with other theories and with actual phenomena and it may
be considered as a starting point for the development of more
elaborate theories.
The basis of this model can be understood from Fig. 3.46.
Saturated water is assumed at the entrance of a tube; a cer
tain amount of pressure decrease will cause the first evapora
tion (first stream tube); a further Ap will cause a further eva
poration (second stream tube); this second Ap also causes expan
sion of the vapor in the first stream tube; and so on.
In Fig.3.46bis the process in a Mollier diagram is shown.
After changing phase the espansion in each stream tube conti
nues independently from the other stream tubes.
If the water is not saturated at the inlet, the authors sug
gest that the conditions up to the saturation curve may be ob
tained from Bernoulli's formula giving for the initial velocity
to be considered as input: 1 2
(1) w = o
p -p ^ - o sat
The basic assumptions are:
- 60 -
- isentropic flow and liquid in equilibrium with the vapor in
each stream tube;
- pressure uniform in each flow section;
- the velocity components perpendicular to the main flow direc
tion negligible.
The equations for each streamtube are:
- mass balance
Y- , = Y. + y. l-l l i
energy balance
f w. ">
l-l| fi-1 2 = Y.
)
entropy conservation
2 i w J v + 41
fi 2 I
2 <, w .
h +-Si gi 2
(2)
(3)
Y. .. s_. n = Y. s_. + y. s . l-l fi-1 l fi 7i gi (4)
Besides the well known symbols we have
i-1 liquid flowrate from which the stream tube "i" follows
(see eq (2)) ;
Y. normalized liquid mass flowrate in stream tube "i"; l
y. fraction of total mass flowrate in i-th stream tube;
i,n number of Ap steps.
The authors also show that by considering equation (2), (3)
and (4) and thermodynamic identity:
hjr = T sr (5)
one obtains Bernoulli's equation applied to the liquid phase.
From equations (2), (3), (4) it follows
r = X—-±— 1=1 p. W.
i,n i,n
n pfn Wn
-1 (6)
- 61 -
3.3.3.20 - RANSOM-TRAPP 1978 778/
This model is only a part of the recently developed code
RELAP 5 /M0D"0M.
In order to obtain chocking, the flow is described by the ove
rail mass continuity equation, by two momentum equations and by
a mixture energy equation as follows:
- mixture mass equation
-k |_a Y ( 1"a ) pf|+ ~t La pg V e 1-0 0 pf wfJ= ° (1)
- vapor momentum equation
aw aw
g I 9 z S 9 x a p + (l-a)g+caCl-o)p
3w 3w 3w. 3w_ _g+w —S L w _ J at fak az g ax
= o (2 )
- l i q u i d momentum equa t ion
(1-cOP, aw aw
+w
at fax +(l-a)|?+co(l-a)p
3wf 3wf 3w 3w +W - * - W i r — *
at sax at fax = o (3 )
- mixture energy equa t ion
a at
a pg s g + ( 1 " a ) pf s f Hi* pg sg wg + ( 1" a ) pf sf wf = 0 (4)
This system may easily be written in terms of four dependent
variables a,p,w ,w . It appears as
A(U) au at
+ B(U) au 3x
+ C(U) = 0 (5)
where A(U), B(U), C(U) are square matrix (4*4 in this case) and
[u~| is the vector of the dependent variables.
By defining the characteristic directions of the system as
the roots A. of the characteristic polinomial: l
AX - B | = 0 (6)
62 -
where
Pi,n= 1-x.
pfn
(7)
gn
and w. i,n
1 7 2|h.-h. + wT _ { i i,nj i_
(8)
In the above expressions M. is the generical thermodynami-
cal quantity risen in the i-th Ap step and evaluated in the n-th
Ap step.
Expression (6) for r (relating to a unit cross sectional area)
is the reciprocal of the sum of the areas of all stream tubes per
unit normalized flow. It results function only of local pressure.
The criticality condition is expressed as:
dr dp
= 0 (9)
The values of r vs Ap (for given p and T ) are shown in r *o o
Fig. 3.47; in Fig. 3.48 r vs x (with p as a parameter) is pre
sented.
With regard to«this model it may be further observed that:
- it bypasses the slip evaluation difficulties found in other mo
dels ;
- it admits differences in velocity along the flow radius, so
that, within certain limits, it may be considered a bidimensio
nal theory;
- it doesn't consider momentum transfer between the streamtubes;
- by adopting very small Ap steps a continuous expansion may be
analytically simulated, even if the decrease of step size may
cause certain instabilities in calculation depending (according
to the authors) upon accuracy of steam tables.
- 63 -
It has been shown that the mathematical condition for cho
ked flow is
A. £ 0 for all j * n (7)
where n is the nunber of equations. This corresponds to the
physical fact that reduction in downstream pressure ceases to r£
suit in increased flowrate.
With reference to this model the following observations can
be made:
- the momentum equation includes the force terms due to relative
motion between the phases;
- the product ca (l-a)p is the coefficient of virtual mass: "c"
has to be fixed each time, although it has a theoretical value
of 0.5 for dispersed flow and for separated flow the value may
approach to zero;
- the energy equation is written in terms of entropy which is
constant for adiabatic equilibrium flow;
- the term C(U) is not given by the authors in equations (1),
(2), (3), (4) since it is assumed that it doesn't affect the
characteristic equation (6) because it doesn't contain deriva
tives of the dependent variables;
- the term C(U) represents the constitutive relations (not given
here): these relations include interphase momentum and mass
transfer, wall heat transfer and friction.
In this model thermodynamic equilibrium is assumed, however
the authors have presented also a similar model valid when
frozen flow is assumed: it is shown in following sections.
The eigenvalues resulting from equation (6) and the mathema
tical discussion of equation (6) are given in App. 3.6.
Fig. 3.49 shows the trend of equilibrium Mach number versus
- 64 -
vapor fraction. The choked conditions are defined as the inter
sections of the lines having c= const, with Mach number unit li*-
ne.
In the bibliography available no other interesting results for
this work are given.
- 65 -
3.3.4 - Non equilibrium models
Non equilibrium models include equilibrium models as a parti
cular case. Non equilibrium effects derive from the physical
fact that two-phase depressurization velocity may be greater
than thermal exchange velocity between liquid and vapor. When
the mixture at the inlet of the broken pipe is subcooled or satura
ted water, non equilibrium consists of a delay in vaporization
and of superheating when there is vapor at the inlet. /26,124,etc/
According to different experimental works the time during which the mixture remains in a metastable state is about
/9/ one millisecond. Starting from such a value, Moody calculated that non equilibrium degree for pipe length greater than 12
centimeters is negligible. /31/
According to other researchers , instead, the non equili
brium value depends on the L/D ratio: i.e. non equilibrium pheno
mena must be considered for L/D£3*-12.
However also in long channels it probably plays an important
role on thermodynamical conditions at the exit section.
Besides, all researchers seem to agree that non thermal equi
librium can appreciably increase both critical flowrate and pro
pagation velocity of rarefaction wave normally generated at the
ruptured section.
Many theories treating non equilibrium have been developed:
some of them do not help more than the empirical formulas to
quickly calculate maximum flow of initially subcooled water; o-
thers make use of very sophisticated models.
Most of these theories adopt empirically determined coeffi
cients: the latter are the number of vaporization nuclei per
- 66 -
unit mass of flowing mixture, a number relating non equilibrium
quantities to equilibrium ones, a number that is needed to take
into account a certain assumed flow contraction, experimentally
observed in single phase flow from orifices.
The feature which is common to all these numbers is that they
are obtained by matching experimental data to theoretical results.
As already said (par. 3.2), we have distinguished these
theories in two groups:
- frozen theories (which admit no variation in mixture composi
tion while flowing);
- non homogeneous non equilibrium theories (in which no assump
tion is given referring to the link between pressure and tem-(*) perature of the two phases) .
With regard to non equilibrium models, the greatest part has
been developed in the last years; in this report we want only to
show the qualitative and quantitative differences existing bet
ween mixture velocities in a homogeneous equilibrium theory and
in I
a frozen theory (See for example Fig.3.50)
3 . 3 . 4 . 1 - "Frozen" t h e o r i e s
The assumptions g e n e r a l l y common to these models a r e :
- the v e l o c i t y r a t i o between the two phases i s g iven;
- no hea t or mass t r a n s f e r t a k e s p l ace between the p h a s e s : thus
the q u a l i t y remains c o n s t a n t throughout the expansion ( t h i s
assumption c h a r a c t e r i z e s "frozen f low") ;
- the vapor expands accord ing to an ass igned law: for example
(X) We actually intend a theory of non equilibrium type when at least one thermodynamical quantity doesn't follow saturation l ine .
- 67 -
according to gasdynamic principles.
All vapor and liquid transformations are then assumed as
independent from one another, as far as the exit section at
least.
3.3.4.2 - BURNELL 1947 78/
It is fairly well noted that maximum flowrate consistently
increases when decreasing L/D ratio and an even greater flowrate takes pla_
ce when subcooled conditions are present just upstream the break,
as in the case of orifices.
The experimental results in this case significantly exceed
the prediction of most of the models presented so far.
In the fourtie s many works have been performed with regard
to this situation and are still valid (see for example ' ' ').
Burnell's work is one of the studies more frequently mentio
ned in bibliography.
He relates the non equilibrium in two-phase flow to the sur
face tension of the liquid droplets that delays the vapor bubble
formation.
In his model the critical flowrate is proportional to the
square root of (p -p 1: in particular he obtains: n VIup *sat r
r * = < 2 P f L P u p - C i - c ) P s a t ] > / 2 CD
The coefficient C is (Fig. 3.51) a function of saturation
pressure. As it may easily be seen from equation (1) the coeffi_
cient "C" decreases in a fictitious manner the value of the satu
ration pressure itself.
Physically speaking this means a certain liquid superheating
- 68 -
is to be considered.
From the above formula a diagram showing r vs p is quickly
obtained when p is known. up
3.3.4.3 - ZALOUDEK 1963
A similar approach has been employed by ZALOUDEK who analy
zed an initially subcooled two-phase flow from short pipes. He
observed two possibilities of choking phenomena:
1) upstream choking with respect to exit section which can be fo_r
med at a so called "vena contracta";
2) downstream choking forming after the exit section, due to prej;
sure built up in turn and to water flashing.
In the first case he gives £he following correlation in order
to calculate V. '
r * - c i { L 2 p f c p U p - p s a t r j > / 2 a )
where the values of empyrical coefficient "C" may vary from about
0.6 to 0.64 .
It is to be noted that physically this coefficient (unlike Bur
nell's coefficient) is equal to a flow area reduction.
The last two theories are assumed to be "frozen" because they
do not consider any change in mixture composition during the
flow. Moreover the formulas expressing maximum flowrate by Bur-
nell and Zaloudek are very similar to those obtained from the pei:
feet gas theory flowing through a cylindrical duct when the as
sumption i) of paragraph 3.3.2.2 is adopted; for convenience we
shall remind this assumption:
- absence of dissipative losses, but admittance for heat exchan
This study too, is based upon an experimental work; all the
results are related to a convergent-divergent exit pipe.
The authors' assumptions are (apart from those valid for frc)
zen flow given in section 3.3.4.1):
- adiabatic expansion of vapor (Y= 1.3);
- no-slip;
- all the kinetic energy in the steam evolves from the vapor ex
pansion;
- flow conditions are controlled by a throat defined according
to gasdynamic principles.
The employed balance equations are:
- mass continuity:
r - - . (i) m
- vapor energy conservation: Y - l
h = h +x v p -Z- ( 1 - r ' ) (2)
o e o g o o y - 1
where "r" is the pressure ratio (p/p ) and v is defined as:
v = x v* + (1 -x )v. (3) m o g ^ o fo
From another form of equation (2) and from the perfect gas
theory (par. 3.3.2.1) it follows:
- V 2 w = L2(VVJ (4)
_x__
* o Y-l
ro
70 -
Moreover from the adiabatic vapor expansion assumption it fol
lows:
v* = v r* Y (6) g go
x and r is given by
* 1 Y *
r* è 2x v p — ^ ( 1 - r 0 (7) X V + fl -X )V 6 1
o g o go *- ->
In Fig. 3.52 the maximum flowrate versus reservoir critical
quality is shown: the reservoir pressure is the parameter.
The strongly marked maximum corresponding to low reservoir-
throat quality can easily be noted; they cannot obviously be ob_
served experimentally.
Moreover the authors say that the assumption of frozen flow
is not as restrictive as one may think, because the time needed -4
to travel through the nozzle is about 10 second. /64/
In the report later presented the authors make an interesting experimental data analysis from which it particularly results that:
- the pressure ratios (p/p ) in the diverging part of the noz
zle greatly decrease when subcooling increases while the flow
rate is constant;
- in the case of initial low quality the pressure ratio increa
ses as the stagnation pressure is reached.
- 71 -
3.3.4.5 MOODY 1969 /79/
The transmission of a pressure pulse in a two phase mixture
is the basis for this two-phase critical flowrate frozen theo
ry.
In this work the formulas are developed for evaluating both
critical flow and sonic velocities in the cases of separated and
homogeneous phases.
Here we shall only describe the critical flowrate frozen thec
ry and illustrate the main results of homogeneous assumption.
The author considers a pressure pulse travelling counter-flow
at velocity "v". The balance equations are written for a moving
control volume which is so thin that neither mass, nor momentum,
nor energy storage occur; neglecting body and potential forces
we have :
- mass conservation:
dr + Vd (—) = 0 (1)
- momentum conservation:
d(r v j + Vd r = -dp M
(2)
- energy conservation:
r(h
2 2 r y
KK + V d(— +r -y)-dp - H
= 0 (3)
In these equations v , v , v«F and v.. are defined in par.
3.3.3.9 and h is given by:
h = xh + (1 -x)kh, g J (4)
- 72
The critical flow in the case of separated phases is obtained
through the subsequent assumptions:
V=0 ; dx=0 ; ds =ds. g f
(5)
from the first of which we have c*) .
M
By solving equations (1), (2) and (3) it follows:
.2 'LlF pg+(T^Ep pfJ
(6)
r = vg tap J v£
(7)
L 3p js
with:,, a=a(k,x) ;, :y • =v -<p) ; vf=y£(p) .
Mpreover x is given through assumption (5), instead k is
assumed to be equal unity or given from the relationship obtai-/l3 9/ ned in the author's reports ' previously described:
k = JL v,
(8)
Taking the above into account, the critical flowrate results
to be a function of both x and p.
In figures 3.53 and 3.54 T is reported versus x ,p being the
parameter. The two figures are related respectively to k from
equation (8) and to k=l, note the very little difference between
these two results.
As already said, in this report Moody also follows a similar
procedure in order to obtain the critical flowrate with homoge
neity assumption. Fig. 3.55 shows the flowrate r in a diagram
(x) One can easily note that this assumption is the same as neglecting ener gy equation.
- 7;
with the same variables as in Figs. 3.53 and 3.54. It may be in-
teresting to note that the results for k=(v /v ) in Fig. 3.54
(separated flow), are quite different from those related to hom£
geneity assumption particularly in low quality region.
3.3.4.6 - HENRY-FAUSKE 1971 720/
The purpose of this model is to solve the problem of two-pha
se critical flow requiring only the knowledge of the stagnation
conditions and, at the same time, accounting for the non-equili
brium nature of the flow.
The duct considered is a DeLaval nozzle.
The conservation equations are:
- mass continuity
G v + G_ v. = A_ w +A w g g f f f . f g g
(1)
momentum conservation
- A dp = d (G w + G _ w J + d F g g f £ w
(2)
In a converging nozzle the wall shear forces are assumed to
be negligible by comparison with the forces due to momentum va
riation and pressure gradient.
With the assumption that at the throat
dr = 0 (3) dp
it results
T2-c d
dp X k + (1-X) f
I (1-x) k v£ + x vg
(4)
* - 74 -
The authors make further assumptions as following and justify
them through experimental arguments (the whole formulation is r£
lated expecially to low quality region):
(5)
— - --^JT"
X =
c
T ' f t
w = g
s gc
fc
X 0
fo
w f =
= s go
= S fo
w
n n (x) p v = p v 7
*o go *c gc
= 0 -1 c
= N i_ - -J c
~ d x E ~ dp
(6)
(7)
(8)
(9)
(10)
(ID
(12)
where x_ is the equilibrium quality given by:
XE = o fE
'fgE
and N is an empirically determined function of x Ec
Moreover from wel l known thermodynamic r e l a t i o n s h i p s we ha-
(X) From this condition (polytropic expansion of vapor) i t follows: dv -
g dp
v —
np
where n is given by the expression:
(1-x) cf/cpg * 1
(1-x) Cf/c' + 1/Y n =
- 75 -
ve ds g
dp M.
n Y (13)
and from the above assumptions the critical flowrate can be wri t
ten as :
x v r=<-^-£+(v - v - )
np g fo
(l-xo)N d s f E xQ c p g ( l / n - l / Y )
S g E - S £E d P ps £gE > (14)
c
This is an equation containing two unknowns (r and p ): in
order to obtain the solution, this equation is coupled with the
momentum equation (2), integrated between the stagnation and
throat conditions:
2 x Y
(1-x )v. (p -p )+-^r v o fo ' o *c Y-l
p v -p v o go c gc
r(i-x^)v. +x v ~i t-—° fo0 °
gd r2 (is)
Equations (14) and (15) solve the problem.
Some results of this theory are presented in Fig. 3.56 and
3.57 where the trends of r and of p /p vs x respectively are 1 c *o o v
shown.
It shall be noted that:
- if N is taken equal unity the solution approximates the
homogeneous equilibrium model;
- if N equals zero the solution approximates the homogeneous
frozen model;
- the function N is:
xT N = <
(TÎT £or ° « xo * °'14
1 for x £ 0.14 o
- 76 -
- when analyzing results from orifices or short tubes the authors
suggest to use value 0.84 for the actual flowrate reduction
coefficient;
- the model assumes neither completely frozen, nor complete equi
librium heat and mass transfer processes.
3.3.4.7 - D'ARCY 1971 780/
The starting point of this theory is the pressure wave propa
gation velocity in a two-phase system. The conditions in which
the wave velocity is zero are assumed as critical conditions.
The fundamental assumptions are:
a) no mass exchange between liquid and vapor phases;
b) the values of x , a , p are assumed to be known: the author G C C
suggests an empirical relationship between x and a so
that only one of these two quantities must be known.
Directly from mass and momentum conservation equations, writ
ten separately for the two phases, the author obtains the follo
wing four equations in the four unknowns (6w , 6w , 6m, <5p) :
p -(l-a) 6w +6m+ fw _-V) -r— f f f dp P£d-a) 6p = 0
p a<5w -6m+(w -V)^—(p a) ôp = 0 g g g dp g
p£(l-a)(wf-V)ôwf+(w-w£)ôm+-rd- p(l-a)
p afw -V) ôw -(w-w ) ôm+-;—(pa)ôp = 0 g g g g dp F
ôp = 0
(1)
(2)
(3)
(4)
where w is an unknown mean velocity so that w 6m is the net
momentum transfer from the liquid to the vapor. Once assumed
6m=0 the value of w becomes "immaterial".
(x) With the symbol 6 the author refers to small perturbations.
- 77 -
The compatibility conditions for these homogeneous linear equai
tions is the vanishing of the coefficient determinant; i.e.:
A = 0 (5)
When equation (5) is satisfied, equations (1) through (4) are
not independent and any of them can be solved, for example in
terms of op.
Taking into account assumption a) it follows:
(w -V)2-^-(p o)--r- (pa) 6m. g ^ dp VMg dp ^ _ Q ( 6 )
6 p -(w- 2w + V) g
By the definition of w it follows that the denominator of
equation (6) is always different from zero. Then: d r ^
7 -rr (p «
dF cpg"> and by substituting this equation in the determinantal equation
(5) we have:
d/dp [p (1-coJ
d/dp [Pg(l-a)] (wf-V)
2= = ~- (8)
Equations (7) and (8) form the compatibility conditions for
equations (1) - (4) and must be satisfied simultaneously.
In order to achieve the solution the author introduces two pa
rameters A and X whose expressions are given in App. 3.7. He
gives in a diagram (Fig. 3.58) X vs A satisfying equations (7)
and (8). The four intersections of the two curves (an hyperbola
and a parabola) so obtained are the two solutions for the sonic
velocities, i.e. (with symbols of Fig. 3.58):
- 78 -
a i = ( V l + V ï ) / 2 C9)
V = CV V2 ) / 2 (10)
Two-phase critical flow is obtained when it is no longer pos
sible for disturbance waves to propagate upstream, i.e. analiti-
cally: r _ v = o
(ii) Y" = 0
These conditions can be found by adjusting T until the V=0
point of the parabola in Fig. 3.58 falls either on one branch of
the hyperbola or another; as a conseguence two admissible values
for r are obtained. They are compared in Fig. 3.59, for a given
reference pressure.
/55/ 3.3.4.8 - ARDRON-FURNESS 1976
A simple frozen theory on the basis of Moody's model is deve
loped for the calculation of what the authors call "upper bound
mass flowrates". /13/ They write all Moody's 1965 equations in order to obtain
(see paragraph 3.3.3-5, eq. (2) that is written here in a dif
ferent form):
Ixw2+i(l-x)w2=x (h -h )+(l-x H h £ -hr )-h. (x-x ) (1) 2 g 2 f oK go ge' oJ K fo fe' fgev o v J
Consideration of these equations shows that since the enthal
py of vaporization is always greater than zero, the kinetic ener
gy flux and hence the flowrate increases with the decrease of
outlet quality. In order to achieve maximum flowrate they assume:
- 79 -
x = x e o
(2)
and isentropic vapor expansion, i.e.
( T) ) VY
P =
P . g go I P0 )
(3)
Moreover the authors assume:
p = const
and, like Moody, the criticality conditions are (*)
9k
3p = 0
(4)
(5)
(6)
Taking into account the thermodynamic identities valid in con
stant quality isentropic flow:
and
8h 3p J s
l - . i
1-x
h -h=x w { ° o go (. p0j i. x0 j
1 -x ^ p Y - l
P£oY Y - l
f iî ï Y
i-l-H the maximum flowrate is easily obtained:
r2-C
P W •go go
Y+ l
7 Y I X J P f
2 " f P . ^ f P o l ^ f l - x P g ^ L lpgoJ IPCJ I X P f
(7)
(8)
(9)
In the expression (9) p i s calculated by applying equations
(5) and (6) to equation (1) , taking into account equations (3)
(x) Condensation (see also PORTER) is forbidden for simplicity.
Jf
>.
> i
}
- 80 -
and (4). In Figg. 3.60 and 3.61 some results are shown; it may be
observed that when:
x 0—:-o
the theory predicts,. Berhouilli's flow.
/78/ 3.3.4.9 - RANSOM-TRAPP 1978
We have already presented an equilibrium model by the same au
thors (Par. 3.3.3.20). This frozen theory and the preceding (by
the authors) actually bound two-phase flow.
In developing this model two balance equations are adjoined to
those given in Par. 3.3.3.20,namely:
- vapor mass continuity
3 Cap) + -4- Cap. wj = 0 (1) 3t 'g' 9x Kg g
- vapor entropy conservation
9 3 -r— (ap s ) + -r— (ap s w ) = 0 (2) 3t g g' 3x Kg g gJ
These two equations with equations (1) through (4) of para
graph 3.3.3.20 forma closed system of six equations whose depen
dent variables are: a.p.w ,w_,s_,s . Two state equations charac-g £ f g
terizing the frozen transformation need to be adjoined to the ba
lance equations; these are the following functional relationships: Pg = f (P.sg) (3)
pf = f (p,s£) (4)
Finally two other transformation equations completely indivi
duate the frozen flow:
81 -
ds g
dp
d s £
= n
0 dp
(5)
(6)
Equations (1) , (2) (of this paragraph) , (1) through (4) of pa.
ragraph 3.3.3.20 and the conditions (3) through (6) solve the pro
blem. The solution is obtained through the characteristics method,
as in the case of equilibrium model. The only mathematical diffe_
rence is that this model produces a sixth order determinantal e-
quation.
In Fig. 3.62 we present the same variables of Fig. 3.49, in
order to compare the quantitative and qualitative differences bet
ween these two diagrams and the different value of velocity rela
ted to the two analytical models.
Apart from the observations made in paragraph 3.3.3.20, it may
also be noted that:
- in the choking flow criterion the only empirical data is the
thermal non-equilibrium degree (in fact the constant c can be
evaluated theorically);
- partial non-equilibrium actually has not be included in calcu
lations, since the lack of reliable vapor generation models:
- the frozen theory applied to EDWARDS' data predicts blow-
down to occur in about half the actual time, while equilibrium
theory well agrees with the experimental results.
3.3.4.10 - Non equilibrium, general theories
/82/ 3.3.4.11 - HENRY-FAUSKE 1970
• These authors started from the consideration that the greater
•
- 82 -
part of the theories developed before their study were equili
brium ones, whose main déficiences were the discrepancies from
experimental data at low qualities and the too high slip ratio
provided.
Stating that the assumption of frozen flow (dx/dp=0) is far
too restrictive, they propose a partially non-equilibrium model
suitable to predict critical flow for equilibrium quality less
than 0.02 approximately. The basic equations of this model are:
- vapor continuity equation
dz
A w g g X V
g = 0
- liquid continuity equation
d dz
Af Wf (l-x)v.
= 0
CD
(2)
- mixture momentum equation
dz = r dz x w +(l-x) L g •<] (3)
in which the wall shear and the hydrostatic head are neglected
in comparison with other pressure terms.
A first criticality condition is:
d r T 0
-"c
(4)
therefore equation (3) may be written
r * - < _d_ dp
xk+(l-x) ' k
(l-x)kv-+xv f gj_
-1 (5)
- 83 -
As in the examined region of the Mollier diagram
V- « V £ g
dv,. dv
dp dp
(6)
with the assumption
k £ 3
and making the possible simplifications consequent to the hypo
thesis 0<x<0.02, equation (5) may be written as:
" -1 > r = - < N
dv dx XC-T-£ + V -r-
E dp g dp
dN g E dp (7)
where N is a non equilibrium parameter defined as
N = kx.
(8)
The term dx /dp can easily be obtained as a function of p
from equation:
ds = 0 (9)
with the further assumptions:
- polytropic expansion process (the authors show tha t the variat ion of n
be tween 1 and 1.3 i n t h e i n v e s t i g a t e d f i e l d ( p * 50 p s i a ) af
f e c t s t h e c r i t i c a l f l o w r a t e of l e s s t h a n 1 $ ) ;
- dv / d p determined from s a t u r a t i o n p r o p e r t i e s ( n = l ) .
E q u a t i o n (7) becomes:
r2 -C •N
cHE
dN - V X^ —;
g E dp
(10)
- 84 -
where T Tt_ is the critical flowrate resulting from hoiaoge-cHE
neous equilibrium theory (see paragraphs 3.3.3.2 and 3.3.3.3).
Moreover in equation (10) the influence of the non equilibrium
parameter N may be clearly seen (infact for N=l, it results
r =r ) . c cHEJ
Regarding to N, it is easily shown that equation (8) can be reduced to
N = f1 n, (11) xE (1-a) vg
However the value of N is obtained esperimentally and, as
a first approximation, it results only as a function of x :
N = 20x c (12) e be
Moreover, the d e r i v a t i v e dN/dp may be w r i t t e n a s
dN dN , fc dp dz dz
and then as dp/dz at the exit is very large and dN/dz appears
to be nearly zero, it follows that dN/dp can be neglected.
Finally, from equations (10) and (12) and from the above con
clusion, it follows:
r = r / / N . c cHE ' e
Some results are shown in figures 3.63 and 3.64; in Fig. 3.65
the ratio r /T TTT, is plotted vs x_ . c cHE Ee
The simplicity of this theory that takes into account slip
and metastable effects is to be observed; the simplicity is coun
terbalanced by the too narrow field of application.
In the same work the authors present also a two component cri_
85 -
tical flow theory which is based on the frozen flow assumption
(dx/dp=0) .
3.3.4.12 HENRY 1970 731/
This other theory developed by Henry applies non-equilibrium
effects to nozzles with high (L/D=12) and very high L/D (L/D >
12) ratios, smooth or sharp edged at the entrance. Now we shall refer
to sharp entrance and initially saturated or subcooled liquid.
The author starts from the following phenomenological observa,
tions:
a) a great pressure drop is present at the inlet;
b) the pressure remains about constant up to L/D-12;
c) for L/D>12 flashing of the mixture generates a momentum pressu
re drop that cannot be neglected.
By further neglecting the quantity d vf/dp and assuming slip
ratio equal unity, from equations (1)T(3) of the preceding sec
tion it easily follows:
r = i / dv
X —r-^- + (V -V ) dp
dx g fJ dp (1)
In order to evaluate the mass transfer term (dx/dp) the au
thor obtains the following correlation from experimental results:
d XE • N — £ - (2)
dx dp dp
(x) In particular, from more or less empirical considerations, the authors reach the formula rc= IcHE
k n o t unexpected because for x=xE, N=l/k (see eq. (8)).
- 86 -
where
N = < 20 x for x < 0,05
and dx /dp, being ds=0, is
dp
(3) for x * 0,05
ds( ds (1-aO- dp dp
'fg (4)
At this point it is necessary to distinguish two situations:
A) L/D =12;
B) L/D > 12.
In the first situation, from the phenomenological assumptions
a) and b) it follows that we have a pressure decay only at the in
let and it may be obtained from:
p - p = Ap. n . o c "inlet
r 2 v f
C fO 2 C 2
(5)
where C is a constant (the orifice contraction coefficient),
whose value is 0.61.
Moreover the last assumption to solve the problem is:
x = 0 c (6)
Equations (1) and (5) which take into account equations (2),
(3), (4) and (6) constitute a set of two implicit equations in the
unknowns p and r.
In situation B), two are the differences from the case A): the
first is related to the fact that equation (5) is no more valid
and, in accordance with the phenomenological aspect c), it is nece£
sary to write a further term in order to obtain p -p ; for homo
geneous flow flashing in a constant area duct the infinitesimal
- 87 -
/128/ frictionless momentum pressure drop (following WALLIS 1969 )
can be written as:
dp= -r d (l-x)v +x v £ g
(7)
and integrating between the inlet (x=0) and the exit section (x=x ),
taking into account equation (5), we obtain:
v, p -p = r *o *c c
2C -=-+x (v -v_ ) 2 e ge fo (8)
The second difference is related to the fact that not even e_
quation (6) is valid any more. Equation (6) is substituted by:
x =Nx <q e E 1-exp
-B(L/D-12) (9)
in which the constant B is obtained from experimental data and
results :
B = 0.0523 (10)
Analogously to situation a) it may be concluded that equations
(1) and (8), with equations (2), (3), (4), (9) and (10), consti
tute a set of two implicit equations in the unknowns r and p .
Some results are shown in figures 3.66 and 3.67. Namely in
Fig. 3.66 r is plotted vs p , L/D being the parameter; in
Fig. 3.67 r vs L/D is shown, p being the parameter. In Fig.
3.68 p /p is shown vs L/D, p being the parameter. Finally, in
figures 3.69 and 3.70 the influence on the flowrate of subcoo-
ling degree is clearly shown.
Up to this point we have treated the sharp entrance: in the
smooth edge case, the author notes that there should be no signi
ficant separation of the main flow from the duct walls; thus the
- 88 -
coefficient C of eq. (5) results equal to one.
Moreover the assumption x=0 at L/D=0 produces a flow pattern
identical to the one at L/D=12 in the sharp entrance configura
tion and therefore the exit quality may be expressed as:
xe=NxEk-exp(-BL/D)j (11)
where B has the same value given in equation (10). Equation
(11) substitutes equation (9) in the solution of the problem for
smooth edged nozzle; some results are shown in Fig. 3.71.
The simplicity and the completeness of this model justify its
diffusion.
/83/ 3.3.4.13 - KLINGELBIEL-MOULTON 1971
The Klingelbiel-Moulton work is based on experimental data in
which, apart from the classical quantities, the two-phase jet
thrust is also measured.
The authors note that condensation rather than vaporization
is the result of high quality isentropic flow. Moreover no model
is really very accurate in the range of quality below about 201.
Therefore two new phenomenologically based models for critical
flow of two phase mixture were formulated.
The first one accounts for the above mentioned condensation,
the second takes into account the fact that some mass of liquid
is entrained like droplets moving at velocity of vapor stream.
We shall describe the latter, because it seems more analytically
developed. It was called Entrained-Separated-Flow (ESF) model.
The conservation equations are:
- 89 -
- mass continuity
r = p w = const, m m CD
- vapor momentum balance
dp A =< g 7
r A +d(r A ) g g g g
(w +dw )-T w A -Dw d(T A )-(l-D)w^d(r A ) g g g g g g g g f g g
> (2)
- momentum equation for the liquid phase
(w£+dw£) (l-D) + (r A,)+d(r Af) (w +dw )D-
-wfAfrf(l-D)-WgrfA£D-Wgd(rfA£)D-wfd(r£Af)(l-D)
dpA£= < r£A£+d(r£A£)
(3)
where D is the constant weight fraction of liquid moving at va.
por velocity and (1-D) is the fraction moving at liquid velocity.
We assumes that evaporation takes place from the entrained
and non-entrained liquid in the same ratio as the weight distribu
tion of the liquid itself. From the above equations it results:
r = I" -dvT (4)
where v is the volume of an entrained-separated system given
by:
|_x2+Dx(l-x)J v ]x +Dx g
a
v£(l-x)2(l-D)2
The expression for the slip ratio (k) is obtained by differen
tiating the kinetic energy expression with respect to void frac
tion or slip ratio, with the assumption that choking involves mi.
nimization of entropy production.
- 90 -
It results
x = D+(l-D) V
g
h* J
J / J
v f x
1 + D 1-x (6)
The formula of D is obtained from experimental data:
f1+0,14 In x D = <
for 0.19<x<l (7)
0.94-0.204 In x for 0.01<x<0.25
Equations (5), (6) and (7) permit r to be a function of p
and x only, e
Some results are presented in Fig. 3.72.
The authors conclude that neither of the two models offer sub
stantial improvements over existent models.
3.3.4.14 - KLINGELBIEL-MOULTON 1971 /83/
Taking into account what has been said in the preceding para
graph, this new method involves calculating critical flowrate
from the energy balance written in the same form as Moody (see
Par. 3.3.3.5). Similarly from mass and energy balance they ob
tain:
r =< 2(h -h.-xh.
•L — xv 12 ,-- ^ + (l-x)v£l ll + x(k2-l)
(1)
The slip ratio and the values of "D" are obtained from ex
pression (6) and (7) of the preceding paragraph respectively.
- 91 -
By substituting the expression of "k" in equation (1), r remains
a function of p and x only and x values may be calculated e e e
from an empirical equation related to the authors' experimental
results; the expression for k is reDorted in appendix 3.8.
Therefore for each value of p (once given h ) a value of r *e 6 o
is obtained.
This formula well agrees with experimental data, with an ave
rage error less than 2% in the quality range 0.01 <x<0.99 and
pressure range 20<p <75 psia. /84/
3.3.4.15 - MALNES 1975 '
The author employs a set of four one-dimensional two-phase
conservation equations, including thermodynamic non-equilibrium.
By simplifying this model with the assumption of steady state
conditions, the pressure gradient goes to infinity at the exit
section when critical two-phase flow occurs: however,in a real
system,irreversible phenomena and two-dimensional effects cause
exit pressure gradient limited, exit pressure higher and exit v£
locities lower than those predicted by one-dimensional theory.
In this synthesis we shall not show the analytical develop
ment of the above but we shall only expose the main model.
The characteristic of this theory is the consideration that:
- the release of dissolved gases is the main mechanism for bub
ble formation during expansion of two-phase mixture.
In order to achieve this a conduction controlled flashing co£
relation is developed.
The balance equations are:
- vapor mass conservation:
- 92 -
9 3re — ap = - — 4 + 3t 3z
*
liquid mass conservation:
3 r i arf *
- mixture momentum conservation:
(1)
(2)
_3_ 3t apgW£ + (1~a)pfWf
= _ JE . JL 3z 3z g g f £ J
- mixture energy conservation:
3E 3t y 3t V^v^i^vptyv
(3)
(4)
where "F" i s a gene r i c term t ak ing f r i c t i o n i n t o account and "E'
( t o t a l i n t e r n a l energy) i s given by the e q u a t i o n :
1 2 1 2 E = ap (e + - wJ + (1 - a) p f (e f + - wf)
(5)
The above system may be solved when the following quantities
are known independently:
- slip ratio
- friction (F)
- flashing O ) .
The author solves equations (1) through (4) by a finite dif
ference technique from upstream reservoir conditions up to the
inlet of a nozzle diffuser. He assumes:
1) slip ratio equal to one ;
2) two-phase friction multiplier given by Becker ;
3) evaporation rate ij; calculated by an original expression
which will be briefly discussed.
- 93 -
The author says that flashing starts essentially from:
a) wall boiling;
b) release of gases;
c) impurities;
d) statistical fluctuations of the vaporization in a superheated
liquid.
Referring to the last point he assumes that "dissolved gases
are the main mechanism of flashing in reactor conditions, surfa
ce boiling is insignificant expecially for large diameter pipes",
justifying this hypothesis with the flashing delay experimental
ly observed in a gas free liquid .
Taking into account the Zuber results for the calculation
of the diameter of a single bubble as a time function, the evapo
ration rate \l> may be calculated by the following equation:
P c \ p_ - - V. <j; = (R -S B + RJ P-g *
° pf l hi fg
(1-cO '3.„2 AT (6)
where R and R„ are dimensionless empirical constants, whose va-° 7 5
lues are 7x10' and 2x10 respectively (for a particular case ) and
1 - 2a for a < 0.5 B = < (7)
0 for a > 0.5
The equa t ions (1) through (4) and ( 6 ) , wi th the assumptions
(1) and ( 2 ) , mathemat ica l ly so lve the problem. Some r e s u l t s a r e
shown in F i g . 3.73 (T vs p , h i s t he p a r a m e t e r ) ; in F i g . 3.74
(x) The author says that typical values of dissolved gases in a LWR are : 0.5 ml/1 in a BWR at 1. bar and 25 °C and 10 ml/1 in a PWR at l .bar and 25 °C.
- 94 -
r versus exit pressure is reported*, finally irt Fig. 3.75 pressu
re and void fraction versus pipe length are shown, for r = 11100
kg/m sec.
/89/ 3.3.4.16 - PORTER 1975
A very complete analysis is thé dnë developed by Porter. He
considers both equilibrium and nônJequiiibrium thermodynamic ef
fects, allowing also for a change in slip ratio and calcula
ting in the two cases, (equilibrium and non-equilibrium) the cri
tical flowrate, in the range from subcooled to superheated sta
tes through saturated state. ' •'•
Moody's theory is taken as à reference ; the equations adopted
are the same as those derived by Moody in treating saturated mix
ture in thermodynamical equilibrium.
In this work we consider the hpn-equilibrium theory only but
we shall also present some results related to the equilibrium
one. ,
I) Mixture in the saturated_field
Fundamental equations:
- mass conservation:
r = const. (1)
- momentum equation:
Td Lk(l-x*)v f+x*v gJ l x * + * -x k
= -dp+2f f L/Dvér 2 d z + E * + k ( 1 ~ * y dz g k(l-x>,nv
(2)
fx) Two values of slip ratio are used: k=l and k= (-=•)
- 95 -
- energy equation:
2 2 * Vt = H + T&(l-x*)v f • x*v J j V + f - J (3)
6 k
where "Q" is the heat exchanged with the exterior per unit length
of exchange surface, "f _" is the Fanning factor for water (f =0.046 -0 2
Re» ' for a smooth pipe) and "<J>" is the two phase friction imU
tipl ier .
In the case of non-equilibrium the author assumes:
- no water flashing;
- isentropic expansion of steam.
When considering the first assumption the initial water flow
is:
r. = (l -x )r fo v o
The vapor expansion causes a partial condensation that may
quantitatively be expressed as:
s - s£
x = -&—£- (4) con s- v J
fg
from which the total water flow results:
r. = (i-x )r + x r(i-x ) = (i-xx )r (5)
fe v o o v con' v o con v J
and the total steam flow results:
r = x x r (6)
ge o con
In this case, therefore, we have:
x* = x x (7) o con
Equations (2) and (3) are derived with respect to "z" and
- 96 -
dp are solved simultaneously to give -p-
field, the results are obtained (as
ming:
k = 1
or v V k = (-*) 3
II) Subcooled water
The only equation is the momentuir
dp = 2£f L/D vf r2 dz
where we have neglected inertial and
III) Superheated steam
- momentum conservation:
2 v
r d(-f) =-dP + 2f£
- enthalpy conservation:
2
H •Sf. = H + ^ r 2
o TA g 2
and —.— . dz
mentioned
i equation
gravity
L/D v r2d g
These equations are differentiated with re dp
may be solved in order to obtain -j*-
The results related to situations
iteratively, for assigned values of
and -r— . dz
I), II),
In the saturation
above) by assu-
(8)
(9)
given by:
(10)
forces.
z
spect
III)
total pressure
and by calculating the inlet pressure through pressure
at each step the two values coincide
In this way the three sets: (1),
', the cal
(2), (3),
culat:
(11)
(12)
to "z" and
are obtained
at the outlet
drops. When
Ion stops.
(7) and (8) or
- 97 -
(9) for saturation non-equilibrium flow; (1) and (10) for sub-
cooled non-equilibrium flow; and (1), (11) and (12) for super
heated steam flow are solved.
Figs. 3.76 and 3.77 show the critical flow versus stagnation
enthalpy in a form similar to Moody's. In Fig. 3.78 and 3.79 the
static pressure is plotted versus stagnation enthalpy. All the a_
bove figures are related to non-equilibrium flow.
EqiT^Librium flow results are presented in Fig. 3.80 and 3.81
in which r vs h and p vs h have been calculated respectively 0 e 1/3°
for k = 1 and k = (v /v_)
By the examination of above figures one can draw the follo
wing basic conclusions:
- in non-equilibrium case, there is not a great difference if <s
quation (8) instead of eq. (9) is used;
- in equilibrium case, a strong dependence of r from subcooling
is observed on the left of the saturated liquid line; in case
of non-equilibrium the strong dependence is observed on the
right of the same curve;
- in every case curves converge towards the same values at the
saturated vapor line, since the assumptions in the saturated
field become identical.
/90/ 3.3.4.17 - RIVARD-TORREY 1975
This work analyzes the transient depressurization phenomenon
entirely.
A two-dimensional time dependent theory is used and the au
thors emphasize some relationships describing:
a) compressibility of the liquid phase;
b) phase transition;
- 98 -
c) heat transfer and friction between the two phases;
d) friction with the duct wall.
The balance equations are written as:
- vapor continuity:
—: (cxp J + V ' (ap w ) = i|> - ty 3t V - v *g-gJ Ye rc
- liquid mass continuity:
Testo aervenuto nel settembre 1980 - vapor-momentum conservation: g
^-(otp w ) + V(ap w w ) = - a V p + F(w_-wJ +\\> w_-4> w + f
- liquid-momentum conservation:
CD
(2)
( 3 ) , (3a)
az Q l - a ) p ^ J + V • £ ( l -a) P£w£wf2=- (1-a)V p + F (w -v^) «l> • w - * ^ + f£
(4),(4a)
- vapor-energy conservation:
ap g
3e —£ + ap w • v —f- + ap w • y e = - p v * I aw + (l-a)w_ |+F(wJ.-w ) + 3t ^ g - g - g ~ J g ^C-1 f g
• (*e-tc)P(vg-vf) • Q(If-Tg) • I ^ H . ^ (5)
- liquid-energy conservation:
3e (l-a) pf-r*+ d-a) p A . Ve. - -(*,-*) L.P(v -v ) + e - e J + R(T -T ) + Jf 3t ' ^ ^ *& * -~f "' "re V L r v ' g "f • ' ~g w£-" " v g f
fvix Ticond
The authors note that some transient phenomena linked to the
Riprodotto in offrt pntêo il Laboratorio Tecnografico délia Direzione Centrale Relazioni Externe <kl CNEN • Viale Rcgina Margherita 125, Roma
- 99 -
microphysics of phase change are necessarily neglected.
Two implicit state equations relating vapor pressure and tern
perature are written as:
P = pg(nsg(P) -1) eg(p,Tg) (7)
T =T + fh (p,T ) -h fp)J /c (8) g s L g^' gJ sg^'-1 ps
and the state function n is given in Appendix 3.9. These re
lations are valid for a > a (a being a fixed value) and for o o &
T < T . g s
The authors also present other forms of equation (7) and (8)
that are not presented here because they yield no quite diffe
rent results.
The transformation equations describing the mass exchange
are:
V2 T f " T s \\> = X B ( l - a ) p , a ( T R) — - — - v a l i d f o r T r > T (9)
e e f ^ s T f - s
\b = 0 o t h e r w i s e e
1/ T - T ij> = A B ( l - a ) a p (T R) - 5 - — & v a l i d f o r T <T (10)
c c g s T g - s
^ = 0 o t h e r w i s e
In these equations "B" is proportional to the contact area
between the two-phases and depends upon an empirical constant
"N" (number of bubbles per unit of mixture volume). The fric
tion between phases is:
- 100 -
F = f <j_p|_cdlwg - wf I + ^ L y + (i - «)v£)jj> B
in which "r" is another function of "N" and "a".
The pipe wall friction is written as:
— 2 2 — 2 T = I -af p u a / 2D I <f> wg L g g g -1 g
(ID
and
T .= -(l-a)Lf.P-:4(l-a)2/ 2DJ4»? wf LfHf "F
(12)
(13)
in which the friction factors f and f _ depend upon the Reynolds
number and the pipe roughness (see App. 3.9). The heat exchange
is analyzed through the function MQM which has not a definite
functional dependence.
After investigating, amongst other things (KACHINA code), the
effects of large relative velocities and large temperature difftî
rences between the phases, the authors conclude that they are ne_
gligible. / ? f\ I
The authors analyzed Edwards experiments and find a good
agreement between experimental and theoretical results over all
the transient.
In conclusion this is one of the most evolute models adopta-
ble also in analyzing depressurization transients in complex ge£
metries. No result is given about the flowrate.
3.3.4.18 - KROEGER 1976 Ml
A complete and analytically consistent analysis is the one
performed by Kroeger. In reviewing the preceding models he points
out two interesting aspects that are not always appropriately con
- 101 -
sidered:
- the vapor generation rate analytical description;
- mathematically self-consistency of a model.
The primary interest of this work is the analysis of expan
sion of initially subcooled water, also considering the flashing.
The main explicit assumption is to neglect dissipation terms in
energy equation; moreover, axial heat conduction, surface ten
sion and uneven distribution of the phases in a plane normal
to flow direction have not been taken into consideration.
This model can be characterized by the following quantity,
called vapor drift velocity:
w = w - w (1) gm g m
where w is the mixture centre of mass velocity, m
The balance equations are:
- mixture mass conservation:
A£(pJ + Ptn~ (AWJ =0 (2) Dt m m 3 z m
- vapor-mass conservation:
D 3 Ap - x =Ai^-— (Ax o w ) (3) m Dt st 3z st m gnr
- mixture momentum conservation:
Ap £• w = -A P- - AP -ET -•£• (A r-^- p w2 ) (4) m Dt m dz m w dz 1-x m gm
st **
- mixture thermal energy:
AP ^ n = A S " + A x * ~ w 7 + sq -'— l~Ax p w hc ~\ -Mm Dt m Dt st p gm dz Hex 3z «- stMm gm fg-1
- Ar ^ 1,- st.,
w + -(- ) w m 2 1 -x gm
w gm
(5)
1-x . st
- 102 -
V J^ _ p m
pgCiyp}
1 " X 4- X 4.
St ( St
f g
with the state equations:
Pf = P£(hf,p) (6)
(7)
(8)
h = (1-x Jh_ + x Ji (9) m v stJ f st g K J
All the quantities are generally not taken along a satura
tion line. In order to solve the set constituted by equation (2)
through (5) four constitutive relationships are further neces
sary; these are written by the author as:
* - vvv^st'V (io)
wgm = g'V^st'V ^
Tw = VVV^st'V (12)
q = f. (h ,h,.,p,x .,w ) (13) ex 4 g f r st m v J
where on the right hand side the derivatives with respect to
"z" and "t" of h,,, h , p, x . can also appear. Since the field ± g st
of study of these equations is essentially referred to low qua
lities (see above) it is not a very restrictive assumption to
write:
h = h . (14) g gsat v -*
In the following we shall briefly present the author's con
clusions about the function f and f , describing the vapor gen£
ration rate (ip) and the vapor drift velocity (w ) . The vapor
- 103 -
generation rate, for homogeneous equilibrium conditions, may be
evaluated by the relationship:
Eq rax. Dp_ _1_ nex
TŒ " pm l3pJc Dt h. A s fg
(15)
where the first term on the right hand side represents the gene
ration rate due to pressure variation and the second term ac
counts for exterior heat addition.
To take into account non-equilibrium phenomenon the author aj5
sûmes:
0 for Xp £ x.
£ r^ (16) 9x
m 9t> Dp. Dt
for > x.
Also another model of vapor generation rate is presented by
the author, through the evaluation of vapor drift velocity.
In this model, the function f? is related to a so called di£
tribution parameter "C ", defined as: o
< Y >
< a >
const
Fr n (17)
where <Y > is the flowing volume concentration. The author shows
that in the examined range of values, Fr >> 1 and then the se
cond term in the right hand side is negligible. Moreover the
distribution parameter C is obtained by interpolating some
[mental data of Zuber' ': expern
c = i + o i9f 3 2 0 6 "F) Lo 1 + u-iyt 3206 J
2.4 (18)
where "p" must be expressed in psia.
- 104 -
The vapor drift velocity equation is then written as
(l-x)+x - pf
w = (C -1) s w (19) gm o pf pf m
(1-x) + x — + x C (—-1) g g
Also another expression for w is obtained by the author gm
and is not presented here.
In conclusion the problem is defined by:
- the constitutive equations (16) and (19) characterizing this
model, and (12) and (13), of which the former has the classi
cal form, and the latter has not been expressed;
- state equations (6) through (9) and eq. (14);
- balance equations (2) through (5).
The set is solved by the method of characteristics in which
the dependent variables are assumed to be: h_, p, x, w and the
independent variables are obviously "z" and "t". The observations
are as following:
- the vapor drift relationships characterize the momentum dis
tribution between the two phases (as the slip ratio in other
models);
- the assumption (14) corresponds to a statement on the energy
distribution between the phases;
- the two relationships for drift flux and for vaporization ra
te fairly increase the free degree of the solution;
- all the chosen constitutive relationships lead to real roots
of the determinantal equation;
- in the available reference only a comparison of theoretical
results with Edwards depressurization experiments is presen-
- 105 -
ted; this comparison is quite good when constitutive relation
ships are opportunely chosen.
/29/ 3.3.4.19 - BOURE -GIOT-FRITTE-REOCREUX 1976
From a mathematical point of view this is undoubtedly the most
developed among the theories presented in this review.
The authors' aim is to present a general model from which any
other consistent model should be obtained; moreover they wish
to point out the effects of some constitutive terms in the balan
ce equations on the critical flowrate evaluation. Regarding to
this, the authors say that in analyzing any thermodynamical tran
sformation, a rational procedure intends to postulate the trans
fer laws (cause) rather than the nature of the transformation ijt
self (depending upon the transfer laws).
The assumptions adopted are:
- two dimensional effects negligible;
- pressure uniform in any cross section;
- diffusive and turbolence effects negligible.
Moreover the importance that these three assumptions may ha
ve is recognized by the authors.
A complete system of six balance equations and seven consti
tutive relations is written as:
- mass cont inui ty equation:
+ 9aG+ f9pKl j n [3pKl 8 A Y 3aG f3pK] 3D
±o — + a — —*-+a — ••' ± o w — + a w —— -*-+ MK9t K[3pj 9t K[3hKJ 3t 1 K3z K K(3pJ 3z AhK p AhK
+ VK irlVGirJ V K ^ J IT±M=-VKWKÂ
C1>la)
only i f K=G p
- 106 -
- momentum conservation equation:
9 aG ±PK "ôTVlC
(dp ^
3p I )
3OL 3p L 3t T K 3t
29aG , 2 1 + \
r3p AhK
3AW Ot 0 -z—
GPG 3 t *Vfc
K
VÊ fdpJ 3Ah
3hK.
K. 3z
3P
r i m
K )
9 w L r
¥z+2aA\1T+
; o n l y i f K=G 3AK
3P,
31V
3Ah I
3t
Ah f only i f K=G
(2,2a)
- e n e r g y c o n s e r v a t i o n e q u a t i o n ;
K h +-r-K 2
I J
8 a G f +a
3 t K
Y 2 ï
h +—-K 2
.1 J
Kl 3p
.
±p„ h +— —
\
^ ^ C s a t . K p -*-+a.j3„w„-—+
3Aw
only i f K=G
+a K
w.,
Vf f ^
*K JV _
3AhK. 3t_±PKWK
2
Vf 3aG W,
Vf
Vi 3W(
V T
2 3wT
3z aGPG 'H 3 W ] 3Aw
3z
r 1
MI
t. ) • Q K = - a K P K \ g C O S 0 - a K P K \
only i f K=G
K A» h +—
K 2
+a w K K
3 P
r 2,
+p a*
sa t
V / Ah K p
K
ap 3z
W_,
Vf 3p
K 3z
(3,3a)
Moreover:
- pressure drop per uni t length ( * )
3x . 3x.
h = T k , o + f ( T T ' T T 3 (4 )
(x) This relationship takes into account friction between the two phases of fluid and between each phase and the duct wall.
- 107 -
- heat transfer to each phase from the exterior per unit volume:
' 3x. 3x.
Qk = «k.o^TT'T^ C5)
- mass transfer from the liquid to the gas phase:
9x. 3x.
M = M + M(-rr f ~ ) (6) o 3t 3z
- momentum transfer from the liquid to the gas phase:
3x. ax. MV = (MV)o + (MV) (-~ , -~) (7)
- energy transfer from the liquid to the gas phase:
3oc. 3x. MH = (MH)o + (MH) (- , ~ ) (8)
In all above equations the subscript K refers both to liquid
(L) and to vapor phase (G), the term A' means dA/dz, and by x.
we intend all the dependent variables (see below). Substituting
equations (4) through (8) in equations (1) - (3) yields a set of
six equations in which the dependent variables are ar, p, w. ,
Aw = w - wT , AhT = hT - hT . Ah0 = hn - h_ . G L L L Lsat G G Gsat
The critical condition results from the formulation of the
model. At the same time it implies the vanishing of the deternu
nant of the coefficients A of equations (1) - (3) and of the de
terminants N. which are obtained from the matrix of A, by subs-
l '
tituting ith column with the column on the right hand side mem
bers of equations (1) through (6J.
In particular the condition:
A = 0 (9)
- 108 -
is a necessary flow criterion and:
N. = 0 (10)
is the "compatibility condition" involving the nozzle geometry
among other terms.
The flow may be considered critical if we verify that, accor
ding to physical experience, the flowrate cannot increase when
the external conditions are varied.
The authors analyze some particular cases, studying the in
fluence of the constitutive laws. The conclusions may be drawn
as following:
- by varying the constitutive laws any consistent model should
be obtained from this theory; in particular, the authors af
firm that this is achievable only when in the transfer terms
the presence of the partial derivatives of the dependent va
riables is allowed;
- the gradients of the dependent variables are generally not in
finite in the critical section;
- in the case of single phase flow investigated by the authors
together with two-phase flow, critical velocities different
from sonic velocity are found: this too is a consequence of
the presence of differential terms in the external constitutif
"A two-fluid model for critical vapour liquid flow"
IJMF, 1978.
/132/ ARDRON K.A., ACKERMAN M.C.
"Studies of the critical flow of subcooled water in
a pipe"
CSNI Spec. Meet., on Trans, two-phase flow, Paris,
1978.
/133/ WOLFERT K.
"The simulation of blowdown processes with conside
ration of thermodynamic non equilibrium phenomena"
Proc. of CSNI Spec. Meet., Toronto, 1976.
/134/ BECKER K.M. et al.
"An experimental study of pressure gradients for flow
of boiling water in vertical ducts"
AE - 86, 1962.
- 164 -
/135/ ZUBER N.
"Recent trends in the boiling heat transfer research"
AMR, Vol. 17 n° 9, 1964.
/136/ RIVARD W.C., TORREY M.D.
"K-FIX a computer program for transient, two dimensio_
nal, two fluid flow"
LA-NUREG-6623, 1977.
/137/ HIRT C.W., ROMERO N.C.
"Application of a drift flux model to flashing in
straight pipes"
LA-6005-MS, 1975.
/138/ HANCOX et al.
"Analysis of transient flow boiling"
Proc. 15 Conf. Heat Transf. - S. Francisco, AECL 1973.
F I G U R E S
m
o
Po , m 0,T 0
l
k-k
Fig. 3.1 - Reference scheme for this work
M
0.5-
1.0-
1.5-
2.0-
3.0-u0 0.2 0.4 0.6 0.8 1.0
gw Q*\N*
F i g . 3.2 - Liepman e t a l . (Par . 3 . 3 . 2 . 1 )
08
06
04
02
M<1
M=1
M>1
i ï 8 p (ARBITRARY UNITS )
Fig . 3.3 - (Par . 3 . 3 . 2 . 1 )
>! \ i
E
CO i _
•4-»
IB CD
RAYLEIGH LINE
arbitrary units
Fig. 3.4 - Perfect gas, flow from cylindrical duct. Ass. i) (Par .3.3.2.2.).
s
4->
ry
un
arb
itra
B-
FANNO LINE
/ /
arbitrary units
A
\
h
f * 0
qex=o
^ c
Fig. 3.5 - Perfect gas, flow from cylindrical duct. Ass. ii) (Par. 3.3.2.2.) .
r ( 2 g c P o p o )
Incompressible liquid
F i g . 3 .6 - (Par . 3 . 3 . 2 . 3 ) .
100.0 p
0.01
1689.5 kN/m , 1100 Btu/lbm
r e f 1(2.326) 105J/kg m
r „ 11000 tbm/sec-ft2
r e f (4882kgm/»ec-«n2
PRESSURE, p / p = 0.25 o ref
X 2.0 4.0 6.0 8.0 10.0 12.0
h / h , o r e i
Fig. 3.7 - Results from HEM (Par. 3.3.3.1
ioao —
o.i
p MOOIbf/in.2 I P r e £ * |689.5 kN/m 2 j
, f 100 Bto/ibm | \ e £ ° 1(2.326) 105J/kgm|
STAGNATION PRESSURE, p Q / p ref/ 0.25
2.0 4.0 6.0 8.0 10.0 12.0
h /h o r e t
F i g . 3 .8 - R e s u l t s from HEM ( P a r . 3 . 3 . 3 . 1 ) .
5000
4000
£ 3000
E -Q
2000
1000
'0 0.2 04 0.6
Fig. 3.9 - Results from HEM (Par. 3.3.3.1).
5000
- 4000
o CD
£ -Q
3000
2000
1000,
po=1000 p s
X0 = 1.52-10" hD=552 BT
s0 =0.753 Bl
ia 2
U/lbm
U/lbm'F
'0 0.1 Xe 0.2
Fig. 3.10 - Results from HEM (Par. 3.3.3.1).
7000
200
Fig. 3.11 - Resul
400 600 800 1000 Pe (psia)
;sults from Lahey-Moody, s =0,753 BTU/lbm °F 'ar. 3.3.3.2").
(1-X)
0.50
0.25
0
y y^y
sly AS
y ^^«--""^ 2
— — —
5.--" ^ ^ ^
3 " * » ^ ^ ^
I--
0 50 100 150 •è/p
AS,rr ,w r
1.0
0.5
0 200
F i g . 3 .12 - B a b i t s k i y 1973 ( P a r . 3 . 3 . 3 1 ) ( s - s ) / s =AS when x i s 2 ) w n s / w s = w r ; 3 ) r r = r n s / r s ; 4 ) ( 1 - x ) from e q . ( 6 ) ; 5 ) ( 1 - x ) from e q . ( 5 ) .
3) . c;i 1 c u l a ted from eq . ( 5) ;
240 w m
200
160
120
80
40
0,
3/
2 /
1
295 31
AS
0X)4
333 353 373 oK
Fig. 3.13 - Babitskiy 1973 (Par. 3.3.3.3). 1)(s0-s) when is calculated from eq.(5),
(kJ/kg°C); 2) w from eq. (2); _ _ 3) w from the equilibrium scheme [w(m/sec)J
c lbm
sec f t 2 jw» -
a 100 TOO 300 400 MO «C ?» KB WO 1000 11» SÎCC 1300 « « ^ ( B T U / l b m )
F i g . 3 .14 - Moody 1965 f P a r . 3 . 3 . 3 . S) .
P c(psia)
KKS
h (BTU/lbm)
F i g . 3 .15 - Moody 1965 ( P a r . 3 . 3 . 3 . 5 )
ri I M
c UJK»
lbm sec f t 2
| l i N
u.n
10 K*
-
-
f s
w ^
ï^ 1
*
- . . . 1 „ . . . . , 1 i
\\\
i • n « i u n i n i n i n IUO I U ?.« 22m 2:00 KM ax jot» L-JO
P c(psia)
F i g . 3 .16 - Moody 1965 ( P a r . 3 . 3 . 3 . 5 )
STAGNATION PROPERTIES
v s
ENTRANCE PROPERTIES
< x , >
r
• ; • /
'J ISENTROPIC u / ENTRANCE
STATION 1
LIQUID
VAPOR
EXIT PROPERTIES
< x 2 > r = r
STATION 2
F i g . 3 .17 - Moody 1966 ( P a r . 3 . 3 . 3 . 6 ) .
1000 3000
Po(psia)
F i g . 3 .18 - Moody 1966 ( P a r . 3 . 3 . 3 . 6 )
p .Cpsia)
F i g . 3 .19 - Moody 1966 ( P a r . 3 . 3 . 3 . 6 )
i i - 1 T
at Z)
VA
PO
— i 1 1 — i —
i i i i o o o
(eisd) od o o
< Z g CO
Z LU
O O tn
m
Kl
1
1 X
TU
RE
i
S
A
ii
i i — i — i
^ s ''S
Y jy
1 1 1 . . . .
—71—
jp^ 3-""""
0
i i i
v— -
-
-
1
§ (eisd)Od
o
o o m
r-LLI
o*;-\— _ l
< Z o ^ CO Z LU
2 O
S j Q
—~ CD
- * - » *— CT (/)
K I I 4->
vO O C7> i—1
O O
<« 1
o r^i
to
ci •H
- I —
o
1
LIQ
M i i
"T - f"
/ *2-
i i
1 1 1 1
^J^^^ ^ 5 _ —
i i j i
o o o (eisd) od
o o
§i=
g CO
I
Oatg Initial Press,, psiq o lo» (100-500) a lnte»med. (700-1200) A High (1200 1800)
-CÔ-L/D=40-
I Reqion
m
Fig ,
? 4 6 8 10 12 14 16
3 . 2 1 - Fauske 1964 ( P a r . 3 . 3 . 3 . 7 )
18 20 L/D
w4 r
eg •P
O 10 >\r
x> 10* r
10
. '
-
r
• ' r-'•
•
1 ' 1 '
X/ «V / ^ v
° \ / / J \ /
£^
1 i 1 •
r •,T T—» r -r-ri
*O~~*-~L//
1 . 1 . 1 . 1
«5"'T~T'"r
5 0 6 0 ^
. » i . >
i
70
• ! ' !
SO 90 i 1 . 1
' I
100,
1
• :
1
: •
-•
» 10 11 12 13
BTU/lbm-10" 2
F i g . 3 .22 - Fauske 1964 ( P a r . 3 . 3 . 3 . 7 )
.001 0.01 0.1 1.0 STEAM QUALITY, WEIGHT FRACTION
F i g . 3 . 2 3 - Levy 1965 ( P a r . 3 . 3 . 3 . 8 )
1000
P fpsia)
100
10
. I I I
"
-
• r
/ f
I i I
' ,
/ /
/
I
X=0.9
•
I
—r-r -
1 1
-rr
JLL
l i
i i
f ""
X-0 .4
...1
,1 l
,l.,
-
1 100 200 400 700 1000 2000 r, 4000
( l b m / s e c f t 2 )
F i g . 3 . 2 4 - L e v y . 1 9 6 5 ( P a r . 3 . 3 . 3 . 8 ) .
I3UU
ho
btu | Ibm
1000
700
400
100 1 01
I00r~
9or
80^
70
•p e
u.
ioV "
50V^
4oY
3(
THE ENDPOIN I 5 1
\ ~ ~ \ .
2oNr**^\s^
1 *sT\ T IS X=104 -^*-
02 !
* \ ^ ^ 2 0 0
/V&OC
1000
' 5 0 0 , *
« * , #
^ 50 <?*
^ = - 1 0 2 0
I 5 1'03 - 5 1 0 4 '
T f i bm/sec f t 2 )
CRITICAL / / POINT 3206 psia
F i g . 3.25 - Cruver e t a l . 1967 (Par . 3 . 3 . 3 . 9 ) .