Вы находитесь на странице: 1из 16

Nuclear Engineering and Design 235 (2005) 457472

Steam generator tube rupture (SGTR) scenarios


A. Auvinen
a
, J.K. Jokiniemi
a,
, A. L ahde
a
, T. Routamo
b
, P. Lundstr om
b
,
H. Tuomisto
b
, J. Dienstbier
c
, S. G untay
d
, D. Suckow
d
, A. Dehbi
d
, M. Slootman
e
,
L. Herranz
f
, V. Peyres
f
, J. Polo
f
a
VTT Processes, Biologinkuja 7, P.O. Box 1602, VTT Espoo 02044, Finland
b
Fortum Nuclear Services, Vantaa, Finland
c
Nuclear Research Institute Rez plc, Czech Republic
d
Paul Scherrer Institute, Villigen-PSI, Switzerland
e
Nuclear Research and Consultancy Group, Arnhem, Netherlands
f
Centro de Investigaciones Energ eticas, Medioambientales y Tecnol ogicas, Madrid, Spain
Received 30 March 2004; received in revised form 6 May 2004; accepted 31 August 2004
Abstract
The steam generator tube rupture (SGTR) scenarios project was carried out in the EU 5th framework programme in the
eld of nuclear safety during years 20002002. The rst objective of the project was to generate a comprehensive database
on ssion product retention in a steam generator. The second objective was to verify and develop predictive models to support
accident management interventions in steam generator tube rupture sequences, which either directly lead to severe accident
conditions or are induced by other sequences leading to severe accidents. The models developed for ssion product retention
were to be included in severe accident codes. In addition, it was shown that existing models for turbulent deposition, which is
the dominating deposition mechanism in dry conditions and at high ow rates, contain large uncertainties. The results of the
project are applicable to various pressurised water reactors, including vertical steam generators (western PWR) and horizontal
steam generators (VVER).
2004 Elsevier B.V. All rights reserved.
1. Introduction and research objectives
Steam generator (SG) reliability and performance
are serious concerns in the operation of pressurised

Corresponding author. Tel.: +358 9 456 6158;


fax: +358 9 456 7021.
E-mail address: jorma.jokiniemi@vtt. (J.K. Jokiniemi).
water reactors. In particular, steam generator tubing
is subject to a variety of degradation processes that
can lead to tube cracking, wall thinning, and potential
leakage or rupture (MacDonald et al., 1996). Over the
last decade, a considerable effort has been spent to
understand these degradation processes and to improve
related preventive and corrective actions as well as
operational aspects. However, steam generator tube
0029-5493/$ see front matter 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.nucengdes.2004.08.060
458 A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472
Nomenclature
AM accident management
AMMD aerodynamic mass median diameter
BC base case
BLPI Berner low-pressure impactor
d
t
tube diameter
D tube equivalent diameter
DF decontamination factor
ECCS emergency core cooling system
EFW early feedwater recovery
EOP emergency operation procedure
FBG uidised bed generator
FW feedwater
GSD geometric standard deviation
MSGTR multiple steam generator tube rupture
N
bins
number of particle bins
N
tubes
number of tubes over which deposition
is considered
NC non-condensable
NPP nuclear power plant
OPC optical particle counter
p absolute pressure
p
cold
pressure in cold collector
p
hot
pressure in hot collector
PSA probabilistic safety assessment
PWR pressurised water reactor
Re Reynolds number
s distance between tubes
SEM scanning electron microscope
SG steam generator
SGCB steam generator collector break
SGTR steam generator tube rupture
SRV safety relief valve
Stk Stokes number
TEOM tapered element oscillating microbal-
ance
VVER Russian-type pressurised water reactor
WP work package
y(k) mass fraction of particles in size class k
Greek letters

ST
(i, k) single tube retention efciency for par-
ticle size k

tbt
ST
single tube turbulent deposition ef-
ciency

imp
ST
single tube inertial impaction efciency
ln(
g
) logarithm of the geometric standard de-
viation
leakage incidents have proven that such occur-
rence cannot be completely ruled out. If a steam
generator tube ruptures during a severe accident,
radionuclides may leak from primary circuit to the
secondary side and bypass the containment. According
to most probabilistic safety assessment (PSA) studies,
a signicant fraction of ssion products are assumed
to ow through an unisolated break in an SG. The
assumption is based on an expert elicitation panel,
since no experimental data of this phenomenon is
available to verify it (USNRC, 1990).
General knowledge on retention of ssion products
in the steam generator tubes and in the secondary
side was poor at the beginning of the project. Most
previous experimental programs have concentrated
on the initial stages of deposition inside tubes. Much
less attention has been paid to situations where the
deposition/resuspension/revaporisation changes as
the deposit layers build up as occurs under expected
accident conditions (Wright, 1994). The understanding
of ssion product retention under realistic steam gen-
erator conditions is needed in order to design efcient
accident management procedures. This is considered
very important, since steam generator tube ruptures
are included in the risk-dominant sequences. Thus, the
rst objective of the SGTR project was to generate a
comprehensive database on ssion product retention in
a steam generator. The second objective was to verify
and develop predictive models to support accident
management interventions during SGTR sequences.
The severe accident management procedures are de-
signed to minimise the release from the defected SG.
Current accident management actions foresee ooding
of the secondary side through the emergency feedwater
system and depressurisation of the primary system in
an attempt to suppress the release of ssion products.
These actions may signicantly reduce the source term
in SGTR of accidents. However, there has been no ap-
propriate database or associated model to estimate it. A
strategic goal of the project was to demonstrate the ef-
fectiveness of the accident management interventions
A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472 459
in reducing the source term even for severe accidents
that lead to a bypass of the containment. The results of
the project are applicable to various pressurised water
reactors, including vertical steam generators (western
PWR) and horizontal steam generators (VVER).
2. Work programme
The project was carried out as parallel studies for
vertical and horizontal steamgenerators. The work was
divided into four work packages (WP). The description
of each WP is given below.
The work in WP1 included the denition of impor-
tant steamgenerator tube rupture sequences for the ver-
tical steam generators (SG) of PWR and the horizontal
SG of VVER-440 and determining the experimental
conditions.
The experimental investigations (WP2) were di-
vided into integral and separate effect studies. The in-
tegral experiments for vertical steam generators were
conducted at PSI in the ARTIST facility, which is a
scaled-down model of the Framatome type SG in op-
eration at the Swiss NPP Beznau 1136 MWth PWR.
HORIZON facility at Fortum was used in the integral
studies of horizontal steam generators. The facility is
a scaled-down model of VVER-440 steam generator
installed in Loviisa NPP (Finland) and Dukovany NPP
(Czech Republic). Aerosol retention mechanisms act-
ing at the secondary side near the break exit were stud-
ied in PECA facility at CIEMAT. Aerosol deposition
and resuspension in the primary side of the SG was
studied in PSAERO facility at VTT. The main objec-
tive in WP2 was to develop integral and separate effect
databases to assess the capability of the accident man-
agement measures, and to develop and verify associ-
ated mathematical models.
The objective of the WP3 was to develop physi-
cal models capable of predicting the local deposition
phenomena in the primary and secondary sides of the
SG. The aim was to build up a simple model, which
could predict the experimentally observed deposition
phenomena in steam generator tube rupture sequences
as a function of ow rate, tube rupture size, aerosol
concentration, aerosol size, etc.
The objective of the WP4 was to include the models
developed in WP3 into the system level codes, and to
see their effect on the aerosol retention in SGTR cases.
In addition to the study of aerosol retention, sensitivity
calculations should assess the effectiveness of different
accident management measures in the reference plants.
Two state of the art reports were also a part of the
work. The rst one described the present knowledge
of aerosol deposition mechanisms in tubes and tube
bundle. The second report was an overview of present
status of SGTR events and procedures in the design
basis and beyond design basis domains for the relevant
power plants.
3. Work performed and results obtained
3.1. State of the art report
The state of the art in SGTR accidents was writ-
ten in two reports (Jokiniemi et al., 2002; Bakker and
Slootman, 2002). The rst report describes the most
important deposition mechanisms in SGTR cases. It
was concluded that even though the deposition mecha-
nisms acting on SGTRscenarios are known, their mag-
nitude and importance in different SGTR conditions is
not understood. Thus, experimental data were needed
to justify the relevant deposition mechanisms and their
magnitude.
The second report gives a general description of
SGTR events both within design basis and beyond de-
sign basis situations. The international projects and lit-
erature on the SGTR phenomena is discussed in the
report. The present way of handling the occurrence
of SGTR events within Borssele, Beznau, Loviisa and
Dukovany NPPs is also reviewed. This includes EOP
and SAM responses. For the Borssele, Beznau and
Dukovany NPPs, the EOP and SAMG responses are
based on the same principles and include the same main
actions. For Loviisa NPP, the structure of the EOPs and
the actions are somewhat different. The EOPs of Lovi-
isa are currently undergoing a wide revision.
3.2. Accident scenarios and boundary conditions
(WP1)
The rst task included analyses of important SG
tube rupture sequences using integral codes. The main
objective was to obtain a basis for the denition of ex-
perimental conditions for the ARTIST and HORIZON
experiments. The second task included the denition
460 A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472
Table 1
MELCOR calculations of the conditions in the faulted vertical SG during ssion product release
SGTR scenario Timing of
interest (1000 s)
Pressure (bar) Steam temperature (K) Flow through SG
breaks (g/s)
SG Water
level
Primary Secondary Primary Secondary Steam Aerosols
Base case (BC) 178188 30 3.1 500 500 1300 1.18 0
3-Tube failure (3T) 89100 18 5.0 475 475 2000 5.16 0
Feedwater recovery (FW) 178180 8 1.1 450 450 600 0.67 0
Early FW recovery (EFW) 139143 33 3.3 500 500 1500 0.05 0
of the experimental boundary conditions based on the
results of the rst task. The nal test matrices for the ex-
perimental studies were determined using the obtained
boundary conditions together with the constraints of
the experimental facilities and project objectives.
3.2.1. Accident scenarios and boundary
conditions of vertical steam generators
For the vertical type SG, the work was performed
mainly by NRG using MELCOR 1.8.3 code. NRG
with the help of PSI extended the existing MELCOR
model of the Beznau NPP. In addition to this SC-
DAP/RELAP5 calculations were performed by PSI in
order to simulate a beyond design basis SGTRaccident
for the Beznau NPP.
The most dominant SGTR accident scenario for
vertical steam generators was based on information
from the Beznau and Borssele PSAs, and on discus-
sions between PSI and NRG. The chosen base case
(BC) scenario presented a single-tube guillotine type
tube failure with a consequentially stuck open safety
relief valve (SRV). The scenario included the avail-
ability of accumulators, of high-pressure emergency
core cooling system (ECCS) until the tank was empty
and of reactor coolant pumps until voiding. Steam
dump at the intact SG was performed on basis of start
of core heat-up. Recovery of feedwater to the failed
SG was assumed after extensive core damage. In this
scenario, high pressure, high steam temperature and
large masses of ssion product were produced. The
resulting boundary conditions, presented in Table 1,
should be considered as maximal values.
To study the effect of a larger break in the SG, a sec-
ond worst-case scenario (3T) with the assumption of a
three-tube failure was analysed. This scenario resulted
in lower pressure in primary side and higher pressure in
secondary side. The owof steamand aerosols through
the break was also higher.
More prototypical accident scenarios for determi-
nation of the experimental boundary conditions were
two scenarios with earlier Feedwater recovery. In the
FW case, the feedwater is recovered after failure of
the top part of the core. In the fourth scenario (EFW),
the feedwater system is recovered after heat-up of the
top part of the core, which also triggers the perfor-
mance of the steam dump. The boundary conditions
based on these two scenarios are also presented in
Table 1.
3.2.2. Accident scenarios and boundary
conditions of horizontal steam generators
The accident scenarios of the horizontal steam gen-
erators were mainly calculated by NRI using MEL-
COR 1.8.3 code. The calculations were based on PSA
Level 2 study performed for the Loviisa NPP. Several
steam generator tube rupture (SGTR) including mul-
tiple tubes rupture (MSGTR) cases were analysed for
Loviisa NPP. For Dukovany NPP steam generator col-
lector break (SGCB) scenarios were analysed. The dif-
ferences between Loviisa NPP and Dukovany NPP in
pressuriser and in secondary system valve design and
settings and in operation procedures were taken into
account. SGCB scenarios for Loviisa NPP had been
calculated earlier by VTT (Pekkarinen, 1996).
The base case scenario was dened by NRI and For-
tum as a double-ended break of the uppermost tube of
the sheet at the hot primary collector. All safety sys-
tems were assumed to work normally. Operator actions
were taken according to Loviisa emergency operating
procedures (EOPs). As a secondary failure, the stuck
open SG safety valve was assumed. Primary system
depressurisation procedure started at the onset of core
damage.
In other scenarios the break location, size and
accident management procedures were varied. The
overviewandthe range of the boundaryconditions have
A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472 461
Table 2
Main results for the horizontal SG accident scenarios
Scenario Average pressure
(bar)
Average steam
temperature (K)
Average ow through break (g/s)
a
# Description
b
Primary Secondary Primary
c
Secondary Steam Aerosols CsI
LO1 Loviisa SGTR stuck-open
SG relief valve
28 1.5 473573 473 150 0.06 0.006
LO1A Like LO1, no primary de-
pressurisation
5095 2 6731073 573873 2000 1 0.1
LO2 Like LO1, break location
cold collector
28 1.5 453523 473 150 0.03 0.003
LO3 Like LO1, break location
other loop
28 1.5 473573 473 200 0.04 0.004
LO5 Main steam isolation and
feedwater failure
210 1.2 473623 473 200 0.04 0.004
LOM1 Loviisa MSGTR Like LO1,
LO2 tubes
210 1.5 453623 453 250 0.2 0.02
LOM2 Like LO1, LO5 tubes 1.59 1.5 473673 453 150 0.7 0.07
DUC1 Dukovany SGCB
stuck-open relief valve to
atmosphere
24 23 473673 453 10004500 6 0.9
DUC2 Stuck-open SG relief valve 27 26 473773 473 10004000 7 0.6
a
From the onset of core damage. The average ow is given, in reality, the ow of gases and aerosols varies by about a factor of 10. It is much
higher at the start of ssion product release.
b
SG primary and secondary side water level zero or negligible for all scenarios.
c
In the hot collector, cold collector for LO2.
been obtained based on these analyses and are shown
in Table 2.
The main purpose of the analyses was to obtain typ-
ical conditions for integral experiments of horizontal
steam generator. Some general conclusion can also be
made from the analyses. Primary system depressurisa-
tion was found to be a very efcient in reducing ssion
product release to the environment. Upon depressuri-
sation, the release was reduced by factor of about 20.
The reason for this is that the opening of the pressure
relief valve opens a path for radioactive material to the
containment instead of releasing it through the break
to the secondary system and to the environment.
3.3. Experimental investigations
3.3.1. Integral studies of vertical SG
The integral tests of vertical steam generators were
conducted at PSI in ARTIST facility, which is a rep-
resentative scaled-down model of Beznau reference
PWRsteamgenerator. The facility consists of a bundle,
shroud, ooding system and aerosol sampling stations.
Only the bundle section of ARTIST (including a break
stage, two far-eld stages and a U-bend section) was
used in this project. Apicture of the facility is presented
in Fig. 1.
The ARTIST test section was directly connected to
DRAGON aerosol generation facility. Aerosols were
produced via uidised bed aerosol generators (FBG) in
conjunction with a venturi injection system. In this pro-
gram, prefabricated TiO
2
powder was used with a pri-
mary particle size of 0.035 m (AMMD). The aerosol
mixture was transported to ARTIST test section by car-
rier gas composed of steam and non-condensable gas
(N
2
) in desired proportions.
A sophisticated aerosol measurement system was
attached at the inlet and outlet piping to characterise
the aerosol particle size and concentration as well as
the gas ow rates, the gas pressure and gas/water tem-
perature. The aerosol characterisation was performed
using state-of-the-art instruments. Two photometers
provided relative aerosol concentration in real-time
at the inlet and outlet. The size distributions were
measured with Berner low-pressure (BLPI) and
Andersen impactors, and the integral concentration
measurements were performed with membrane lters.
Amore detailed description of the facility can be found
in the SGTR project deliverable (Dehbi et al., 2000).
462 A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472
Fig. 1. Photo and a schematic representation of the ARTIST test section.
The actual conditions for the experimental tests
were derived from the RELAP5/SCDAP calculations
(G untay et al., 1999) and from report by Bakker
(2001a), while keeping in mind the practical limits im-
posed by ARTIST facility. Five tests comprised the PSI
EU-SGTR experiments. The test matrix is shown in
Table 3. The rst three tests dealt with the aerosol re-
tention in the break stage under dry (A01 and A05)
and wet (A02) conditions. Test A05 was a repetition
of the test A01. The other two tests addressed accident
management (AM) issues, whereby the SGbundle goes
from a fully dry state to a fully ooded state. Test A03
was performed with a non-condensable (NC)-rich car-
rier gas, while test A04 was performed with a steam-
rich mixture. An axis-symmetric guillotine break was
used and located 300 mm above the tube sheet in the
middle of the bundle. The aerosol AMMDs at the inlet
were 2.253.70 m, while at the outlet, the AMMDs
were in the range 0.490.84 m.
The following conclusions can be drawn from the
investigations of integral effects in the ARTIST 4 m
height scaled bundle (Table 4):
With dry bundle and full ow representing the break
stage conditions, there is strong evidence that the
TiO
2
aerosols (AMMD24 m, 35 nmprimary par-
ticles) disintegrated into much smaller particles be-
cause of the sonic conditions at the break. The
breakup promoted particle escape from the sec-
ondary and lowered the overall DF, which was typ-
ically small, i.e. between 2.5 and 3. Further investi-
gation is needed in order to determine the inuence
of the type of aerosol on the disintegration process.
With dry bundle, and small owreproducing the far-
eld velocities (test A03), the DF was of the order
of 5, implying better decontamination than with the
full ow. This could be explained by the somewhat
lower particle disintegration than witnessed with the
larger ow. The far-eld retention implied a DF of
the order of 1.9 per stage, which, for SG with 9 or
more stages, can translate in overall DF of several
hundreds, when the break is located near the tube
sheet.
With a bundle ooded just above the break and a
steam/non-condensable mixture (test A02), the DF
was between 45 and 112 for the full ow and 482
for the small ow (typical of far-eld). This implied
again that the far-eld stages are more efcient at
trapping aerosols than the break stage.
For the far-eld conditions, under a ooded bun-
dle and in the presence of steam (test A04), the DF
was roughly of the same order regardless of the wa-
ter height, i.e. in the range from 482 to 1081. A
large fraction of the aerosols was scrubbed at the
A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472 463
T
a
b
l
e
3
E
x
p
e
r
i
m
e
n
t
a
l
m
a
t
r
i
x
f
o
r
t
h
e
i
n
t
e
g
r
a
l
t
e
s
t
s
o
f
v
e
r
t
i
c
a
l
s
t
e
a
m
g
e
n
e
r
a
t
o
r
s
T
e
s
t
P
r
i
m
a
r
y
s
i
d
e
S
e
c
o
n
d
a
r
y
s
i
d
e
G
a
s

o
w
r
a
t
e
(
k
g
/
h
)
N
C
m
a
s
s
f
r
a
c
t
i
o
n
(
%
)
S
e
c
o
n
d
a
r
y
s
i
d
e
A
e
r
o
s
o
l
s
i
z
e
r
a
n
g
e
(
A
M
M
D
)
(

m
)
P
r
e
s
s
(
b
a
r
)
T
e
m
p
e
r
a
t
u
r
e
(

C
)
P
r
e
s
s
(
b
a
r
)
T
e
m
p
e
r
a
t
u
r
e
(

C
)
W
a
t
e
r
h
e
i
g
h
t
(
m
)
W
a
t
e
r
t
e
m
p
e
r
a
t
u
r
e
(

C
)
A
0
5
,
A
0
1
a
(
d
r
y
)
4
.
6
2
3
1
.
0
2
3
6
5
0
1
0
0
N
/
A
N
/
A
1

3
A
0
2
(
w
e
t
)
4
.
0
1
4
5
1
.
1
9
0

9
5
3
4
0
3
0
1
.
3
9
0

9
5
1

3
A
0
3
(
d
r
y

w
e
t
A
M
N
C
-
r
i
c
h
)
1
.
2

1
.
6
2
0

4
0
1
.
0
2
0

4
0
1
1
0
1
0
0
0

3
.
8
2
0

4
0
1

3
A
0
4
(
d
r
y

w
e
t
A
M
s
t
e
a
m
-
r
i
c
h
)
1
.
2

1
.
5
1
7
0
1
.
1
9
0

1
6
0
b
6
3
2
3
0

3
.
8
9
0

9
5
1

3
a
A
e
r
o
s
o
l
m
e
a
s
u
r
e
m
e
n
t
d
e
v
i
c
e
s
p
a
r
t
l
y
f
a
i
l
e
d
.
b
A
t
d
r
y
s
t
a
r
t
i
n
g
c
o
n
d
i
t
i
o
n
s
.
break level because of strong diffusiophoresis and
impaction of the incoming jet on the water interface.
The additional water head beyond the break stage
had only a secondary inuence on the magnitude of
decontamination.
For the far-eld conditions, under a ooded bundle
and in the absence of steam (test A03), the DF in-
creased exponentially from 124 to 5739, when the
water height in the bundle increased from 1.30 to
3.6 m. The aerosol removal rate was roughly con-
stant with height, and hence the DF was solely a
function of residence time in the water pool (water
height).
When steam was present in the carrier gas under
ooded secondary (tests A02 and A04), condensa-
tion inside the tube caused aerosol deposition and
produced blockage near the break, with subsequent
increase in the primary pressure. This has implica-
tions for real plant conditions, as aerosol deposits
inside the broken tube will cause more ow to be
diverted to the intact tubes, with corresponding re-
duction in the source term to the secondary.
3.3.2. Separate effect studies of vertical SG
The separate effect studies of vertical steam gener-
ators were conducted in the CIEMAT PECA facility,
which was properly modied and conditioned for that
purpose (Peyres et al., 2002). The aim of these stud-
ies was to determine aerosol deposition into the break
stage near the break location.
The PECA facility set-up consisted of gas and
aerosol injection lines, the vessel with tube mini-bundle
and associated instrumentation and sampling stations.
A schematic picture of PECA facility is presented in
Fig. 2. The test mini-bundle was a scaled mock-up of
the rst stage of the steam generator tube bundle. It
consisted of a squared arrangement housing inside a
total of 117 tubes plus four supporting rods placed in
the corners. The mini-bundle allowed two possible lo-
cations of the broken tube. One place was just at the
centre of the structure and the other place was three
tubes away from the centre.
There was one sampling at the injection line for the
optical particle counter (OPC) aimed at determining
the aerosol size distribution and quantifying the mass
concentration at the inlet. Within the vessel atmosphere
eight samplings were taken to six lters and two cas-
cade impactors, fromwhich the mass concentration ex-
464 A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472
Table 4
Results of tests conducted in the ARTIST facility
Type Test AMMD (m) Water level
above tube
sheet (m)
DF Phenomena
In Out Disintegration Impaction Diffusiophoresis Rise-zone
inertial
removal
Break stage
Dry A05 2.25 0.72 0 2.52.9 X X
Wet A02 3.14 0.84 1.30 45.7112 X X X
Far-eld
AM NC-rich A03 3.70 N/A 0 4.9 X
1.20 124 X X
2.30 1251 X X
3.60 5739 X X
AM steam-rich A04 2.92 0.49 0 4.6 X
1.33 482 X X X
2.55 1081 X X X
3.80 514 X X X
iting the tube mini-bundle was estimated. In addition to
this, the mass deposited on several selected tubes was
collected and extrapolated to estimate the total aerosol
mass in the mini-bundle.
The design of the experimental matrix came from
the analysis of the prototypical boundary conditions
estimated with MELCOR and SCDAP/RELAP5
codes and practical limitations imposed by the facility
(Pekkarinen, 1996; Bakker, 2001b). The boundary
conditions were room temperature and inlet pressure
of 2.8 bar. The carrier gas was air and the aerosol
product used was prefabricated TiO
2
particles. The
experimental matrix covered two types of break (guil-
lotine and sh-mouth), two possible location of the
break (central and periphery), two possible break ori-
entations (facing tube and facing diagonal), and three
different inlet gas ow rates (75, 150 and 250 kg/h).
With the sh-mouth break type, two different broken
areas were also covered. These were sh-mouth 1D
and sh-mouth 0.5D, where D denotes the tube equiv-
alent diameter. The experimental matrix is shown in
Table 5.
Table 5
Test matrix for the separate effect experiments of vertical steam generators
Test Break type Break location Break orientation Gas ow rate (kg/h)
Fish Guillotine Central Periphery Facing tube Facing diagonal 75 150 250
1 X
a
X X
2 X
a
X
c
3 X
b
X X
4 X
b
X X
5 Repetition of test 2
6 X
b
X X
7 X
b
X X X
8 X
b
X X X
9 X
b
X X X
10 X X X X
11 X X X X
12 X X X X
a
0.5D sh mouth.
b
1.0D sh mouth.
c
Reduce ow rate to a value at which ow velocity is equal to that of test #4.
A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472 465
Fig. 2. Schematic representation of the PECA facility.
The main result of the separate effect studies was a
rather low global retention in the mini-bundle. Within
the range of boundary conditions tested, the reten-
tion was always below 20% (Peyres et al., 2003). The
amount of retention was much lower than estimated for
particle deposition by inertial impaction. Deposition of
aerosols on the tubes was not uniform along the mini-
bundle. The nearest tubes to the break showed deposits
forming crusts while the outer tubes showed a thin layer
of aerosols. As presented in Fig. 3, the highest amount
of deposit was found from the mini-bundle with the
lowest ow rates 75 and 100 kg/h. Evidently, deposit
removal by erosion and resuspension signicantly in-
uenced the results with higher ow rates.
The aerodynamic mass median diameter (AMMD)
of particles decreased from6 mat the inlet to approx-
imately 3 m at the outlet of the facility. The decrease
in the particle size indicated that upon collision with
the surfaces agglomerated TiO
2
particles fragmented.
This result also indicated that deposit erosion by parti-
cle impaction was an important process near the break
location.
Important differences between break type and orien-
tation were found only at the lowest owrate (75 kg/h).
In case of a guillotine break, the pattern showed a
square symmetry. The tubes up to third neighbours col-
lected almost the 70%of the mass retained by the mini-
bundle. In sh-mouth tests, the deposition patterns had
466 A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472
Fig. 3. Retention in the mini-bundle vs. inlet gas ow rate.
a triangular shape, where the tubes located far from the
break point had a low individual deposition. However,
far-eld tubes taken together represented up to 40% of
the total retention in the mini-bundle.
3.3.3. Integral studies of horizontal SG
Integral experiments of horizontal steam genera-
tor were conducted in HORIZON facility, which is a
scaled-down model of horizontal SG used in VVER-
440 (Fig. 4). The objective of the studies was to gather
Fig. 4. A photograph of HORIZON facility.
data on aerosol behaviour in the primary side of the SG
tubes in different ow conditions.
The experiments carried out in the HORIZONfacil-
ity are shown in Table 6. Aerosols used in the experi-
ments were generated by vaporising CsI in the vertical
high-temperature ow reactor. Depending on experi-
ment, a ow of steam or air carried aerosol particles
through the reactor into a ow mixer. In the mixer
aerosol ow was mixed with superheated main steam
ow. Measured particle size (AMMD) ranged from0.8
to 2.7 m and the geometric standard deviation (GSD)
of the size distribution from1.4 to 1.6 at the inlet of the
SG (see Table 7).
The fraction of deposited aerosols was measured
online by tapered element oscillating microbalance
(TEOM) during the experiments. After the experi-
ments, the amount of deposited particles was deter-
mined by chemical analysis of the lters and SG tubes.
In addition, calculations based on thermalhydraulic
results and on aerosol AMMD were carried out. These
results are summarised also in Table 7.
The results of aerosol deposition on the primary side
of the horizontal SG were compared with the values
obtained from the calculations with the existing de-
position models. It appeared that the current models
were adequate at Reynolds numbers (Re) below 5000,
but gave too high deposition velocities at Re above
A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472 467
T
a
b
l
e
6
E
x
p
e
r
i
m
e
n
t
s
o
f
H
O
R
I
Z
O
N
f
a
c
i
l
i
t
y
c
a
r
r
i
e
d
o
u
t
w
i
t
h
i
n
S
G
T
R
p
r
o
j
e
c
t
E
x
p
e
r
i
m
e
n
t
S
p
e
c
i
a
l
c
o
n

g
u
r
a
t
i
o
n
C
o
n
d
i
t
i
o
n
s
D
e
s
c
r
i
p
t
i
o
n
C
s
I
v
e
s
s
e
l
M
i
x
e
r
N
u
m
b
e
r
o
f
t
u
b
e
s
A
e
r
o
s
o
l
m
e
a
s
u
r
e
m
e
n
t
f
r
o
m
t
h
e
h
o
t
c
h
a
m
b
e
r
a
p
h
o
t
(
k
P
a
)
p
c
o
l
d
(
k
P
a
)
S
t
e
a
m

o
w
r
a
t
e
(
g
/
s
)
F
u
r
n
a
c
e
s
t
e
a
m

o
w
r
a
t
e
(
g
/
s
)
A
i
r

o
w
t
h
r
o
u
g
h
f
u
r
n
a
c
e
(
l
p
m
)
b
T
u
b
e
R
e
y
n
o
l
d
s
n
u
m
b
e
r
0
1
-
A
I
I
I
I
3
8
I
2
5
0
2
5
0
5

6
0
9
4
0
A
e
r
o
s
o
l
e
x
p
.
,
S
G
T
R
,
i
n
t
a
c
t
t
u
b
e
s
0
4
-
A
I
I
c
I
I
I
3
8
C
2
7
0
2
7
0
2
3
1
d

3
8
0
0
A
e
r
o
s
o
l
e
x
p
.
,
S
G
C
B
0
6
-
A
I
I
c
I
I
I
1
C
4
0
5
2
3
0
2
5
1

1
4
0
0
0
0
A
e
r
o
s
o
l
e
x
p
.
,
S
G
T
R
,
d
e
f
e
c
t
t
u
b
e
s
0
7
-
A
I
I
I
V
2
C
2
9
0
2
3
0
2
5
1

7
0
0
0
0
A
e
r
o
s
o
l
e
x
p
.
,
S
G
T
R
,
d
e
f
e
c
t
t
u
b
e
s
0
8
-
A
I
I
I
V
2
C
3
1
0
2
5
0
2
5

6
0
7
1
0
0
0
A
e
r
o
s
o
l
e
x
p
.
,
S
G
T
R
,
d
e
f
e
c
t
t
u
b
e
s
0
9
-
A
I
I
I
V
3
8
C
2
5
0
2
5
0
2
4

6
0
4
1
0
0
A
e
r
o
s
o
l
e
x
p
.
,
S
G
C
B
,
s
e
c
.
s
i
d
e

o
o
d
e
d
a
C
:
h
o
t
c
h
a
m
b
e
r
;
I
:
i
n
l
e
t
o
f
t
h
e
h
o
t
c
h
a
m
b
e
r
.
b
I
n
n
o
r
m
a
l
c
o
n
d
i
t
i
o
n
s
(
p
=
1
0
1
3
2
5
P
a
;
T
=
2
7
3
.
1
5
K
)
.
c
A
m
e
t
a
l
l
i
c
p
i
e
c
e
a
d
d
e
d
a
b
o
v
e
t
h
e
C
s
I
v
e
s
s
e
l
t
o
h
e
a
t
u
p
t
h
e
s
t
e
a
m
b
e
f
o
r
e
e
n
t
e
r
i
n
g
t
h
e
v
e
s
s
e
l
.
d
M
o
m
e
n
t
a
r
i
l
y
s
e
t
t
o
3
g
/
s
.
Fig. 5. A schematic picture of the PSAERO facility.
70 000 compared to the experimental results. Turbu-
lent impaction was considered to be the main deposi-
tion mechanism at high Re. However, the effect of re-
suspension, which became signicant also at high Re,
should be added to the models. In all experiments, the
deposited aerosol fraction per unit length had a peak
value at the tube bend, which conrmed the relative
importance of particle inertia on deposition. Aerosol
deposition was signicantly increased, when the sec-
ondary side was ooded with water. Nevertheless, still
most of the aerosol injected into the tubes was trans-
ported as aerosol out of the tubes. Therefore, ooding
was considered to inuence mainly the secondary side
retention mechanisms such as pool scrubbing.
3.3.4. Separate effect studies of horizontal SG
Separate effect experiments of horizontal steam
generators were conducted in PSAERO facility. The
separate effect experiments were designed to com-
plement the integral experiments conducted with the
HORIZON facility. The objective of the experiments
was to gain mechanistic understanding about aerosol
behaviour in the steam generator tubes.
In PSAERO facility (Fig. 5), the aerosol behaviour
was studied in a straight 3-m long stainless steel tube.
The inner diameter of the tube was 13 mm and the
length of the measured section was 2 m. Aerosol de-
position and the movement of the deposited material
was determined by activating the aerosol in a nuclear
reactor and applying sequentially placed scintillation
detectors in online measurements. The nal deposition
prole was obtained after the experiment by scanning
the facility with a similar gamma detector. When inac-
tive aerosol was used in an experiment, the deposition
was determined by sampling with quartz bre lters
from the inlet and outlet of the steam generator tube.
468 A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472
Table 7
Aerosol size distribution in the hot chamber of the HORIZON facility and the amount of deposited aerosols in the tubes as fraction of the amount
of injected into the tubes
Experiment Particle size distribution
in the hot chamber
Deposited fraction
AMMD (m) GSD TEOM Filter Chemical analyses Calculated
a
01-A 2.7 1.5 0.30 0.32 0.23 0.21
04-A 1.4 1.6 0.20
b
0.51 0.03 0.03
06-A 1.9 1.6 0.90 0.97 0.19 1.00
07-A 12 N/A N/A 0.94 0.23 0.94
c
08-A 0.8 1.4 0.25 Err.
d
0.10 0.25
09-A 1.3 1.6 N/A N/A 0.25 N/A
a
Calculated with the real particle size assuming density of 4510 kg/m
3.
b
Point estimate made from 290 to 320 min.
c
Selected value corresponds to particle size of 0.67 m.
d
Results in negative value.
In the experiments, polydisperse copper particles
were used as aerosol source. In last experiments, cop-
per aerosol was coated with either dry or liquid NaOH.
With a scanning electron microscope (SEM), the par-
ticles were observed to be separate, nearly spherical
and dense. The particle size distribution was deter-
minedwithBerner low-pressure impactors (BLPI). The
aerosol mass size distribution was bimodal and it could
be presented as a sum of two lognormal distributions.
The smaller peak, with aerodynamic mass median di-
ameter (AMMD) of 0.66 m and logarithm of the ge-
ometric standard deviation (ln
g
) of 0.65, contained
17% of the aerosol mass. The AMMD of the larger
peak was 8.31 m and ln
g
was 0.88. Aerosol mass
size distribution did not change during the experiments.
The test matrix for the resuspension experiments
is presented in Table 8. In PSAERO experiments, the
aerosol was always deposited with a constant gas ow
rate. In the rst two experiments, deposition and resus-
pensionphenomena were studiedusinga highowrate.
From experiment 3 onwards, the ow rate during the
deposition phase was low. After the deposition phase
the gas ow rate was increased stepwise and the de-
position prole in the tube was measured online using
radioactive tracer. The effect of material properties on
the depositionresuspension phenomena was studied
by modifying the surface of the particles in experiment
5 and by changing gas composition in experiment 6.
As a result, the mass of particles in the tube was ob-
tained as a function of time and location. Spatial reso-
lution of the developed online technique was 2 cm and
time resolution was 5 s. From that information, local
resuspension rate into a pure gas ow could be cal-
culated. Several conclusions could be made from the
results:
Very little aerosol deposited in experiments con-
ducted with a constant high gas ow rate. With ow
rates of 100 and 200 l/min particles hitting the sur-
face mainly bounced back to the gas stream. The
result was very similar to that found in HORIZON
and PECA experiments. Also, even if the ow rate
Table 8
The test matrix for the separate effect studies of horizontal steam generators
Experiment Dep. ow
rate (l/min)
Resusp. ow
rate (l/min)
Inlet Re Gas inlet
temperature (

C)
Gas Aerosol material Concentration
(g/m
3
)
1 200 200 15500 218 N
2
Cu 7.1
2 100 100 7700 224 N
2
Cu 5.6
3 38.6 40100 468011700 22 N
2
Cu 10.4
4 55.6 6080 69809310 22 N
2
Cu 6.7
5 60 6080 44905980 252 N
2
Cu +NaOH 4.1
6 60 6090 44106660 239 N
2
+H
2
O Cu +NaOH 3.6
A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472 469
during the deposition phase was low, less than mono-
layer of particles remained on the surface, when the
ow was increased up to 100 l/min.
When the ow rate during the deposition phase was
decreased, the amount of particles deposited on the
surface increased. According to deposition mod-
els, the main deposition mechanisms were turbulent
eddy impaction and settling. However, the deposi-
tion prole in the tube was much smoother than that
estimated with the deposition models. The probable
reason for the discrepancy was that the impaction
of large particles must have caused erosion. A sig-
nicant fraction of the already deposited particles
were knocked of from the surface and subsequently
deposited further downstream. The total amount of
deposit was also signicantly underestimated by tur-
bulent impaction models, which use dimensionless
particle relaxation time as parameter (Papavergos
and Hedley, 1984). It seems that the effect of ow
Reynolds number should be taken into account in
the modelling of turbulent impaction. Also, the high
density of particles applied in the experiments may
have inuenced the results more than estimated by
the model.
The ow rate during the deposition phase had a
very signicant inuence on the strength particles
adhered to the deposit layer. Particles deposited in
a higher ow rate were much harder to resuspend
than was the case with a lower ow rate. It is likely
that particles impacting on the surface packed the de-
posit and increased the number of contacts between
particles. Therefore, particles depositing with higher
momentum would result in a stronger deposit struc-
ture. In a higher ow rate, the deposit layer was also
subjected to stronger removal forces. Thus the min-
imum force, in which particles could adhere to the
surface, increased with the ow rate.
In most experiments, deposit removal took place
mainly immediately after the ow rate in the tube
was increased. In such occasion, a small fraction of
particles was resuspended into pure gas stream near
the inlet of the tube. The fraction of removed de-
posit increased further downstream as the number
of particles in the gas ow increased. It is thus likely
that the resuspended particles caused erosion of the
deposit layer. In most experiments, erosion was the
dominant removal mechanism, probably because the
particles were rather large.
In experiment 5, copper particles were coated with
dry NaOH, which made the surface of the particles
rough. Increased surface roughness decreased the
adhesion force of the particles. Unlike in other ex-
periments, signicant resuspension took place in ex-
periment 5 even during a constant low gas ow rate.
Thus resuspension was strongly time-depended. Par-
ticles resuspended from the surface also deposited
further downstream. As a result, the deposition pro-
le in the experiment had a wavelike form. The
peak of the wave moved slowly downstream during
the experiment. In other experiments, the amount
of deposit decreased exponentially as the distance
from the tube inlet increased. The difference in the
deposition prole was likely due to different parti-
cle removal mechanisms. In experiment 5, the main
removal mechanism was resuspension into the gas
stream, whereas in other experiments particles were
mainly removed by erosion.
Results fromthese experiments were very well com-
parable to previous studies on resuspension con-
ducted with polydisperse aerosol (Biasi et al., 2001).
However, experiments with monodisperse aerosol
without an exception result in a much higher resus-
pension. It is likely that the adherence of polydis-
perse aerosol is much better, because particles in the
deposit layer have more contacts to other particles
than is the case with monodisperse aerosol. A major
problem in resuspension modelling is that the effect
of polydispersity is not taken into account. How-
ever, the diameter of a particle is customarily a very
important parameter in these models. Therefore, pa-
rameters derived from experiments, conducted with
monodisperse particles, should be used with caution
in models describing the behaviour of polydisperse
aerosol.
3.4. Model development
3.4.1. Aerosol deposition model for the primary
side of the steam generator
In this task, a one-dimensional steady-state model,
AERORESUSLOG (Ludwig, 2002), for aerosol de-
position in the primary side of steam generator tubes
was developed. Based on the most important accident
scenarios, the turbulent impaction, thermophoresis and
gravitational settling were considered to be most im-
portant deposition mechanisms. These mechanisms are
470 A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472
fairly well understood and validated with experimental
data. In contrary to deposition, the mechanisms respon-
sible for resuspension are less well understood, and
more difcult to predict. In order to calculate resuspen-
sion, a quasi-static moment balance model by Reeks
and Hall (2001) was selected for AERORESUSLOG
code. The model seemed to be in reasonable agree-
ment with previous experimental results (Biasi et al.,
2001).
In the model, the ow eld of steam and/or air is
calculated using correlations for heat and mass trans-
fer. The mass concentration of the aerosol entering the
tube is a lognormal mass-size distribution divided into
different size classes. Particle deposition due to ther-
mophoresis, turbulent impaction and gravitational set-
tling is calculated in each size bin. Thereafter, the re-
suspension fraction of deposited particle mass is esti-
mated and the mass concentration for each size bin is
corrected with the net deposition.
Integral experiments of horizontal steam gen-
erators were calculated with AERORESUSLOG.
The calculated results were in good agreement with
measured values at low (Re =940) or intermediate
Reynolds numbers (Re =3800), where no resuspension
is expected. However, the values calculated at high Re
(70 000140 000) including resuspension were incon-
sistent with the experimental results from HORIZON
and PSAERO experiments. The reason for different
values is most likely that the resuspension model is
developed for less than monolayer coverage of parti-
cles on a surface rather than for a deposit layer. It does
not take into account several important parameters
such as erosion, dependence of adhesion on deposit
physicochemical properties, polydisperse particle
size distribution and system geometry. Thus, a better
model is needed to describe resuspension in SGTR
conditions.
3.4.2. Aerosol deposition model for the secondary
side of the steam generator
In this task, a model to calculate aerosol deposition
in the near-eld of tube breach under dry conditions
was developed. The model is based on lter concept,
which means that aerosol owing through a bundle of
obstacles is submitted to forces that tend to clean up the
gas by removing particles onto obstacle surfaces. Two
major hypotheses lie under this approach. First, gas
is seen as a viscous uid owing transverse to tubes.
Second, ltration is considered uniform at any plane
perpendicular to incoming gas ow direction.
Under foreseen SGTR conditions, the major
deposition mechanisms in the near-eld are turbulent
deposition and inertial impaction. The former domain
extends over a Stokes number (Stk) ranging from 0
up to 0.1. From this upper bound to higher Stokes
numbers, inertial impaction becomes dominant. A
database to develop individual models for turbulent
eddy deposition and inertial impaction was set up
based on literature survey (Wong and Johnstone,
1953; Ilias and Douglas, 1989). More than a hundred
experimental measurements were compiled and
from them the following expressions for single tube
ltration efciencies were derived:

tbt
ST
= 4.38 10
2
+7.13 10
2
ln(Stk) (1)

imp
ST
=
0.75
1 +29.31 exp(3.85Stk
0.5
)
(2)
Using these expressions, total retention efciency in
the near-eld of the tube breach was

TB
= 1 exp

4d
t
(d
t
+s)
4(d
t
+s)
2
d
2
t

1 +(1)
N
tubes
+1

N
tubes

N
bins

k
y(k)
ST
(i, k) 1

(3)
in which N
tubes
is the number of tubes over which
deposition is considered (i.e., ltration depth),
ST
(i, k)
the individual efciency of a single tube i for particles
of size k, N
bins
the number of particle size classes, y(k)
the mass fraction of particles in size class k, d
t
the tube
diameter and s the distance between tubes.
The equations were incorporated into a highly mod-
ular FORTRAN-90 code, called ARISG-I (aerosol re-
tention in steam generators). One of the key variables
in the assessment of lter efciency of the tube bun-
dle was shown to be the gas velocity. As ARISG-I is
just a rst step forward in the modelling, major contri-
butions to further develop the model would be needed
on aspects such as: in-bundle gas velocity, individual
mechanisms responsible for aerosol deposition in the
far-eld of SG and removal of deposited particles from
surfaces. In addition, a more extensive validation exer-
cise of the model should be carried out.
A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472 471
3.5. Plant evaluations
3.5.1. Plant evaluations of the vertical steam
generator
Plant evaluations for the vertical SGwere performed
based on the calculations done in WP1 and the experi-
ence received during the SGTRproject. The purpose of
these MELCOR calculations was to determine the ef-
fect of the developed ssion product retention models
on the calculated ssion product releases to the envi-
ronment, and the impact of accident management mea-
sures.
In the cases studied, the impact of the newly de-
veloped models on the calculated release was rather
small. New models predicted higher deposition to the
primary side of the broken SG, which was compensated
by decreased deposition to the secondary side. How-
ever, it was shown that the developed models could be
implemented into MELCOR using control functions.
The models could be applied also in other scenarios, in
which retention might be more signicant.
Accordingtothe calculations, accident management
measures inuenced not only the deposition of ssion
products, but also the thermal-hydraulics, the sequence
of the events and thus the ssion product behaviour. It
was observed that early injection of feedwater to the
brokenSG, highfeedwater owrate andits highlevel in
the brokenSGdecreasedthe release tothe environment.
3.5.2. Plant evaluations of the horizontal steam
generator
A two-tier approach was used in the plant eval-
uations of horizontal SG. The plant analysis was
performed rst with MELCOR 1.8.3 neglecting the re-
tention inside the broken tube and on the near-eld tube
bundle secondary side. Particle retention was then stud-
ied in detail using the models developed in WP3 with
the help of MELCOR 1.8.5 and modied AERORE-
SUSLOG codes (Dienstbier, 2003). Besides of LO1
(single tube break) and LOM2 (5-tube break) scenarios
(Table 2), providing no retention on the primary side
of the SG tubes, similar scenarios were analysed with
break location at half of the SG tube length.
The results indicate that the retention inside the bro-
ken tube would probably be very small compared to
the effect of depressurisation as mentioned in Section
3.2. Although turbulent deposition velocity was high,
mechanical resuspension removed most deposited par-
ticles. The effect of deposition near the break location
in the secondary side could be described by a moderate
decontamination factor (DF) of about 1.41.7.
An important accident mitigation factor that takes
place, even if no accident management is taken into ac-
count, is the aerosol deposition on the secondary side
far-eldtubes andSGshell. This phenomenonwas con-
rmed for vertical type SG in the ARTIST experiment
(Dehbi et al., 2003), and it was included in all plant cal-
culations for horizontal SG (Dienstbier and Duspiva,
2000; Dienstbier, 2003).
4. Conclusions
The objective of the SGTRproject was to provide an
experimental database of aerosol particle retention in
SGTR sequences. The work included development of
simple steady-state models, which were applied in the
estimation of ssion product retention into the steam
generator. These models were incorporated into system
level code MELCOR. A number of SGTR scenarios in
reference PWR and VVER-440 plants were studied in
model calculations. In these calculations, the effective-
ness of different accident management strategies was
assessed in this kind of accidents.
The SGTR project made an important step forward
in resolving uncertainties of various physical models,
especially regarding aerosol mechanical resuspension.
Work to nd a more exact prediction of the effect
of aerosol retention in the steam generator during
an SGTR accident should be continued. The project
highlighted areas where future work should be con-
centrated. These include more focused, separate effect
studies of aerosol retention in the break stage and
far-eld stages, including the effects of thermophoresis
and aerosol material. Extension of the investigations to
upper structures (separator and dryer) is also advisable
and would allow a thorough understanding of aerosol
phenomena in the whole steam generator. Lastly, in
order to model particle removal from the surfaces,
dynamic models for deposition and resuspension
should be developed.
References
Bakker, P.J.T., 2001a. MELCOR analysis of Beznau SGTR accident
scenarios determination of boundary conditions for vertical SG
experiments within EU-SGTR Project. SAM-SGTR-D010.
472 A. Auvinen et al. / Nuclear Engineering and Design 235 (2005) 457472
Bakker, P.J.T., 2001b. Steam generator tube ruptures: experimental
boundary conditions. SAM-SGTR-D005.
Bakker, P.J.T., Slootman, M.L.F., 2002. Overview Report on
Operational Aspects of SGTR Accidents. SAM-SGTR-
D011.
Biasi, L., de los Reyes, A., Reeks, M.W., de Santi, G.F., 2001. Use
of a simple model for the interpretation of experimental data
on particle resuspension in turbulent ows. J. Aerosol Sci. 32,
11751200.
Dehbi, A., Suckow, D., G untay, S., 2000. Integral tests in a vertical
steam generator bundle: description of the test matrix. SAM-
SGTR-D006.
Dehbi, A., Suckow, D., G untay, S., 2003. Integral tests in a vertical
steam generator. SAM-SGTR-D024.
Dienstbier, J., Duspiva, J., 2000. SGTRscenarios calculation results.
SAM-SGTR-D002.
Dienstbier, J., 2003. SGTR scenarios results including new aerosol
retention models and the assessment of accident management.
SAM-SGTR-D027.
G untay, S., Birchley, J., Suckow, D., Dehbi, A., 1999. Aerosol trap-
ping in a steam generator (ARTIST): an investigation of aerosol
and iodine behavior in the secondary side of a steam generator.
In: Proceedings of the 27th Water Reactor Safety Information
Meeting, Bethesda, November.
Ilias, S., Douglas, P.L., 1989. Inertial impaction of aerosol particles
on cylinders at intermediate and high Reynolds numbers. Chem.
Eng. Sci. 44 (1), 8199.
Jokiniemi, J., Ludwig, L., Herranz, L., 2002. Review of aerosol de-
position mechanisms which may be relevant in SGTRconditions.
SAM-SGTR-D021.
Ludwig, W., 2002. AERORESUSLOG: model for depositionresus-
pension in the primary side of a steam generator in tube rupture
accidents. SAM-SGTR-D022.
MacDonald, P.E., Shah, V.N., Ward, L.W., Ellison, P.G., 1996. Steam
generator tube failures. NUREG/CR-6365, INEL-95/0393.
Papavergos, P.G., Hedley, A.B., 1984. Particle deposition behaviour
from turbulent ows. Chem. Eng. Res. Des. 62, 275295.
Pekkarinen, E., 1996. Evaluation of the ssion product release to en-
vironment using the MELCOR-code in four Loviisa containment
bypass sequences. VTT Energy Research Report ENE4/41/96.
Peyres, V., Polo, J., Herranz, L.E., 2002. PECA facility conditioning
and set-up for the SGTR Project. SAM-SGTR-D007.
Peyres, V., Polo, J., Herranz, L.E., 2003. Separate effect studies of
vertical steam generators. SAM-SGTR-D023.
Reeks, M.W., Hall, D., 2001. Kinetic models for particle resuspen-
sion in turbulent ows: Theory and measurement. J. Aerosol Sci.
32, 131.
USNRC, 1990. Severe accident risks: an assessment of ve U.S.
nuclear power plants, vol. 2. Sandia National Laboratory Report
NUREG-1150.
Wong, J.D., Johnstone, H.F., 1953. Engineering experimental station.
University of Illinois, Technical Report number 11.
Wright, A.L. (Ed.), 1994. Primary System Fission Product Release
and Transport. NUREG/CR-6193.

Вам также может понравиться