ML20209E517
| ML20209E517 | |
| Person / Time | |
|---|---|
| Site: | Oconee  |
| Issue date: | 04/30/1985 |
| From: | Boyack B, Dearing J, Henninger R, Nassersharif B LOS ALAMOS NATIONAL LABORATORY |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-A-7228 LA-10351-MS, NUREG-CR-4140, NUDOCS 8506240647 | |
| Download: ML20209E517 (112) | |
Text
-
- LA-10351-MS
+.
I_
+
t Dominant Accident Sequences in Oconee-IPressurized Water Reactor a...,,, w o,> u o o o a o jipp ADC)CV 0 9 0 0 ','i
e
(]@ //1\\c \\ 65' W}'y] @ Los AlamosNntional Laboratory
<xab@) /r JX1 Ll i+) f) Los Alamos.Nm Mexico 87545 l
w l
l
L.-
W'
- *9 __
r An AMyunadve Aminn/Egust Opputunky Employw :
i.
Prepared by Jean Martinez and Wendy Rowley, Group Q-9 t
NOTKE Tus w ees papesed as as sessem d eek epsessed by a aguery dihe Unised Seesse C _
Peseber es Websd teesse C.
ese est essary deseuf, er any W ihmer sagtsyest ashee say oerremy, espeesed er haphmL w eesumme any heel Embey a soupsenhey Ist any thed puny) ess.e es seenho d ash mes.d any -
apparsam pseduse er pousse dhsissed h alde supost a sepaussse ihm he ses by mush IMrd pony would eel bdMage phendy eseed eghet t
...m
NUREG/CR-4140 LA-10351-MS R4 Dominant Accident Sequences in Oconee-l Pressurized Water Reactor J. F. Dearing R. J. Henninger B. Nassersharif Compiled by B.E.Boyack B. Nassersharif Manuscript submitted: January 1985 Date published: April 1985 Prepared for Division of Accident Evaluation Office of Nuclear Regulatory Research Us Nuclear Regulatory Commission Washington, DC 20555 NRC FIN No. A7228 S86m_Ld _
n @ Los Alamos NationalLaboratory
.O U G/ Los Alamos,New Mexico 87545
C0KfENTS LIST OF FIGURES...........................................................
vii LIST OF TABLES............................................................
x GLOSSARY..................................................................
xi ABSTRACT..................................................................
1 1.
INTRODUCTION........................................................
1 II.
SELECTION OF DOMINANT SEQUENCES.....................................
3 III.
SYSTEM DESCRIPTION..................................................
6 IV.
LOFW SEQUENCES........................................s.............
13 A.
Base Case.......................................................
14 B.
Normal Plant Response...........................................
15 C.
Operator Actions for Accident Management........................
20 D.
Equipment Studies...............................................
26 E.
Conclusions for L0FW............................................
30 V.
SMALL-SMALL BREAK L0CA..............................................
33 A.
Introduction....................................................
33 B.
Assumptions Used During Accident Progression....................
34 C.
TRAC Calculations of Accident Sequences.........................
34 1.
Break Size of 102-mm (4.02-in) Diameter.....................
34 2.
Break Size of 31-mm (1.2-in) Diameter.......................
39 3.
Break Size of 22-mm (0.87-in) Diameter......................
43 4.
B r e a k S i ze o f 11 -mm ( 0. 4 3-i n ) D i ame t e r......................
43 D.
Extensions of TRAC Calculations to once-Through Water Limit...........................................................
47 1.
Solution of Global Continuity and Energy Equations..........
50 2.
Extensions of TRAC Solutions................................
50 E.
Conclusions for Small-Small Break L0CA..........................
52 VI.
LOFW-I NITI ATED TRANSI ENT WI T110VT SCRAM..............................
54 A.
Base Case.......................................................
55 B.
Normal Plant Response...........................................
60 C.
Equipment Studies...............................................
64 D.
Conclusions for ATWS............................................
70 VII.
INTERFACING SYSTEMS L0CA............................................
70 A.
Introduction....................................................
70 B.
Base Case.......................................................
71 C.
Normal Plant Response...........................................
73 D.
Operator Intervention (No LPIS).................................
74 E.
Operator Intervention (No LPIS and one HPIS)....................
74 F.
Conclusions for Interfacing Systems L0CA........................
76 VIII.
SUMMARY
CONCLUSIONS.................................................
80 V
~.
REFERENCES................................................................
81 APPENDIX A. TRAC VERSION..................................................
82 I.
INTRODUCTION...................................................
82 A.
TRAC Characteristics.......................................
83 1.
Variable-Dimensional Fluid Dynamics....................
83 2.
Nonhomogeneous, Nonequilibrium Modeling................
83 3.
Flow-Regime-Dependent Constitutive Equation Package................................................
84 4.
Comprehensive lleat-Transfer Capability.................
84 5.
Consistent Analysis of Entire Accident Sequences..............................................
84 6.
Component and Functional Modularity....................
84 B.
Physical Phenomena Treated.................................
85 C.
Significant Changes from Previous TRAC Versions............
85 D.
Planned Improvements.......................................
86 II.
TRAC VERSIONS USED IN THIS STUDY...............................
86 APPENDIX B. SENSITIVITY OF RESULTS TO UNCERTAINTIES.......................
87 APPENDIX C. TIMING STATISTICS.............................................
88 APPENDIX D.
INPUT M0DELS..................................................
89 I.
TRAC-PF1 Oconee Model for T (B )MLU Sequence Transients........
89 3
3 II.
TRAC-PF1 Oconee Model for S 11 Sequence Transients..............
91 3
III.
TRAC-PF1 Oconee Model for T KMU Sequence Transients............
93 2
IV.
TRAC-PF1 Oconee Model for V Sequence Transients................
93 vi
LIST OF FIGURES 1.
Oc o n e e - 1 dom i na n t a c c i d e n t s e q u e nc e s.............................. 4 2.
Oconee-1 RCS......................................................
9 3.
Oconee-1 LPIS.................
.................................. 10 4.
Oconee-1 IIPIS....................................................
12 5.
Core average pressure'during base case LOSP-induced LOFW transient..................................................
16 6.
Core average liquid and saturation temperatures during base case LOSP-induced LOFW transient....................
16 7.
Core liquid-volume fraction during base case LOSP-induced LOFW transient.....................................
17 8.
Maximum cladding temperature of an average rod during base case LOSP-induced LOFW transient....................
17 9.
Core average liquid and saturation temperatures during LOSP-induced LOFW event with normal plant response.......
19 10.
Core average liquid and saturation temperatures for feed initiated at 120 s follcwing a LOSP-induced LOFW event......................................................
21 11.
Primary pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event.......................
23 12.
Primary coolant temperature for feed and bleed initiated at 120 s following a LOSP-induced LOFW event..........
23 13.
Cold-leg mass flow for feed and bleed initiated at 120 s following a LOSP-induced LOFW event....................
25 14.
Primary temperature and pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event.......... 25 15.
Primary temperature and pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW esent (throttling begins at 3600 s)...................................
27 16.
Primary temperature and pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event
( PORV s i z e t w i c e a c t u a l )........................................ 2 7 17.
Core average pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event (one llPI pump).......................................................
29 18.
Core average liquid and saturation temperatures for feed and bleed initiated at 120 s following a LOSP-induced LOFW e ve n t ( o n e llPI p um p )....................................... 29 19.
Primary pressure for auxiliary-feedwater recovery at 900 s following a LOSP-induced LOFW event....................
32 20.
Primary-coolant liquid and saturation temperatures for auxiliary-feedwater recovery at 900 s following a LOSP-induced LOFW event.......................................
32 21.
Mass flow out 102 -mm-d i a m ( 4 - i n. ) b r e a k.......................... 36 22.
Mass flow in single llPI line, 102-mm-diam (4-in.) break case.....
37 23.
Core average liquid temperature, 102-mm-diam (4-in.) break case.. 37 24.
Pressurizer pressure. 102-mm-diam (4-in.) break case.............
38 25.
Core liquid volume fraction. 102-mm-d i am (4-i n. ) break case...... 38 26.
Low-pressure-injection mass flow, 102-mm-diam (4-in.) break case. 39 27.
Mass f l ow o u t 31 -mm-d i am ( 1. 2 - i n. ) b r e a k......................... 40 38.
Mass flow in single llPI line, 31-mm-diam (1.2-in.) break. case.... 41 29.
Core average liquid temperature. 31-mm-diam (1.2-in.) break case. 41 vii
l l
1 30.
Pressurizer pressure, 31-mm-diam (1.2-in.) break case.............
42 31.
Core l iqu id volume f rac t ion, 31-mm-diam ( 1. 2-i n. ) break case...... 42 32.
Mass f l ow ou t 2 2 -mm-d i a m ( 0. 87-i n. ) b r e a k......................... 44 33.
Mass flow in single llPI line. 22-mm-diam (0.87-in.) break case.... 44 34.
Core average liquid temperature, 22-mm-diam (0.87-in.)
break case.......................................................
45 35.
Pressurizer pressure, 22-mm-diam (0.87-in.) break case............
45 36.
Core liquid volume fraction. 22-mm-diam (0.87-in.) break case.....
46 37.
Mass flow out 11-mm-diam (0.43-in.) break.........................
47 38.
Mass flow in single llPI line. 11-mm-diam (0.43-in.) break case.... 48 39.
Core average liquid temperature, 11-mm-diam (0.43-in.)
break case.......................................................
48 i
40.
Pressurizer pressure, 11-mm-diam (0.43-in.) break case............
49 41.
Core liquid volume fraction, 11-mm-diam (0.43-in.) break case..... 49 42.
Extension of TRAC results with global solution of mass and energy equations for various break diameters, no llPI throttling...................................................
51 43.
Ef fect of throttling IIPI to maintain 25-K primary subcooling j
on primary pressure for 31-mm-diam (1.2-in.)
break...............
52 44.
Effect of throttling liPI to maintain 25-V primary subcooling on primary average temperature for 31-nm-diam (1.2-in.) break.... 53 45.
Ef fect of throttling IIPI to maintain 25-K primary subcooling on primary pressure for 22-mm-diam (0.87-in.)
break..............
53 46.
Effect of throttling ilPI to maintain 25-K primary subcooling on primary pressure f o r 31 -mm-d i am ( 1. 2-i n. ) br e a k............... 54 47.
Core average pressure during base T KMU transient.................
56 2
48.
Core average liquid and saturation temperatures during base T KMU transient.............................................
56 3
49.
Reactivity feedbacks during base T KMU transient..................
58 2
50.
Reactor power during base T KMU transient.........................
58 2
51.
Core average pressure during T KMU transient with normal 2
s y s t e m r e s p o n s e.................................................. 61 52.
Core average liquid and saturation temperatures during T KMU transient with normal system response......................
61 3
53.
Reactor power during T KMU transient with normal system 2
response.........................................................
62 54.
Reactivity feedbacks during T KMU transient with normal 2
systemresponse..................................................62 55.
Core average pressure during T KMU transient with a 50%
2 reduction i-AFW at 1500 s.......................................
65 i
56.
Core average liquid and saturation temperatures during T KMU transient with a 50% reduction in AFW at 1500 s............
65 2
57.
Reactivity feedbacks during T KMU transient with a 50%
2 reduction in AFW at 1500 s.......................................
66 58.
Reactor power during T KMU transient with a 50% reduction in AFW at 1500 s.... 2............................................
66 59.
Core average pressure during T KMU base transient with 2
failure of the PORV to reclose at 247.5 s........................
67 60.
Core average liquid and saturation temperatures during T KMU base transient with failure of the PORV to reclose 3
af247.5s.......................................................67 61.
Reactor power during T KMU base transient with failure 2
o f t h e PORV t o r e c l o s e a t 2 4 7. 5 s................................ 6 8 viii 4
y_
y,
_-_,..__~w_,
62.
Reactivity feedbacks during T KMU base transient with 2
f a i l u r e o f t he PORY t o r e c l o s e a t 247. 5 s....................... 68 63.
Core liquid volume fraction during T KMU base transient 2
wi t h fa i l u re of the PORY t o rec los e a t 247. 5 s.................. 72 64.
Primary-system pressure during V sequence base and parametric Cases...........................................................
72 65.
lireak flow dur ing V sequence base and pa rame t ric ca ses........... 75 66.
Maximum cladding temperature of average rod during V sequence base and parametric cases.......................................
75 67.
Oconee-1 V sequence event tree...................................
79 68.
Duration of effective core cooling during V sequence base and parametric cases................................................
79 D-1. TRAC noding of B&W Plant for L0FW................................
90 D-2. Schematic of TRAC input model for Oconee-1 small-small break analysis.................................................. 92 D-3. TRAC noding of Oconee-1 for ATWS.................................
94 D-4. TRAC-PF1 noding of Oconce-1 f or the "V" seque nce................. 95 ix t
g, m---
e-e.v-.
.-ww,-
y,ewn,,,--ew--e-~nwr,vv---,,
ua,_a_
l t
f LIST OF TABLES 1
1.
SYMBOLS USED IN FIGURE 1........................................
5 11.
EXPECTED CONSEQUENCES PER RELEASE NORTHEAST RIVER VALLEY COM PO S I T E S I T E................................................
7 t-1I1.
DOMINANF ACCIDENT SEQUENCES AT OCONEE-1......................... 7 IV.
SEQUENCES TO BE ANALYZED WITH TRAC FOR OCONEE-1.................
8 V.
000 NEE-1 REACTOR PROTECTION TRIPS..............................
14 i
VI.
EVENT SEQUENCE FOR LOSP-INDUCED LOFW BASE TRANSIENT (NO HPI)...15 VII.
EVENT SEQUENCE FOR LOSP-INDUCED LOFW EVENT FOR NORMAL PLANT RESPONSE......................................................
18 VIII.
EVENT SEQUENCE FOR FEED INITI ATION AT 120 S FOLLOWING A LOSP-INDUCED LOFW EVENT.......................................
20 IX.
EVENT SEQUENCE FOR FEED-AND-BLEED INITIATION AT 120 S FOLLOWING A LOSP-INDUCED LOFW EVENT...........................
22 X.-
EVENT SEQUENCE FOR FEED-AND-BLEED INITIATION AT 120 S WITH i
1 HPI PUMP FOLLOWING A LOSP-INDUCED LOFW EVENT................ 28 XI.
EVENT SEQUENCE FOR RE W VERY OF AUXILIARY FEEDWATER AT 900 S FOLLOWING A LOSP-INDUCED LOFW EVENT.....................
31 XII.
TIMING OF EVENTS IN FOUR SMALL-SMALL BREAK LOCAS (SECONDS l
A FT E R B R EAK ).................................................. 3 5 XIII.
REACTIVITY DATA FOR A FULL-POWER BEGINNING-OF-CYCLE CORE.......
57 XIV.
EVENT SEQUENCE FOR T KMU BASE TRANSIENT (NO HPI)............... 59 XV.
EVENT SEQUENCE FOR T KMU TRANSIENT WITH NORMAL SYSTEM RESP 0NSE...................................................... 63 I
XVI.
EVENT SEQUENCE FOR T KMU BASE TRANSIENT WITH FAILURE OF 2
TH E PORY TO RECLOSE AT 2 4 7. 5 S................................ 69 XVI I.
INTERFACING SYSTEMS LOCA EVEKr SEQUENCE (BASE CASE)............
73 XVIII.
INTERFACING SYSTEMS LOCA EVENT SEQUENCE (NORMAL PLANT R E S PON S E CA S E )................................................ 7 6 XIX.
INTERFACING SYSTEMS LOCA EVENT SEQUENCE (OPERATOR INTERVENTION CASE--No LPIS)................................................
77 XX.
INTERFACING SYSTEMS LOCA EVENT (OPERATOR INTERVENTION CASE--
NO L P I S AND ONE II P I ).......................................... 7 8 i
f 1
X I
i
..,.__m_
l GLOSSARY 4
AFW Auxiliary Feedwater (same as EFW)
/rWS Anticipated Transient Without Scram B&W Babcock & Wilcox BWR Boiling Water Reactor BWST Borated Water Storage Tank CFS Core Flooding System CFT Core Flooding Tank ECC Emergency Core Cooling ECCS Emergency Core Cooling System EFW Emergency Feedwater EFWS Emergency Feedwater System lillASWS High Head Auxiliary Service Water System ilPI High-Pressure Injection i
ilPls High-Pressure injection System INEL Idaho National Engineering Laboratory LOCA Loss-of-Coolant Accident LOFW Loss of Feedwater LOSP Loss-of-Offsite Power LPI Low-Pressure Injection LPIS Low-Pressure injection System MFWS Main Feedwater System MOV Motor-Operated Valve MTC Moderator Temperature Coefficient NDT Nil-Ductility Transition NRC Nuclear Regulatory Commission ORNL Oak Ridge National Laboratory xi
..~._
.~
PORY Power-Operated Relief Valve PRA Probabilistic Risk Analysis 1
P/T Pressure / Temperature PWR Pressurized Water Reactor 4
Reactor Coolant Pump RCS Reactor Coolant System kIIR Residual Heat Removal I
RSSMAP Reactor Safety Study Methodology Applications Program SASA Severe Accident Sequence Analysis SG Steam Generator SIS Safety-Injection System SNL Sandia National Laboratories SSS Safe Shutdown System SV Safety Valve TMI Three Mile Island TRAC Transient Reactor Analysis Code i
TSV Turbine Stop Valve i
r i
I i
I xii i
, ~, -. _.. _ _ -.
l 2
DOMINANT ACCIDENT SEQUENCES IN OCONEE-1 PRESSURIZED WATER REACTOR by i
J. 17
Dearing,
R. J. llenninger, and B. Nassersharif Compiled by B. E. Boyack and B. Nassersharif ABSTRACT t
A set of dominant accident sequences in the Oconee-1 pressurized water reactor was selected using probabilistic risk analysis methods.
Because some accident scenarios were similar, a subset of four accident sequences was selected to be analyzed with the Transient Reactor Analysis Code (TRAC) to further our insights into similar types of accidents.
The sequences selected were loss-of-feedwater, small-small j-break loss-of-coolant, loss-of-feedwater-initiated transient without
- scram, and
!nterfacing systems loss-of-coolant accidents.
The normal plant response and the impact of equipment availability and potential operator actions were also examined.
Strategies were developed for operator actions not covered in existing emergency operator 4
guidelines and were tested using TRAC simulations to evaluate their effectiveness in preventing core uncovery and i
maintaining core cooling.
L I.
INTRODUCTION The accident at Three Mile Island (TMI) led to the reorientation of Nuclear Regulatory Commission (NRC) research needs.
Attention has been focused on the potential for equipment / instrument failure and its consequences during anticipated t ransients.
Normal recovery procedures can be ineffective during certain severe or multi-fault transients, and operator responses become extremely important in controlling accidents and mitigating their severity.
Therefore, it is critical to identify and understand complex system transients 1
j coupled with multiple plant system failures, accidents that may result in fuel 1
l damage and fission product transport, depressurization accidents, and the safety implications of the man-machine inte-face.
The NRC initiated the Severe Accident Sequence Analysis (SASA) program to improve the unoerstanding of reactor accidents in order to develop better strategies to prevent, manage, and mitigate severe accidents.
Ilest-estimate state-of-the-art computer codes like the Transient Reactor Analysis Code (TRAC) developed by los Alamos National Laboratory are used to gain insight into a spectrum of safety issues.
In addition, the deterministic analyses provide insights about the course of a transient, event timing, the importance of equipment availabilities, and the impact of selected operator actions or inactions.
The SASA studies are performed by four national laboratories: Los Alamos is investigating the early phases of pressurized water reactor (PWR) accidents:
Idaho National Engineering Laboratory (INEL) is studying the early phases of both boiling water reactor (IlWR) and PWR accidents: Oak Ridge National Laboratory (ORNL) is analyzing the earlier and later phases of IlWR accidents and fission product transport, and Sandia National Laboratories (SNL) is studying the later phases of PWR accidents and containment structural response.
The contribution of Los Alamos to the SASA program has required delineation of potentially severe accident sequences at specific nuclear power plants and thermal-hydraulic simulations of the plant response to equipment failure and operator actions during the accident.
These were accomplished by performing computer simulations using TRAC, an advanced, best-estimate computer program for the analysis of accidents in light-water reactors.
TRAC f ea t u re s either a one-or a three-dimensional treatment of the pressure vessel and its associated internals; a two-fluid nonequilibrium hydrodynamics model with a noncondensable gas field and solute tracking; flow-regime-dependent constitutive equation treatment: optional reflood tracking ca pabi l i t y for bottom-flood and falling-film quench fronts; and consistent treatment of entire accident sequences including the generation of consistent initial conditions.
A more detailtd description of TRAC is presented in Appendix A.
This report describes investigations of the dominant accident sequences identified by the Reactor Safety Study Methodology Applications Program (RSSMAP) in the Oconee PWR ' Oconee-1 is a 886 MW Ilabcock and Wilcox (IkW) plant, owned and operated by Duke Power Company, which is located on Lake Keowee, South Carolina.
Iour accident sequences were analyzed: a loss-of-feedwater (LOFW) 2 i
transient; a combined LOFW and loss-of-offsite-power (LOSP) transient:
a small-small break loss-of-coolant accident (LOCA) and the interfacing systems LOCA.
Section 11 discusses the selection of the dominant accident sequences.
A summary description of the Oconee-1 PWR system is presented in Section III.
where the safety systems are described.
Section IV presents the LOFW calculation and the parametrics studied.
A base case calculation is performed to establish event timing.
Three parametrics were calculated to study normal plant response without operator actions.
operator actions for accident management, and the effects of various plant equipment in accident progression and mitigation.
Section V presents the calculated results for a small-small break LOCA.
The accident is initiated by a cold-leg break followed by failure of the high-pressure recirculation system.
Four break sizes were studied:
10.2 cm (4.02 in.),
3.1 cm (1.2 in.),
2.2 cm (0.87 in.),
and 1.1 cm (0.43 in.)
in diameter.
Calculations were performed to determine the time period fo
.sh i c h once-through cooling was available before the high-pressure recirculation system was required to prevent core damage.
Section VI presents the calculated results for a LOFW-init iated t ransient without scram.
A base case and two parametrics were calculated.
Normal plant response and the effects of different plant equipment in accident progression and mitigation were studied.
Section Vil describes the interfacing systems LOCA calculations performed.
A base case was calculated to establish event timing, and the parametrics studied normal plant response and possible operator actions to prolong once.-through cooling of the reactor core.
Section VIII is an overview summary of the insights gained.
Appendix A contains a description of TRAC and the versions of TRAC used in this study:
a sensitivity study of the results to uncertainties: Appendix C, Appendix 11, computer time statistics, and Appendix D.
the input models developed for the various calculations described here.
II.
SELECfl0N OF DOMINANT SEQUlKES i
The probabilistic risk analysis (PRA) studies prepared for the Oconee #3 plant in RSSMAP' were used as the basis for selecting the dominant accident l
sequence for this SASA study.
Figure 1 summarizes the results of RSSMAP study.
The amount of radioactive core inventory released determines the release category." The probability of each event occurring is tabulated in Fig.
1.
i Acronyms used in Fig. I are defined in Table 1.
The probability of core 3
m,. - r
-y
..-..-e.-%.
- - - - - ~
--w.-,..-~e-
,-.s.
.,.--.,ew.-
-.--w,... -. -.. - - - - - - - -
-.. - ~
.=
L C[T$NY 1
2 3
4 5
6 7
7 0x104
- 8.8x 104 (6.0x 104 T MLU 6
TiMLU 71.0x 104
- 1.5x 104 c 1.0x 10-*
v v<4.0x 10-*
Ti(B )MLU 71.1x 10
- 1.6x 10 c 1.1x 10 4
4 4
2 5
4 4
7 5x10
- 8.0x10 c5.5x 10-*
T.MO-l{
7 0x104
- 7.3x 10-e c5.0x 104 Sell 5
SD a6.7x 10 71.3x 10-8
- 4.9x 10 c5.4x104 4
4 T.MO-Fl{
725x 10-*
- 3.7x 10-8 c2.5x 10-*
7 1x10
- 3.1x 10-*
c2.1x 10-*
4 S Fil 2
l S.Fil a 1.3x 10.e
- 9.5x 10-8 c 1.0x 10-*
4 T MLUO 74.1x 10
- 5.9x 10-e c 4.1 x 10
i T.KMU 4
7 9x10
- 5.7x 10-8 c3.9x 104 3
SD a2.0x 104 4
4 74.0x10
- 1.5x 10 c 1.6x 104 I
SD 77.0x 10
- 1.0x 10-e c7,ox 104 4
TiMLUO 72.7x 104
- 3.9x 10-8 c2.7x 10-8 f
4 4
T MLUO 75.5x 10
- 8.0x 10-*
c5.5x 10 4
7 5x10
- 1.1x 10
c7.5x 10 7
4 T MO-D CAT f,1Y 1.1x 104 1.0x 10-*
2.9x 104 9.7x 10-e 4.6x 104 7.3x 104 3.5x 104 10-3_
10-4_
10-5_
10-8_
10-7_
10-8_
1 2
3 4
5 6
7 Release Category i
i Fig.
1.
Oconee-1 dominant accident sequences.
4 I
I l
TABLE I J
SYMBOLS USED IN FIGURE 1 Initiatine Events-Tg - LOSP Transient T2 - L ss f Power Conversion System Transient Caused By Other Than a LOSP T3 - Transients with the Power Conversion System Initially Available S; - Intermediate LOCA (10" < D ( 13.5")
S2 - Small LOCA (4" < D 4 10")
S3
. 9na ll-Sma ll LOCA (D ( 4")
V- - Interfacing Systems LOCA System Failures (B ) - Emergency Power System 3
D
- Emergency Coolant Injection System F
- Containment Spray Recirculation System 11
- Emergency Coolant Recirculation System K
- Emergency Feedwater System, Recovery of Power Conversion System and Iligh llead Auxiliary Feedwater System M
- Power Conversior. System (Normal Operation)
Q
- Reclosure of Pressurizer Safety / Relief Valves U
- Iligh-Pressure injection System 0
- Failure of Reactor Building Cooling System 1
Containment Failure Moses a - Vessel Steam Explosion 0 - Penetration Leakage 7 - Overpressure Due to llydrogen Burning e - Base Mat Melt Through i
5
~
meltdown is estimated to be about 8 x 10-' per year.
Release categories 6 and 7 involve melt progression through the containment floor and result in small consequences as compared with the other release categories, as indicated in Table II.
Therefore, the probability of serious consequences (categories I through 5) is about 4x 10-*/ year.
Categories 1,
4, and 5 have such low probabilities that they contribute a small portion to this probability; if they are neglected, the probability is still about 4x 10-5/ year for meltdown with serious consequences.
If specific accident sequences with probabilities of less than 1 x 10-*/ year are omitted f rom categories 2 and 3, the remaining sequences have a probability of about 3 x 10-'/ year; these sequences, listed in Table III, dominate the risk at Oconee based on the RSSMAP results.
Many of the dominant accident sequences in Table Ill are similar.
Table IV lists a set of sequences which, when analyzed in detail with TRAC. will provide information useful in understanding the progression of all the dominant sequences at Oconee.
Four sequences were selected: T (Il )MLU. LOFW-induced g
3 LOSP; T KMU, LOFW without scram: S 11, small-small break LOCA; and V, interfacing 2
3 systems LOCA.
III.
SYSTEM DESCRIPTION Oconee-1 is a 886 MWe, two-loop, four-cold-leg Il&W PWR which is located on Lake Keowee, South Carolina.
The plant is owned and operated by Duke Power Company.
The construction permit was applied for in November 1966 and received in November 1967.
The full-term license was issued in February 1973, and initial criticality was achieved on April 19 1973.
This was the first large IMW plant (not counting Indian Pt. 1) to go critical.
The reactor coolant system (RCS) of Oconee-1 is shown in Fig. 2.
The RCS is arranged in two heat transport loops, each with two shaft-scaled reactor coolant pumps (RCPs) and one vertical once-through steam generator (SG).
An electrically heated pressurizer is used to maintain the RCS pressure.
Significant design parameters and reactor protection systems of the Oconee-1 plant are described in the remainder of this section.
The overall emergency core-cooling system (ECCS). which is designed to prevent core damage over the entire spectrum of LOCA break sizes, consists of the high-pressure injection system (llPIS), the low-pressure injection system (LPIS), and the core-flooding system (CFS).
6
TABLE II EXPECTED CONSEQUENCES PER RELEASE NORTilEAST RIVER VALLEY COMPOSITE SITE Latent Early Cancer Fatalities Property Damage Category Fatalities per year (10*$)
1 45 118 2130 2
7 67 2440 3
0.4 55 990 4
0 18 340 5
0 6
200 6
0 1
170 7
0
~0 170 f
TABLE III IXMINANT ACCIDENT SEQUENCES AT OCONEE-1 f
Release Probability /
Sequence Category year T MLU 3
1.0 x 10-*
3 V
2
<4x 10 ~*
T (B )MLU 3
1.1 x 10-*
3 3
T MQ-il 3
5.5 x 10"*
2 S 11 3
5.0 x 10~'
3 SD 3
1.3 x 10~'
g T MQ-Fil 2
2.5 x 10-*
2 S Fil 2
2.1 x 10-*
3 T MLUO 3
4.1 x IO"'
2 T KMU 3
3.9 x 10~'
2 T MLUG 3
2.7 x 10**
g 3.3 x 10-'
7
r--
TAllLE IV SEQUENCES To llE ANALYZED WITil TRAC FOR OCONEE-1 V
T (B )MlU j
3 T MO-II 2
T KMU 2
SH 3
SD 3
The CFS is a passive, self-contained, self-actuating system.
It is designed for a large LOCA and floods the core when the RCS pressure drops below 4.24 MPa (600 psig).
The system consists of two independent paths, each consisting of a core-flooding tank (CFT) or accumulator, two check valves, a normally open motor-operated isolation valve in series with the check valves,
and associated piping.
The CFT has a total volume of 39.93 m (1410 ft 3) and a 3
normal borated water (2270 ppm boric acid) inventory of 29.45 m (1040 ft 3).
3 The tank design pressure is 4.93 MPa (700 psig) and is normally pressurized with nitrogen to 4.24 MPa (600 psig).
The two 0.36-m (14-in.) series check. valves (rated at 17.34 MPa (2500 psig)] prevent high-pressure coolant from entering the CFTs during normal operation.
The motor-operated check valve at the tank outlet is fully open during normal plant operation, and its position is indicated in the control room.
Depressurization of the RCS below 4. 24 MPa (600 psig) will cause discharge of borated water directly into the reactor vessel under the driving force of-the pressurized nitrogen in the tanks.
1 The LPIS (Fig. 3) consists of two separate trains for delivering borated i
water to the RCS following a LOCA.
Ilora t ed wa t e r is drawn from the borated water storage tank (llWST) through a single suction header.
The llWST has a total 3
capacity of 1469 m (388000 gal) and contains 2200 ppm borated water.
The LPIS lines connect to the core-flooding nozzles on the opposite sides of the vessel.
8
I lii l@ i i
il 2
a i1
- 4 Y
l 41 o I i
l o
a 3
5 ti y
1" 2,
.r e 1
5 3g.
(
s s
_a e
e a
p e
h 5
g T
i l
t1 l4 a
41 g
o
!* i l!
ll' e
I 9
17.34 MPa/618 K 3.55 MPa/422 K o
(2500 peig/650*F) j (500 Ig/300'P) l Normally Open 424 MPa (600 psig) j Normally Closed Core Flood kePmssun l
1 Safety Valves Pumps j
Exchangersg
_ L92 i
Heat 4
_/
' r3 f
O From i
To j
Borated hf$
j(0254rn) 10" Piping b/
Nozzles j
O Storage 5
OE g-y l
\\3.81 F
em (1.5")
424 MPa f
LJ I
/
VR To Pressurizer Spray Head i
Inside RB Outside RB Containinent Containment Fig. 3.
Oconee-1 LPIS.
c yJ
?
Each flow path can deliver borated water to the reactor vessel at a flow' rate of 0.189 m /s (3000 gpm).
The BWST is isolated from the LPIS pumps during normal 3
When plant operctions by two parallel. normally closed, motor-operated valves.
the RCS pressure falls t o 3. 55 MPa (500 psig) or the reactor building pressure rises to 0.13 MPa (4 psig), the LPIS is sutomatically actuated.
The injection mode continues until the ilWST is approximately 94%. empty at which time a low water level is annunciated in the control room.
Upon receipt of this alarm. the operator must realign the LPIS io' recirculate water from the react'or building sump through the residual heat removal (RilR) heat exchangers and core-flooding
~ '
nozzle into the reactor vessel.
The HPIS (Fig; 4) is designed to prevent core uncovery for small LOCAs (highe r,- pres sur e s ),
and to delay core uncovery for intermediate-sized LOCAs.
The llPIS can also be us'ed for core cooling following a non-LOCA reactor shutdown (feed-mode cooling).
Feed-mode cooling will be utilized only if normal and emergency secondary heat renokalviatheSGsis.notavailable.
Three pumps are available for borated water injection into the'RCS.
Each HPI pump can deliver 0
0.0284 m /s (450 gpm) at 11.82 MPa (1700 psig) and 305 K (90 F).
Borated water 3
is drawn from the BWST through a single suction header and is pumped through injection lines that penetrate the reactor building (RB) on opposite sides.
Each injection line splits into two lines to provide four injection paths to the four RCS cold legs (between the RCP and the reactor vessel inlet nozzles).
The initiated at a low RCS pressure of 10.4 MPa (1500 psig)I or a containment llPIS is overpressure of 0.13 MPa (4 psig).
The llPIS valves will reach full open within 6 s after receiving an actuation signal.
The IIPIS is actuated on an automatic control system signal but will not turn off automatically. 'The HPIS operation must, be manually terminated.
The llPIS may also be started manually from the control room.
The emergency feedwater system (EFWS) is designed to remove the decay heat from the RCS after shutdown when the main feedwater system' (MFWS) is unavailable.
If the EFWS is unavailable,,ihe high head auxiliary service water system (IlllASWS) may be used.
Successful SG cooling can be accomplished by a flow of 0.0315 m /s (500 gpm) supplied by either system.
The EFWS can feed to 3
either or both SGs under automatic or manual initiation and control.
The system consists of two separate feed trains and a combined, suction source with a normal 3
inventory of $68 m (150000 gal).
Each train can supply 0.0341 m /s (540 gpm) 3 if the turbine-driven pump is used or 0.0315 m /s (500 gpm) if the motor-driven 3
11 y
1
.n g
~ - -,. - - - ~ -, -,
7,,.
G N Normally Open N Normally Closed X
/
ofs y
From N
11 r ted old i
MEs N
W ter Storage
(
.-t/=
x x
O i
W~Q y
/
W'
>W Iligh Pressure Inside RB *-+ Outside RB Pumps Containment Containment Fig. 4.
Oconee-1 IIPIS.
4 pumps are used.
The EFWS is automatically initiated by either low MFW discharge pressure f rorr.rboth MFW pumps or pump trip signals f rom both pumps.
The operator can monitor FFWS flow, discharge pressure, and cooling water flow for all EFWS throughphant instrumentation in the control room.
Both the EFWS and MFWS pumps rep'res$nted as tabular boundary conditions of flow versus pressure or time are in the TRAC model.
The lillASWS is a subsystem of the Oconee safe shutdown system (SSS).
The a backup if all othen f eedwate r systems are unavailable.
It lillASWS provides single 0.142 m /s (2250 gpm) motor-driven pump that has the 3
consists of a capacity to provide adequate shutdown cooling to the SGs of all three Oconee units simultaneously.
The lillASWS AC power is generated by the SSS and does not rely on other power systems utilized at the Oconee station.
The IIIIASWS is initiated by remote manual control.
This system was not modeled in the TRAC calculation.
The react'r protection trips are listed in Table V.
A subset of this list o
was used for each transient calculation.
IV.
LOFW SEQUENCES We first examine a series of LOFW transients.
These transients are initiated by a LOSP which results in a loss of the power conversion system (T M) g followed by a failure of the emergency AC sources (11 ).
The loss of both the 3
normal and emergency AC sources causes a station blackout.
Failure to recover the power conversion system and failure of the emergency feedwater and high-head auxiliary service water systems (L) will lead to SG dryout when the initial secondary inventory is evaporated.
Finally, a failure of the llPIS (U) is assumed.
More details concerning this sequence can be found in Ref.
1.
In the actual plant, the LOSP terminates the normal feedwater, which rapidly decays to zero as the main feedwater pumps coast down.
We have assumed that the main feedwater flow dropped to zero instantly at the start of the transient to determine the minimum time to a key event during a T (B )MLU 3
3 transient, the time of SG dryout.
A diagram of the noding scheme used to model the plant plus other modeling details can be found in Appendix D.I.
13
TABLE V OCONEE-1 REACTOR PROTECTION TRIPS Trip Setpoint Comment Reactor scram Low primary-system Control rods drop pressure (0.5 s insertion (13.20 MPa, 1915 psia) time)
AFW on Low primary-system pressure (11.14 MPa, 1615 psia)
HPI on Low primary-system 30 s delay pressure from EOC actuation (11.14 MPa, 1615 psia) signal RCPs off Same as llPI signal 60 s delay 90 s delay after HPI actuation Accumulator Pressure differential check valves across the valve open AP < 50 kPa (7.3 psia)
LPI on Low primary-system pressure (1.45 MPa, 210 psia)
A.
Base Case The first LOFW sequence examined was the dominant sequence, identified as T (B )MLU in the Oconee RSSMAP study.
The event sequence for this base 3
3 transient is given in Table VI.
Closure of the turbine stop valves (TSVs) and interruption of main feedwater resulted in secondary-system pressurization to 7.23 MPa (1049 psia) and opening of the steam-line secondary safety valves (SVs) at 2.1 s.
Failure of the EFWS and lillASWS lef t the SGs unable to remove decay heat by 70 s.
This is shown in Figs. 5 and 6, which give the average pressure and liquid temperatures in the core region.
With the loss of heat sink the primary begins to heat up to saturation.
Expansion of the primary coolant resulted in the power-operated relief valve (PORV) opening at 494 s.
(8.2 min),
followed by filling the pressurizer steam space at 950 s (15.8 min).
Steam 14
TABLE VI EVENT SEQUENCE FOR LOSP-INDUCED LOFV BASE TRANSIEI.T (No HPI)
Time Event (s)
(min) 0 0
LOSP.
TSVs close RCPs coast down Loss of all feedwater 0.5 Trip reactor 2.1 Secondary SVs open
'70 1.2 SGs ineffective as heat sink 494
- 8. 2 PORY opens
"*600 10.0 SGs empty 950 15.8 Pressurizer full 1654 27.6 ECC signal on containment overpressure (HPI fails) 1800 30.0 Voiding in the core 2200 36.7 Primary SVs open 2500 41.7 Tops of candy canes voided 3000 50.0 Rapid heating of core 3400 56.7 Core empty 3750 62.5 End of calculation flowing from the primary system through the PORY pressurized the containment building to 0.129 MPa (4 psig) at 1654 s (27.6 min).
This resulted in an emergency core-cooling (ECC) actuation signal at 1654 s (27.6 min).
In this sequence, we assumed that the high pressure injection (HPI) system failed to deliver any water.
Without IIPI wa t e r, a s can be seen in Fig. 7, boiling began in the core region at 1800 s (30 min), and by 2200 s (36.7 min) the core region and upper portions of the primary system were saturated.
Saturated expansion of the primary coolant opened the primary SVs (Fig. 5).
At 3000 s (50 min) the fuel rod temperature, which had been close to the liquid saturation temperature corresponding to the relief-valve setpoint pressure, began to increase rapidly as the fuel rods uncovered as seen in Fig. 8, and at 3400 s (56.7 min) the core was empty.
The calculatien was terminated at 3600 s (60 min).
B.
Normal Plant Response This section discusses the sequence in which the plant responds as designed with automatic actuation of safety equipment, but with no operator action.
Compared to the base case transient, this sequence assumes the recovery of emergency AC power to operate the HPIS.
The event sequence is given in 15
m.
a.
PRIMARY SRW open 17.S -
f g
W
-2500 87 -
p Lie
.,A
.g.
p l
j E
v v
i w
w l
w.s--
Saturated expansion
--2400 m
m W
a:
g-n.
w
-2300 w
o a
4 4
5 is.s -
5
-2200 W
g e-8 o
o u
M.S -
O ineffbellye
--2l00 84 0
900 1000 1500 2000 2500 3000 3S00 4000 TIME (s)
Fig. 5.
Core average pressure during base case LOSP-induced LOFW transient.
640
{
3 PORV & SV maintain pressure W
s30-M
, v,./..f,s-c
...sy,..
- M E'I[-[-[
h a
-es0 m
M 620-M N
j 5
g go.
640 z
a c
g LIQUID TEMPERATURE y
3 600--
--- ---- SATURATION TEMPERATURE
--620 l
m 590 -
--600 5
5 4
4 M
L
-u0 M
o
- -SCs ineffective o
o v
570 O
S00 1000 200 2000 2500 3000 3500 4000 TlWE (s)
Fig. 6.
Core average liquid and saturation temperatures during base case LOSP-induced LOFW transient.
16
12 gSystem saturates 5
Eo 0a-Boiling gins E
wg 0.6 -
8>
c 0.4 -
8 3
0.2 -
8o Core empty
~
-02 1000 1500 2000 2500 300C 3500 4000 O
SCO TlWE (s)
Fig. 7.
Core liquid-volume fraction during base case LOSP-induced LOFW transient.
8c0 7
v v
750-
-~*
g y
7 0
5 S
I i
W 700--
--500 g
8 8
7 i
w w
~
70g g
Rods at liquid temperature
~
~ Rods
(
uncovered 2
2 600
-600 E
E SGs ineffective 2
2 550-1000 1500 2000 2500 3000 3500 4000 0
600 TlWE (s)
Fig. 8.
Maximum cladding temperature of an average rod during base case LOSP-induced LOFW transient.
17
-w- - - -
Table VII.
The sequence is the same as the base case discussed in Sec. IV.A until the containment overpressure signal occurs at 1654 s (27.6 min) and !!PI delivery began.
The primary system at 1654 s (27.6 min) was heating up but was still subcooled and liquid full because of liquid expansion.
Two llPI pumps delivered sufficient water to avert saturation make up water lost through the PORV, and cool the core.
This can be seen in Fig. 9, which gives the core average liquid and saturation temperatures.
The system was cooled as follows at 1984 s (33.1 min).
The 305-K (90 -F)
IIPI wa t e r a t a flow of 6.6 kg/s (5.2 x 10' lb,/h) entered each of the four cold legs and mixed with the cold-leg flow of 114 kg/s (9.0 x 10' lb,/h) at 617.5 K (652.5 F).
The temperature resulting from this mixing process was 607.7 K (635 F).
This cooler water flowed in the normal direction down the downcomer and up through the core where the temperature increased to 617.5 K (652.5 F).
The warmed water flowed into the hot legs, through the SG and back to the cold legs.
Because the primary was liquid full, a volume equal to the llPI flow was relieved through the pressurizer and out the PORV.
The flow induced by the density difference between the water warmed in the core region and the water TABLE VII EVENT SEQUENCE FOR LOSP-INDUCED LOFW EVENT FOR NORMAL PLANT RESPONSE Time Event (s)
(min)
O
- LOSP, TSVs close RCPs coast down Loss of all feedwater
- 0. 5 Trip reactor 2.1 Secondary SVs open
'70
- 1. 2 SGs ineffective as heat sink l
494 S.2 PORY opens
'600 10.0 SGs empty 950 15.8 Pressurizer full 1654 27.6 ilPI actuated by containment overpressure 5000 83.3 End of calculation.
system full and subcooled 18
sso r,s.~, w.~:im:.:.:.: v. ret.r;.,r,:M,W1' VN'AnWM!d41
/,:
PORY cycles to maintain pressure
.see
. l
/
p
"~
- HPt activated
[
_-s40 e
I 2
s00--
--820 g
5 5
c c
$90-5 LIQUID TEMPERATURE g
--- -- --- SATUR ATION TEMPERATURE 3
E M
M 2
O ineffbellVe
~
-500 570 O
S00 1000 1500 2000 2S00 3000 3500 4000 4S00 S000 S500 l
Tid.(s)
Fig. 9.
Core average 1iquid and satura( ion temperatures during LOSP-induced LOFV event with normal plant response.
cooled by llPI flow was sufficient to keep water in the vessel well mixed.
The maximum rod temperature of 625 K (666 F) was attained just before the llPI system i
began to deliver water.
At the end of the calculation at.5000 s (83.3 min). the primary system a rate of temperature was 606 K (632 F), 20 K (36 F) subcooled. and cooling at
- 14. 7 K/h (26. 4 F/h).
In an auxiliary calculation. the effect of having three liPI pumps operating instead of two was determined.
It was found that the cooling rate was 24.0 K/h (43.2 F/h) and the temperature was at 597 K (616 F).
The cooling mode established is termed " feed" cooling.
In this mode, the PORV cycles automatically to hold the system near 16.9 MPa (2451 psia).
Feed cooling i
can be maintained until the source of the IIPI water, the BWST, is depleted.
The BWST contains 1.47 x 10' kg (3.24 x 10' lb )
f water; at a
flow of m
l approximately 27 kg/s (2.1 x 10 lb /h), this mode can be maintained for 54000 s 5
m (15 h).
Thus, the operator has 15 h to establish some other cooling mode.
Several alternate cooling modes are discussed in the next section, which 1
considers operator actions for the management of the accident.
19
C.
Operator Actions for-Accident Management in the previous section, feed cooling was actuated automatically by the system at 1654 s (27.6 min).
The first operator intervention to be considered is early feed initiation.
In this sequence, the operator ranually actuated two llPI pumps 120 s (2 min) after the loss of feedwater.
The event sequence for this simulation is given in Table Vill.
Initiation of the llPI system resulted in pressurization of the primary system, and the PORV opened at 313 s (5.2 min).
The pressurizer was filled by 600 s (10 min).
The primary liquid temperature increased until 2800 s (46.7 min), at which time it began to decrease as the i
energy removal capacity of the flow through the core exceeded the decreasing core decay heat.
No ne t mas s wa s lost from the system, and the system was at least 36 K (65"F) subcooled throughout the transient, as is shown in Fig. 10.
With cooling established by once-through flow of IIPI water, the calculation was terminated at 3600 s (60 min).
The major effect of early feed actuation is to maintain a larger subcooling margin than the automatic feed case, 36 K (65 F) compared to'7 K (12.6 F).
Feed cooling is a limited mode of " feed-and-bleed" cooling to be discussed next.
Feed-and-bleed cooling is initiated when the operator opens the PORV to TABLE Vill EVENT SEQUENCE FOR FFED INITIATION AT 120 S FOLLOWING A LOSP-INDUCED LOFW EVENT Time Event (s)
(min)
O LOSP TSVs close, RCPs coast down, Loss of all feedwater 0.5 Trip reactor 2.1 Secondary SVs open
'70 1.2 SGs ineffective as heat sink 120
- 2. 0 2 IIPI pumps actuated 313 5.2 PORV opens
[
600 10.0 Pressurizer full
'700 11.7 SGs empty 2800 46.7 Maximum system t empe rat:.re 3600 60.0 End of calculation, system full and 30 K subcooled 20
eM i
i a
e l
- , q,;. as.s
- *- ;.=*.-lt,*,.-s
.*
- r. v-
- v.-
'm.,~.,,a.
-s40 l
e20-
/
PoRV opena and rnalntains pressure 6
l D
.,/pSGs ineffective w
w g
gio
-540 g
G ac ac LicVID TEMPERATURE y
w o00--
SATURATION TEMPERATURE
- -520 3
e-e-
5 5
MO -
c c
--sco g
g R
B a
a 7
u0
-20 0
550 1000 1500 2000 2$00 3dOO 3$00 4000 TlWE (s)
Fig. 10.
Core average liquid and sat ration temperatures for feed initiated at 120 s following a LOSP-irduced LOFW event.
reduce system pressure (bleed) shortly after the IIPI system is actuated (feed) by the operator.
The event sequence for feed-and-bleed cooling initiated at 120 s (2 min) is given in Table IX.
The primary pressure, as pictured in Fig.-11, decreased at 120 s (2 min) until the pressurizer filled at 600 s (10 min).
As the flow through the PORV changed from steam to water, the volumetric flow decreased.
Lower volumetric flow resulted in less pressure relief and the system pre ure increased.
At approximately 1200 s (20 min), the flow through the core was sufficient to remove the decay power.
Figure 12, f
which gives the liquid temperature, shows the cooling rate increased at 1200 s
.(20 min).
With sufficient cooling, the pressure also began to decrease at 1200 s (20 min).
As the system was cooled, the pressure decreased to the end of the calculation at 16000 s (4.4 h).
in a 1 quid-full system is determined by the llPI pump The pressure In other w]ords, the pressure sought is that which achieves a characteristics.
2 balance of volumetric flow.
According to the Burnell choked-flow model used in TRAC-PF1, the flow out the PORY increases with liquid subcooling.
Thypressure of the system thus drops, increasing the llPI flow to match the outficy through
(
the PORY.
The flow in the Burnell model is proportional to the squarl root of
..{
21
TABLE IX EVENT SEQUENCE FOR FEED-AND-BLEED INITIATION AT 120 S FOLLOWING A LOSP-INDUCED LOFW EVENT Time Event (s)
(min)
O
- LOSP, TSVs close, RCPs coast down, Loss of. all feedwater
- 0. 5 Trip reactor 2.1 Secondary SVs open 120
- 2. 0 Initiate feed and bleed SGs empty 600 10.0 Pressurizer full Boiling in core SVs open 16000 264 End of TRAC-PF1 calculation 27000 450 BWST empty (alternate cooling mode required) the difference between the coolant and the saturation pressures.
The pressure therefore decreases at a decreasing rate as the primary depressurizes.
This can be seen in Fig. 11.
The mass flow in the pressurizer-loop cold legs decreased until approximately 8000 s (2.2 h).
when cold llPI water that had been accumulating in the cold legs " fell" into the loop seal.
This manifests itself as a flow reversal that can be seen in Fig. 13.
Following two such flow reversals, oscillations began in the pressurizer-loop cold legs.
Oscillations of this type have been observed and analyzed previously and were found to be reasonable for a one-dimensional representation of flow in the primary piping.3 The effect of the flow reversal was to arrest the cooling process for several hundred seconds as hot water was brought into the core region from the l
pressurizer-loop hot legs and SGs.
The flow oscillations then resulted in mixing of IIPI water with water in the pressurizer loop.
Note that the water in the nonpressurizer loop, for which the split cold legs were not modeled, was 22
-2200 g4 Plow sufficient for cooling i
9 TRAC
~
j
- 800 k
- SIMPLE MODEL y
-1600 y
8 8
y A-Loop oscillations begin
-1400 l
Pressurizer full N.
-1200 t
-1000 s
10000 15000 20000 25000 30000 0
5000 TIME (s)
Fig. 11.
Primary pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event.
as0 p Open PORV. Initiate HPl t
800-I Cooling rate equal to decay power
-600 7
E
{
A-loop coki-leg cecillations begin 33,_
f y
-500 y
TRAC y
M 300
- SIMPLE WODEL p
3 2
N
-400 N
9 3
490-e p
O O
8 N
_3oo J
'N 4eo-
-200 l
3s0 0
s000 10000 150uo 20000 25000 30000 j
TIME (s) l Fi. 12.
d Primary coolant temperature for feed and bleed initiated at l
120 s following a LOSP-induced LOFW event.
23
flowing and already well mixed before this event occurred.
Cooling and depressurization then continued to the end of the TRAC-PF1 calculation at 16000 s (4.4 h).
In order to continue the simulation to the end of once-through cooling (depletion of the BWST), a fast-running simplified model was used.
With this model, adequate simulations of a liquid-full primary can be performed and used to extrapolate TRAC-PFI results.
The results of the simple model are included in Figs. 11 and 12.
It can be seen that the simple model matches the TRAC results quite well, certainly well enough for performing extrapolations.
The simple model was run until the BWST was depleted at 27000 s (7.5 h).
The operator must consider further action sometime before 7.5 h to establish a stable and permanent cooling mode.
In one such cooling mode, known as high-pressure recirculation, water is drawn from the containment sump by the low-pressure injection (LPI) system.
The water flows through the RilR heat exchangers and into the llPI system f rom which it is pumped back through the primary system.
If the operator can successfully align valves, high-pressure recirculation can cool the plant indefinitely.
Another, and more desirable, permanent mode of cooling is the use of the RilR system.
This mode of cooling is more desirable because the cooling circuit is closed, llot water flows from the hot legs, through the RilR heat exchangers, and returns to the top of the vessel downcomer.
The maximum pressure and temperature at which the RilR system can be operated are 2.4 MPa (350 psia) and 422 K (300 F),
respectively.
In what follows, we will determine whether this RilR system pressure and temperature combination can be achieved using a feed-and-bleed procedure.
With the primary being cooled by the llPI system, the reactor operator is instructed to maintain the system pressure / temperature (P/T) combination within the limits given in Fig. 14.' Specifically, the operator is required to keep the P/T within the "RCPs off" region shown in the figure.
When an adequate subcooling margin (25 K, 45 F) is obtained, the operator is to continually reduce the llPI flow rate to maintain the limits of Fig. 14.
Figure 14 also i
shows as a
dashed line the calculated P/T trace of the unthrottled 1
feed-and-bleed cooling mode.
This P/T trace leaves the RCPs-off region at approximately 7000 s and is approaching the nil-ductility-transition (NDT) condition when the BWST is depleted at 27000 s (7.5 h).
The figure also shows that the P/T combinat ion at which the RilR system can operate is not achieved.
l' 24
10c0
-2000 800-
-RCPs trip at Os
-1600 600 Q
-1200 7
i
}
6 400-O
-a00 3
Flow reversal 3
--400 200-0--
--o
_--400 A-loop flow oscillations
__.00 0
50'00 10d00 15dOO 20 BOO 25000 30000 TlWE (s)
Fig. 13.
Cold-leg mass flow for feed and bleed initiated at 120 s following a LOSP-induced LOFW event.
caoo c400-RCPs OFF
,k moo-
/
m 2 1800-
- Asoos, NDT 1000 1400-1200-1520s...
ce toco-2s000. BRITTLE A.
FRACTURE SUBCt,0 LING o soo-CE a00-4 400-200-RIIR 200 soo 460 s6o e6a voo RC TEMPERATURE ('F)
Fig. 14.
Primary temperature and pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event.
25
4 The operator must thus act by 7000 s to maintain the limits prescribed by j
RCPs-off region in Fig. 14.
To determine whether the system can be kept in the RCPs-off region and the RHR operating conditions can be achieved, a simulation of a controlled throttling of the HPI system was begun at 3600 s (60 min) into the feed-and-bleed procedure initiated at 120 s.
An optimal strategy in which subcooling was held at appro'ximately 25 K (45 F) was simulated by means of the simple model.
The P/T combination time history of this simulation is shown in Fig. 15.
This strategy resulted in a lower pressure (2.8 MPa, 406 psia) when the BWST water was depleted at 50000 s (13.9 h), but the temperature was 479 K (403 F).
The higher temperature was a result of more throttling of the HPI to I
reduce the system pressure: thus, system cooling was less.
The BWST water was extended by 6.4 h using this strategy, but the desired P/T combination was not achieved.
The cooling and depressurization capability of the existing combination of HPI and PORY was limited by the amount of flow through the system that was possible.
In the next section we will consider a larger PORY that does not have the flow limitations of the existing PORV.
D.
Equipment Studies In this subsection we will discuss three simulations that relate to 4
equipment.
In the first simulation, calculated with the simple model, we considered a PORY that had twice the flow capacity of the existing PORV.
The-P/T trace for this simulation is given in Fig. 16.
The larger PORY flow resulted in a rapid depressurization to 2.4 MPa (350 psia) by 1000 s (16.7 min).
The system temperature was reduced to 413 K (284 F) by 29240 s (8.1 h), when the BWST water was depleted.
Thus, the P/T combination for RHR operation could be achieved if relief capacity approximately double that of the current PORY were F
available.
J In the second simulation, we considered feed and bleed with diminished HPI capacity.
Nominal HPI flow is delivered to the primary by two HPI pumps.
In this simulation, it was assumed that only one HPI pump was available.
The event sequence is given in Table X.
The core average pressure is given in Fig. 17.
j The system behavior is the same as that seen in the feed-and-bleed case of the previous section until 120 s (2 min) when the PORY was opened.
The pressure then dropped until the liquid at the top of the core began to boil at about 500 s (8.3 min).
The tops of the hot legs were saturated at about the same time, and the void fraction there briefly rose to 0.06.
The pressurizer filled 4
i 26
l 2400-RCPs OPP-y 2200-
/
acco-k tano-
- bscos bteco-NDT -
f
$ 1400-
/ p' a: 1000-BRITTLE
/8"'/',/
1200-
/
8.
FRAcrURE /
/
SUBCOOLING O soo-
'12.*
g
'/
ooo-1 aco-RHR a00 a00 4bo -
ebo abo 700 RC TEMPERATURE (T)
Fig. 15.
Primary temperature and pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event (throttling begins at 3600 s).
mo 400-RCPs OFF y
m-
/
=-
n 2 1800-l:soas
$tooo-NDT -
/
$ 1400-B--
/
a; toco-BRITTLE FRACTURE /
/
SUBCOOLING S.
m _-
'/
eco-j
,...*icoco, RHR iimos 200 300 -
ebo Sbo abo 700 RC TEMPERATURE (T)
Fig. 16.
Primary temperature and pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event (PORY size twice actual).
27
TABLE X EVENT SEQUENCE FOR FEED-AND-BLEED INITIATION AT 120 S WITil 1 HPI PUMP FOLLOWING A LOSP-INDUCED LOFW EVENT Time Event (s)
(min)
O
- LOSP, TSVs close, RCPs coast down, Loss of.all feedwater 0.5 Trip reactor
'700 11.7 SGs empty 700 11.7 Pressurizer full 3300 55.0 Boiling in core Primary SVs open Tops of candy canes voided 5000 83.3 End'of calculation 7000a 116.7 Recovery of subcooling if lost aEstimated at 700 s (11.7 min), and water began to flow out the PORV.
The reduced volumetric flow associated with water flow resulted in a pressure increase that stopped the boiling in the core at 900 s (15.0 min).
With all of the voids in the system collapsed, the pressure increased more rapidly until it reached 12.6 MPa (1827 psia) at approximately 1100 s (18.3 min).
Once-through flow at l
12.6 MPa (1827 psia) was insufficient to cool the core, and the coolant j
temperature began to increase as shown in Fig. 18.
At 3300 s (55.0 min), the temperature at the top of the core reached the saturation temperature and 28 l
I 1
4-
-2250 15 i pOpen PORV, initiate llPI (1 pump)
/
-2100 14 -
9 9
-1950 n.
2 15-a wg
~1800 g
12 Dolling resumes -
y N
W A
-1650 Q-11
,p Doiling steps
-1500 10 Prcmurlier ftiI
-1350 9-e 1500 2000 2 BOO 4000 3500 4000 4500 5000
, - - r- - T-O SCO 1000 Tiut (s)
Fig. 17.
Core average pressure for feed and bleed initiated at 120 s following a LOSP-induced LOFW event (one llPI pump).
620
-850 615 -
Q 610 -~
-640 p
v v
I w
w
-{
.-630 605
_,,,,,,.. * * * * - - - - - - - - - - * ~ ~ ' ~ ~ ~ ~ '
600-- i
/'
--620 W
I 2
W i
W g
_ 't
-- SiO 2
585-
\\
l 9
4 se0-i uoUID TEMPERATURE N
5 i,
i
----- SATURATION TEMPERATURE e
.', /
4 4
585-
'., /
-soo 580-
-500 575 1000 1500 2000 2500 3000 3500 4000 4500 5000 0
500 i
TlWE (s)
Fig. 18.
Core average liquid and saturation temperatures for feed and bleed initiated at 120 s following a LOSP-induced LOFW event (one HPI pump).
29
boiling began.
The calculation was ended at 5000 s (83.3 min).
At 5000 s (83.3 min), the void fraction at the top of the core was 0.013.
The tops of the hot legs remained full, however.
Thus, flow around the primary loop was not interrupted.
The peak fuel-rod temperature of 605 K (629 F) was reached at 3300 s (55.0 min) and is estimated to remain at that level until 7000 s (1.9 h).
At 7000 s (1.9 h), the decay power is estimated to be at a level that can be removed by once-through cooling with no boiling.
At this time, the mass loss rate will be reduced to zero and cooldown will begin.
Thus, early actuation of one IIPI pump in conjunction with opening the PORY prevented a large mass loss.
Although some boiling occurred, the tops of the primary loops remained essentially full.
In the final simulation of this transient it was assumed that auxiliary feedwater (AFW) was restored 900 s (15.0 min) after the LOSP initiator.
The event sequence is given in Table XI.
Main feedwater coasted down in 16 s, instead of instantly, and in this calculation the turbine-bypass system was available.
At 900 s (15.0 min), AFW flow of 34 kg/s (2.7 x 10' lb,/h) to each SG was restored.
When the SG liquid level corresponding to 50% of the normal operating range was attained, the flow was throttled to maintain that level.
The primary and secondary pressures and primary temperature are given in Figs. 19 and 20.
Restoration of AFW results in depres,surization and cooling of the system until 1850 s (30.8 min) when the AFW flow was throttled.
The cooler AFW also cooled and depressurized the secondary side, as can be seen in Fig. 19.
As the flow was throttled, the secondary side repressurized, the turbine-bypass valves. opened, and primary cooling by natural circulation was established.
This is illustrated in Fig. 20, which shows a slow cooling of the primary liquid.
The AFW flow required at the end of this simulation is only 6.7 kg/s (5.3 x 10' lb,/h) for each SG.
Additional cooling is possible by increasing the AFW flow and manually opening the turbine-bypass valves.
It has thus been shown i
that restoration of AFW before the primary loops are voided also restores cooling of the primary system.
E.
Conclusions for LOFW The LOFW studies have provided insights about the course of the sequence, event timing, the importance of equipment availabilities, and the impact of selected operator act ions or inactions.
The minimum timing of critical events has been calculated for the basic station blackout transient, T (B )MLU.
The i
3 3
following conclusions have been reached abcut the importance of equipment.
l 30 l
I
TABLE XI EVENT SEQUENCE FOR RECOVERY OF AUXILIARY FEEDWATER AT 900 S FOLLOWING LOSP-INDUCED LOFW EVENT Time Event (s)
(min)
O LOSP.
TSVs close, RCPs coast down, Loss of all feedwater
- 0. 5 Trip reactor
- 2. 0 Secondary SVs open 768 12.8 PORV opens 900 15.0 SGs empty 900 15.0 Restore AFW Pressurizer full HPI actuated 1850 30.8 SG filled to 50%
of operating range 2500 41.7 Long-term cooling established 4700 78.3 End of calculation 1.
Two IIPI pumps actuated automatically on containment overpressure at 1654 s (27.6 min) are sufficient to maintain subcooling and cool the plant; 2.
One HP1 pump is sufficient to cool the plant if initiated early (<120 s (2 min)], but saturation and some boiling in the core region will occur; and 3.
A long-term cooling mode using the RilR system cannot be achieved by cooling and depressurizing in a
once-through cooling mode (feed-and-bleed cooling) without additional primary relief capacity or additional water supplies after the BWST is empty.
31
'8 AFW restored is
-2250 SG D PRIMARY' 34
- SG D SECONDARY 2000 T
T a
d 12 -
-1750 E
W S
-1500 g
w w
w E
w E'
-1250 8-s.
FGs tilled to 50~.
-750 4
2500 4000 J500 4000 4500 5000 0
SCO 1000 B00 2000 TlWE (s)
Fig. 19.
Primary pressure for auxiliary-feedwater recovery at 900 s following a LOSP-induced LOFW event, em i
^\\
~
s20-
-a60
/.~
m A
g
,/
LIOUID TEMPERATURE
-s40 D
i
SATURATION TEMPERATURE y
g
=
Q s00--
-620 Q
m g
w w
590-A restored 6no E
N_
5s0 -
5_
h "O
l N
(
q 570~
SGs ne ective Q
m
-560 m
$go _
N SGs filled to 5(r4
-340 1
550 0
500 1000 1500 2000 2>00 3000 3500 4000 4500 5000 TlWE (s) l Fig. 20.
Primary-coolant liquid and saturation temperatures for l
auxiliary-feedwater recovery at 900 s following a
LOSP-induced LOFW event.
32
In addition, conclusions were reached about several operator-initiated actions.
1.
The availability of the llPIS and its use early in the transient in a feed-only cooling mode prevents any net primary mass loss and maintains large subcooling margin; a
2.
Operator initiation of feed-and-bleed cooling early in the transient increases the cooling rate by increasing the flow through the system; and 3.
Restoration of AFW permits depressurization and cooling of the system.
If initiated before 900 s (15 min), natural circulation can be quickly restored because the primary is liquid full.
V.
SMALL-SMALL BREAK LOCA A.
Introduction We now examine a series of small-break loss-of-coolant transients that are initiated by a random break of less than 102-mm (4-in.) diameter in the primary-system cold-leg piping.
The initiator is designated by S.
During the 3
course of the accident, the high-pressure emergency coolant recirculation system is assumed to not function (system failure designated 11 ) and the steam generators to remain isolated.
The supply of once-through borated water will allow the llPI system to function while the operator attempts to either restore high-pressure emergency coolant recirculation or bring the primary to conditions under which the low-pressure recirculation system is functional.
The overall sequence is designated S 11.
M re details can be found in Ref.
1.
A diagram of 3
the noding scheme used to model the plant, plus other modeling details, can be found in Appendix D.II.
The response of the primary system is strongly determined by both break size and llPI flow.
For small breaks it may be necessary to limit the amount of IIPI flow in order to depressurize the primary.
This strategy will extend the life of the once-through cooling water, but will also decrease the cooling rate of the primary system.
Calculations were performed to indicate the effect of varying break size and operator throttling of the HPI on several important factors that will help determine the outcome of the accident.
These include 33
- 1) effectiveness of the llPI in keeping the primary full, 2) ability to reach conditions of low-pressure recirculation, and 3) time that once-through cooling water remains available.
B.
Assumptions Used Durine Accident Progression Four small-small break accidents were analyzed.
The in,tial conditions for these differed only in the size of the break.
The common logical controls for the system (transient boundary conditions) are as follows:
If the time that the pressurizer pressure falls below the reactor trip setpoint of 13.1 MPa (1900 psia) is designated t then the reactor g,
and the main feedwater are tripped at t
+ 0. 5 s, the turbine stop 3
valve is closed at t 3 + 1 s, and the AFW is turned on at t i + 30 s.
If the time that the pressurizer pressure falls below the llPI trip se t point of 11.135 MPa (1615 psia) is designated t2, then the IIPI is turned on at t2 + 35 s, and the RCPs are turned off at t2 + 50 s.
The actual times of these events for the four small-small break accidents-analyzed are given in Table XII.
As would be expected, the larger the break size, the faster the initial depressurization and reactor scram.
For the smallest break size, the reactor did not scram until over 900 s (15 min) after the 11-mm-diam (0.43 in.) break occurred.
This table should be used to interpret the plots of various system parameters that are given for each specific accident in the sections that follow.
C.
TRAC Calculations of Accident Sequences 1.
Break Size of 102-mm (4-in. ) Diameter.
This is the largest break size in the "sma ll-small" class of break-initiated accident sequences.
Figures 21 through 26 show, respectively, TRAC-calculated values for break outflow, inflow through one IIPI, core average liquid temperature, pressurizer pressure, core liquid volume fraction, and low-pressure-injection mass flow, from the beginning of the accident to 2364 s (39.4 min).
The break mass flow (Fig. 21) began by flashing the subcooled liquid in the cold-leg piping at ~700 kg/s (5.56 x 10' lb,/h).
By 200 s the flow became two-phase because of steam flow through the vent valves into the downcomer and the vessel side of the cold leg near the break.
By 1500 s (25.0 min) the break flow was still two-phase and highly unstable, with an average value of 80 kg/s (6.3 x 10' lb,/h).
34
TABLE XII TIMING OF EVENTS IN FOUR SMALL-SMALL BREAK LOCAS (SECONDS AFTER BREAK)
Break Trip Main Diameter Reactor Feedwater, Begin Enable Enable Turn off l
(mm)
Trip 14-s Coastdown Close TSV AFW HPI Pumps
[0.5 s after
[0.5 s after low-
[1.0 s after
[30 s after
[35 s after
[50 s after low pressure pressure reactor low pressure low pres-HPI signal]
HPI signal) signal]
trip signal]
signal) sure signal]
)
102 8.1 8.1 8.5 37.6 50.6 65.6 31, 96.4 96.4 96.9 125.9 143.3 158.3 22 200.4 200.4 200.9 229.9 247.7 262.7 11, 926.3 926.3 926.8 955.9 973.4 988.3 1
l N
800
-1750 N-
-1500 600-
-1250 9
500-h
-1000 400-R R
g Two-phase flow out brealc begins g
_ 73o g
300-g W
LPI initiation
/
W o
2
-500 2
200-10 0 - ~
0--
--O.
250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s)
Fig. 21.
Mass flow out 102-mm-diam (4-in.) break.
The inflow through one llPI (Fig. 22) was nearly one-quarter of tm total of the four IIPIs, which means that the total llPI flow at 1500 s (25.0 3;n) was 5
-65 kg/s (5.2 x 10 lb,/h).
The primary was thus losing tr ' :. s at 5
~15 kg/s (1.2 x 10 1b,/h).
The pr ima ry was sti11 cooling (Fig. 23), 90 wever, because there was a net energy loss (outflow - inflow + generation).
line system pressure (Fig. 24) decreased rapidly initially, until it reached the saturation pressure of the hottest fluid in the primary
(~7.5 MPa, 1088 psia).
The depressurization then followed the saturation pressure as the system cooled.
This behavior contrasts with that of the smaller small-break accidents described next, in that the llPIS did not have the capacity to refill and " pump up" the system.
The CFT began slowly dumping water into the core at ~700 s (11.7 min) as the pressure fell below 4.24 MPa (600 psig) and the LPIS came on at 1810 s (30 min) as the pressure f e l l be l ow 1. 45 MPa (210 ps i a ).
The CFT and LPI water (subcooled relative to the primary fluid at the injection sites) caused violent condensation oscillations (Figs. 25 and 26), and significantly reduced the numerical time step.
The ca lculat ion was ended as soon as it became apparent i
that the vessel was refilling.
If the low-pressure recirculation system were l
36
20 4o 17.5 -
35
-30 12.5 -
9 o
.x o
io.
-20 u) 7.5 --
m N
Q 2
2 5-
--10 2.5 -
0-
- -0
-25 0
250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s)
Fig. 22.
Mass flow in single IIPI line, 102-mm-diam (4-in.) break case.
600 600 580-8
- 570 p
v w=
560-w g
g<
540
=
540 -.
510 W
3w a5 520 -~
480 o
j U
g
_ 43o 500-h
- 420 f5 si W
480-f 8
-3eo 4a0-
-360 440 0
250 500 750 1000 1250 1500 1750 2000 2250 2500 TIME (s)
Fig. 23.
Core average liquid temperature, 102-mm-diam (4-in.) break case.
37
14 -
-2000 12-
-- 950 10-Saturation pressurs
.m 9
~
8-f 6-7so E
4-
-500 2--
-250 0
A.*
A A
A
& 240 & 20 TlhE (s)
Fig. 24.
Pressurizer pressure, 102-mm-diam (4-in.) break case.
LOS 1-0.95-7 0
N
~
E
\\
0.s0 -
)
Y I
0.as-go LPI Initiation Accumulator j
g 0.e0-huection o
Begms l
3 o
tJ 0.75 --
J Y
~
0.65-l 0.60 0
250 500 750 1000 1250 1500 T7b0 2000 2250 2500 TIME (s)
Fig. 25.
Core liquid volume fraction. 102-mm-diam (4-in.) break case.
38
c.
J p
16 0 wo-
- 3'S q
,2n _--
_ -vo 10 0 --
--225 so--
1
-- 18 0 o
D i
3 h
m so--
- -us g
V2 40--
--90 gn.-
--<s g_.
--g
-20 0
250 500 750 1000 1250 1500
!?50 2000 2250 2500 Ti m (s)
Fig. 26.
Low-pressure-injection mass flow, 102-mm-diam (4-in.) break case.
not operational, the supply of once-through cooling would be exhausted 9000 s
(-2.5 h) after accident initiation.
2.
Break Size of 31-mm (1.2-in.) Diameter.
Figures 27 through 31 show, respectively, the break outflow, IIPI inflow, average' liquid temperature in the core, the pressurizer pressure, and the core liquid volume fraction from the beginning of the transient to 10000 s (28 h).
For this size break and smaller, the llPI system has the capacity to keep the primary system full, and did so after an initial period of vapor generation in the vessel.
5 The flow out the break (Fig. 27) began at 77 kg/s (6.1 x 10 lb /h), but m
as the liquid subcooling decreased (as the primary depressurized), the magnitude 5
of the choked flow rapidly decreased to a minimum of 32 kg/s (2.5 x 10 lb /h).
m When the llPI came on at 143 s, the local subcooling at the break increased rapidly, increasing the choked outflow.
Boiling in the vessel produced large oscillations in the outflow that did not stabilize until ~1500 s (25 min).
Although vapor generation ceased at ~600 s (10 min), a large vapor bubble remained in the upper head until ~5500 s (1.53 h), when it finally condensed, causing a large perturbation in the break flow.
Condensation of the vapor 39
l bubble was slow because flow into the upper head was limited by the small bypass flow areas into the upper head.
The mass flow in a~ single llPI line (Fig. 28) was relatively constant after the IIPls came on at 143 s, showing only perturbations due to condensation in the vessel.
The core average liquid temperature (Fig. 29) shows the initial rapid drop to saturation conditions and boiling, and then recovery of subcooling after llPI initiation.
The core average temperature fell to 470 K (390 F) at 10000 s (27.8 h).
The pressurizer pressure also shows the sudden drop to saturation i
conditions, the recovery of subcooling, and the previously discussed perturbation due to condensation of the bubble in the vessel upper head at
~5500 s.
The core liquid volume fraction (Fig. 31) shows the effect of boiling unti1 600 s.
No more vapor was generated until termination of the TRAC calculation at 10000 s.
Possible scenarios for the final outcome of this and the next two break sizes will be discussed later.
80 i
7S-
-- 16 5 i
70-
- 15 0 65-l Condensation of
-l Q
vapor bubble in upper head 9
so.
{
E 55-
- -i20 o
I m
m
'I Q
W 50-t 2
-105 L
f 43 l
40-
- '80 l'
55-_
-_3 30
[
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 TIME (s)
Fig. 27.
Mass flow out 31-mm-diam (1.2-in.) break.
40
,w, n
m
us 15 -
30 12.5 --
-25 T
?
30 b,
6 20 c
W 3
y 7.5 -
g
- 15 m
en W
W 2
5-2 1
2.5 -
-.s 1
I i
0-
--O
-2,5
~
~5 0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 i
M (s)-
Fig. 28.
Mass flow in single IIPI line, 31-mm-diam (1.2-in.) break case.
4 600
-600 1
580-i b
Fg.ttaration
- 570 v
w t
m-W g(*
Re. coven subcooling g
a<
-540 m
I:
540-50 o
l 5
oU 3
d
--4s0 i
520-w t
2 j
f
- 450 j
500-g<
- 420 8
4a0-
,390 460 0
9000 2000 3000 4000 5000 6000 7000 8000 9000 10000 TlWE (s) t Fig. 29.
Core' average liquid temperature, 31-mm-diam (1.2-in.) break case.
i 41 i
Ty,
.a-*-
16
-2250 14-l 12-
-- 950 7
~"
10-Saturation 8-Condensation of steam in upper head
-1000 6-
- 750 4-
-500 2
i i
i o
e sooo acoo 4000 sooo sooo woo oooo oooo soooo TnE(s)
Fig. 30.
Pressurizer pressure, 31-mm-diam (1.2-in.) break case.
10025 1-l f
~
Wtid void generation 7O 0.9950-I E
0.9925-l W
0.9900-l o 0.9875 -
5o3 0.9850-0.9s25-0.9800--
0.9775-0.9750 0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 TIME (s)
Fig. 31.
Core liquid volume fraction. 31-mm-diam (1.2-in.) break case.
42 1
1
3.
Break Size of 22-mm (0.87-in. ) Diameter.
Figures 32 through 36 show, respectively, break outflow, llPI inflow, core average liquid temperature, pressurizer pressure, and core liquid volume fraction from the beginning of the transient to 10000 s (2.78 h).
These are the same parameters shown in Figs. 27 through 31 of the 31-mm break case, and the behavior is qualitatively similar except for event timing.
The break mass flow (Fig. 32) again dropped rapidly initially as local subcooling decreased, then increased slowly to a relatively constant value after the llPIS came on at 248 s.
In this transient, much less vapor was generated in the vessel, and there were no large perturbations due to delayed condensation of vapor pockets in the vessel head.
The mass flow in a single IIPI line (Fig. 33) reached an initial maximum based on saturation pressure during the short period of boiling in the core, increased.
then slowly decreased af ter 1000 s (16.7 min) as the primary pressure The core average liquid temperature (Fig. 34) fell rapidly after reactor 4
trip until it reached the temperature determined by the system-wide heat balance
- at 2000 s (33.3 min).
This temperature was nearly constant for several hundred seconds after 2000 s (33.3 min), but the slow decrease in decay heat level and increase in break flow (due to increased subcooling) began to cause the core 3
average (and primary system average) temperature to fall slowly.
The pressurizer pressure (Fig. 35) rapidly fell to saturation, but after the last voids condensed at -1000 s, the IIPIs were able to " pump" the pressure up again because of the small size of the break.
A stable pressure plateau was reached at which the break flow and IIPI flow were equal af ter 3000 s.
The core liquid volume f raction (Fig. L;) dropped from unity for a shorter i
period than in the previous transient, because the llPI was cble to repressurize the system faster with a smaller break.
4.
Break Si ze of 11-mm (0. 43-i n. ) Di ame t e r.
Figures 37 through 41 show, i
respectively, break outflow, IIPI inflow, core average liquid temperature, pressurizer pressure, and core liquid volume fraction from the beginning of the transient to 10000 s (2.78 h).
These are the same parameters shown in the last two cases.
The flow through the break (Fig. 37) decreased initially as the primary i
slowly depressurized and subcooling decreased.
When the reactor tripped at 926 s the pressure fell rapidly, as did the break flow.
IIPI initiation at 973 s 43
?
45
~
40-t
\\
80 35-I
^4 o
70 5
5 O.,
30-h h
60 25-m E
2 3g 3
Minimum subcooling, 20-minimum flow
-40 15 -
30 0
1000 2d00 3dOO 4N00 5dOO 6dOO 7800 8dOO 9dOO 10000 TIME (s)
Fig. 32.
Mass flow out 22-mm-diam (0.87-in.) break.
16 35 Minimum s stem ressurr, maximum 'PIin ow
-30 12-
-25 10-3
.7
-20 v
g o
8-h
-15 h
m 6-m N
2 10 2
4
'S g.
0-
- -O
-2 0
1000 2000 3000 4000 5000 6000 7000 8000
'J000 10000 TIME (s)
Fig. 33.
Mass flow in single llPI line, 22-mm-diam (0.87-in.) break case.
1 44
620 600 590-
-.600
'580-F b
g
-580 y
ce p
N U
g 570 -
-560 g
se0-
-540 m-3 b
-m 540-W I
W
-50' 5
4 530-N U
g
_-480 g
O
$g
--460 500 0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 TIME (s)
Fig. 34.
Core average liquid temperature, 22-mm-diam (0.87-in.) break case.
15
-2100 14-l -
1800 12-9 E
~
g
-1650 i
S 11-Q
-1500 10 -
n.
-1350 9-Saturation pmssure
~
-1200 8-
-m50 0
1800 2d00 3dOO 4dOO SdOO 6000 7dOO BdOO 9800 10000 TIME (s)
Fig. 35.
Pressurizer pressure. 22-mm-diam (0.87-in.) break case.
45 l
i
1.0005 i
4 s-5g 0.9995-
~
E 0.9%u -
B 9
h 0.9985-8 W
8 0.9980-Initial void generation 0.9975-0.9970 i
i 0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 TNE (s)
Fig. 36.
Core liquid volume f raction, 22-mm-diam (0.87-in. ) break case.
(Fig. 38) caused the pressure to increase, subcooling to increase, and thus the choked break flow to increase until the pressure reached the PORV setpoint.
The core average liquid temperature (Fig. 39) decreased rapidly after the reactor trip until ~1800 s (30 min) when, in a manner similar to the 22-mm break case, it reached a temperature determined by the system-wide energy balance.
This temperature was increasing after ~2000 s (33.3 min), because the small break and IIPI flow were not providing sufficient cooling to remove the decay heat.
When the PORV opened at 2100 s (35 min), the rate of temperature increase began to decrease, and by 6000 s (100 min), the PORV and break were together removing the core decay heat, and the system began to cool.
Any break smaller than 11 mm will also cause the PORV to open, with similar subsequent behavior.
ilPI flow allowed less cooling of the primary, resulting in an increasing average temperature.
When the pressure (Fig. 40) reached the PORV setpoint of 16.8 Mpa (2440 psia) the PORV opened, providing enough added energy release to slowly begin cooling the system.
46 l
Figure 41 shows that boiling occurred in the core for only a very short
~
period in this transient.
In the 102-mm-diam (4.02 in.) break case, the outflow was too large for the llPI to match, and the primary system lost inventory.
In the 31 -mm-di am break case, the break was just small enough that the llPI flow could refill the system, taking over 5000 s (83 min).
In the 22-mm-diam break case, less than 1000 s (16.7 min) was required, and in this case, the 11-mm-diam The llPIS break, the break was so small that only ~100 s (1.7 min) was required.
pumped the primary up to the PORV setpoint of 16.8 MPa (2437 psia), and the PORV cycled to maintain this pressure.
The system would slowly cool until once-through cooling was exhausted.
D.
Extensions of TRAC Calculations to Once-Through Water Limit Three of the four calculations discussed in Sec. IV.C (the 31, 22-and 11-mm-diam break cases) were ended 10000 s (2.8 h) after the initial break.
In each of these cases the primary system wac slowly cooling and depressurizing.
An important question is whether the system will reach the conditions of pressure and temperature (1.45 MPa and 422 K, 210 psia and 300 F) at which the low-pressure recirculation system can function before the available supply of 13 12 -
~
-25 11 --
10-
- -20 s-Reactor trip j
W 5
h 8-m m
Q N
7-
-15 3
1 6-l 5-Minimum system subcooline.
-iO
\\/ minimum outflow 3
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 TIME (s)
Fig. 37.
l Mass flow out 11-mm-diam (O.43-in.) break.
47
1 16
-35 14 -
Inflow decreases as
'30
~
system pressurizes to POllV setpoint 12-
-25 W-J
~
-20 8-
\\
3 h
(n 6-c.-
-15 N
g Q
2
-10 2
4-2
-s I
O
- -O
-2 0
1000 2000 3000 4000 5000 0000 7000 8000 9000 10000 TIME (s)
Fig. 38.
Mass flow in single llPI line, 11-mm-diam (0.43-in.) breu case.
c00
- 615 b
~
-600 C
G U
E sea.
_-sas S
a.
I W
y
~
- 570 57o.
g U
UB U
Core temperatum sleterminn!
-555 g
, positive system-wide energy bga anco g
s60 M
- 540 8
g sso-o o
- 525 540 0
1000 2000 J000 4000 5000 6000 7000 8000 9000 10000 M (s) l Fig. 39.
Core average liquid temperature. 11-mm-diam (0.43-in.) break case.
48
J 18 i
-2600
~
- ' 1 f-
--POHV Setpoint pressure '3 2<0o 16
{
-2200 14 -
7 f
9 k
l
~
d
-1800 3
12 -
W 8
D E
-1600 8
E 10.
4
. goo 8-~
^N
-jooo 6
0 1000 2000 3000 4000 5000 0000 7000 8000 9000 10000 TIME (s)
Fig. 40.
Pressurizer pressure, 11-mm-diam (0.43-in.) break case.
L0005 i_ _
h 0.9995-5 4b 0.9990-
$g C.9385-8 O
C.9980-
$.9975-0 yInitial void generation 0.9970-0.9965 0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 TIME (s)
Fig. 41.
Core liquid volume fraction. 11-mm-diam (0.43-ir.) break case.
49 F
,a n
)
once-through cooling water is depleted.
To answer this question using the TRAC model would require excessive amounts of computer time.
Linear extrapolation of TRAC resul t s is also not advisable for large times.
An intermediate pa t h wa s taken; namely, a very simple mode l wa s developed that accepted the final TRAC solution as an initial condition, and extended it in time at minimal cost.
1.
Solution of Global Continuity and Enersry Equations.
Two necessary conditions for the use of this simple model technique are that the primary be full, or nearly full, of water, and that this water be the rmally we ll-mi xed.
The final TRAC solution was taken as the initial condition, and the transient conservation equations for mass and energy were solved on a system-wide basis.
The mass-conservation equation was solved for primary liquid mass with constant system
- volume, and specified llPI inflow and break outflow.
The energy-conservation equation was solved for the primary system plus the SGs (including the effective heat ca pac i t i e s of the core and other solid components), with specified IIPI inflow and break outflow.
The calculation of break outflow used the same method used in TRAC for choked and friction-controlled flow of a flashing subcooled liquid.
The calculation of IIPI inflow used the same IIPI characteristics used in TRAC.
Liquid water property evaluations are identical to those in TRAC.
The overall solution is fully implicit. iterative and allows transition from choked to friction-controlled break flow for a cycling PORV that maintains constant system pressure and for arbit rary cont rol of IIPI throttling.
2.
Extensions of TRAC Solutions.
The 31,
22-and 11-mm-diam break calculations were extended with the above method to the time that once-through cooling water is depleted.
These extensions were begun well before the end of the TRAC calculations to compare results.
These comparisons are shown in Fig. 42.
l As is expected, Fig. 42 shows that smaller breaks required less llPI flow, so the supply of once-through water lasted longer.
Breaks less than 16-mm (0.63-in.) diameter caused the PORV to maintain a constant pressure and thus constant llPI flow.
Agreement between the global extensions and their respective TRAC cases was good considering the simplici., of the model.
50 l
l Each of these calculations ended when the BWST emptied; in none of these did the pressure approach the level of 1.45 MPa (210 psia) necessary for operation of the low-pressure recirculation system.
These results assumed no operator throttling of the llPI.
This can be an effective method for reducing system pressure, as is shown in Fig. 43.
This figure shows the effect of throttling the llPI such that the primary remains 25 K subcooled.
The time that the once-through cooling water reserve was depleted was extended f rom 26000 to 43000 s (7.2 to 11.9 h). and the pressure decreased to near the LPI setpoint of 1.45 MPa (210 psia).
The effect of this throttling is shown in Fig. 44, as the temperature did not decrease nearly as rapidly and did not approach the low-pressure recirculation design criterion of 422 K (300 F).
The effects of IIPI throttling for the 22-mm-diam break are shown in Figs. 45 and 46.
The operator throttled the IIPI beginning at 10000 s to achieve a primary subcooling of 25 K (45 F).
The pressure dropped rapidly, but did not COO 575 11 mm
/
',k/
~~
GLODAL MODEL 550-.-
g
.......... TRAC 525
',Nmm
}
'h 500 O
7 s1G m m,
g7,.
u Q
in m m l
M, 450- 31 mm N.
l Q
mmm 47s 27 m m
~
U Once-thrntich water limit.
400-
~
31 mm i
375-0 10000 20000 30000 40000 50000 60000 Fig. 42.
Extension of TRAC results with global solution o f ma s s ard energy equations for various break d i an.c t e r s,
no llPI throttling.
51
approach the low-pressure recirculation setpoint, and the temperature remained near 500 K (440 F) (Fig. 46).
Throttling for smaller break sizes can lead to increasing system temperature.
E.
Conclusions for Small-Small Ilreak LOCA The 511 IJEA sequence can be divided by break size into two groups whose 3
behavior is quite different.
Sequences with initial break sizes smaller than
~38-mm d iam (~1.5 in.) will remain pressurized by the llPI (after an initial period of steam generation in the vessel) and will cool at a rate determined by the magnitude of flow out the break, or out the break and the PORV if the break is small enough (<16-mm diam).
The period of available once-through cooling can be extended by operator throttling of the llPl. but conditions necessary for low-pressure recirculation cooling are not niet before once-through water runs out.
Sequences with initial breaks larger than 38-mm-diam cannot be pressurized by the llPIS, but will depressurize at rate determined by the dependence of the a
saturation pressure on the temperature of the hottest fluid in the core as boiling occurs and the saturation temperature slowly falls.
The action of the e
No thmtuing 7- ~l
Throttling to
~ -N maintain 25-K 7
\\
. Flow becomes
""
- W z
,... - ' un choked 9
.,no 6-ir 4
o
~
'\\
W 3
5-
\\
10 i
8
\\.
E 4--
\\.,
_-s00 y<
3--
.,N..,
_- 4s0 W
8 2--
. -300 l
l
'l -
-20 O
S000 10000 15000 20000 25003 30000 35000 40000 45000 TIME (s)
Fig. 43.
Effect of throttling IIPI to maintain 25-K primary subcooling on primary pressure for 31-mm-diam (1.2-in.) break.
52
54o N,
No thruttling 52o-.-
Throttling to
- -'80 inaintain 25-K subcooling h
b
\\
W N
soo.
--44o g
x f5 W
~..,,
4oo n.
4ao-3 h
beo g
e a
4eo.
5
-360 g
a S
w g
-320 4
4 W
oO
-280 0
400-
-240 3ao 0
5000 10000 15000 20000 25000 30000 35000 40000 45000 TIME (s)
Fig. 44.
Effect of throttling IIPI to maintain 25-K primary subcooling on primary average temperature for 31-mm-diam (1.2-in.)
break.
14
., goo 12 7g
-tsco
?
g 10.
d
~
No throttling O
- Throttii
-teco maintain K
El subcooling i
E
-itoo 8-d i
-tooo i
6-g
-soo 4-'
-~***
-4eo 2
O 10000 20000 30000 40000 Socco 60000 TlWE (s) rig. 45.
l Effect of throttling IIPI to maintain 25-K primary subcooling l
on primary pressure for 22-mm-dian (0.87-in.) break.
I 53 r
5e0
~
- 520 540--
b C
t N
~
-4s0 w
520-a-
j2
<t g
500--
..un 3w 9
W D
480-
~~400 0
0
~
No throttling 3
d ThrvUling to d
460-Inaintain 25 K -
o g
~
subcooling
-360 E
W 440-U O
~
-320 "o
U U
420-
~
-280 400 0
10000 20000 30000 40000 50000 60000 TNE (s)
Fig. 46.
Effect of throttling IIPI to maintain 25-K primary subcooling on primary pressure for 31-mm-diam (1.2-in.) break.
low-pressure injection can refill the primary, but this will be a temporary reprieve if low-pressure recirculation cannot be made available.
VI.
LOFW-INITIATED TRANSIENT WITil0UT SCRAM in this section, we examine a sequence for which a loss of the power con-version system is caused by other than a LOSP.
This initiator with failure of the power conversion system is designated T M and is followed by a failure of 2
the reactor protection system (K) and the failure of the llPIS (U).
This se-quence is the type known as anticipated transients without scram ( AWS).
The designated T KMU and more details can be found in total accident sequence is 2
Ref.
1.
With offsite power available, the RCPs operate throughout the transient.
Data for the transient were obtained primarily from Ref. 5 and IWW.
The latter
[
also provided transient results for comparison.
Depending upon the cause of the l
loss of main feedwater, coastdown times can vary from approximately 1s to l
- This information supplied by Randy Ellison (October 27, 1983).
54
l 15 s.5* We assumed a 1 s main feedwater coastdown and that the TSVs also closed by 1 s.
AFW, when fully available, begins to deliver water at 15 s with full flow (DS kg/s, 1.02 x 10' Imb/h) at 31 s.5* We assumed full AFW flow began at 27.5 s consistent with Ref. 6.
AFW was throttled to maintain 509 of the normal SG operating range should that level be achieved.
Reactivity feedback coefficients that made full use of the interdependencies now available in TRAC-PF1 (Ver. 11.0) we re generated by means of the LEOPARD code.' Reactivity coefficients and other reactivity-related data are given in Table XIII, where they are compared to values obtained from Ref. 5.
The most important coefficient is the moderator temperature coefficient (MTC) because heating of the coolant is the only prompt negative feedback.
The MTC used in this analysis is conservative relative to those of the references because the feedback is reduced. The difference in value is a result of the fact that it was generated for a new or " fresh" core, whereas those of the references were generated for a core that had some burnup.
The MTCs appearing in the references represent values that will not be exceeded over 99% of the life of an equilibrium cycle.
The fuel coefficient is also slightly more conservative since the fuel will cool, resulting in a reactivity insertion.
Voiding and boron concentration reactivities were not considered directly in Refs. 6 and 7, and therefore could not be compared.
This accident sequence is designated T KMU.
More details can be found in 2
Ref.
1.
A diagram of the noding scheme used to model the plant, plus other modeling details, can be found in Appendix D.III.
A.
Base Case The event sequence for the base case transient is given in Table XIV.
No llPI was available for this case.
The coastdown of main feedwater in 1s left the SGs unable to remove the core power, and resulted in pressurization of both the primary and secondary systems.
At 4.6 s, a primary system overpressure signal generated a reactor trip signal but the reactor failed to scram.
The rapid primary-system pressurization and heating at the start of the transient is shown in Figs. 47 and 48.
The PORV opened at 9.9 s followed shortly thereafter by the SVs.
Feedback from heating of the coolant resulted in a lowering of the power.
Lowering the power, in turn, produced in a flattening of the temperature profile from the fuel centerline to the coolant, such that the coolant temperature increased and the average fuel temperature decreased.
The coolant l
55
- This information supplied by Randy Ellison (October 27, 1983).
w
---e,,-.,n-
,w,,
21 i
-3000 Power reduced by coolant ternperature feedback 20-q
-2850 Q
o_
Pressurizer filled with liquid g
b O
19 -
y
-2700 y
Saturation
]
18 -
y E
h
-2550 w
w o
17 -
0 4
-2400 w"
4 4
16 --
g g
m x
O
-2250 o
U U
15 -
Power increases as a result of cooling
- 210 0 0
2b0 4b0 6b0 8b0 1000 12b0 14'00 16b0 18b0 2000 TIME (s)
Fig. 47.
Core average pressure during base T KMU transient.
3 650 n
n b
-700 b
W 640-(;
g 3
,f'.
3
\\
W 1 \\
-oeo j
630-p 2w 2w H
LIQUID TEMPERATURE w
~**
620
- -------- SATURATION TEMPERATURE z
O o
C D
--840 610 --
w m
m w
w 600-
--620 m
W m
W k
I Q
590-w
-600 w
a x
O o
U u
580 o
200 400 600 800 1000 1200 1400 1600 00 2000 TIME (s)
Fig. 48.
l Core average liquid and saturation temperatures during base l
T KMU transient.
2 I
56 l
w TABLE XIII REACTIVITY DATA FOR A FULL-POWER BEGINNING-OF-CYCLE CORE This Report Ref. 5 Ref. 6 Fuel Temperature
-2.57 x 10-5
-2.38 x 10-'
a Coefficient (
/K)
MTC(
/K)
-1.06 x 10-*
-1.89 x 10-*
-1.80 x 10-*
b Void Coefficient (
/a)
+.054 a
a Boron Concentration
-1.03 x 10-'
-9.09 x 10-5 a
(bk7 ppm)
Coefficient
/b Boron Concentration 1345 ppm 1383 ppm a
for Criticality
" Not available.
b Note that the value of the void coefficient is positive for zero void because of boron removal when voiding begins.
and fuel temperature feedbacks shown in Fig. 49 resulted in the power that is shown in Fig. 50.
Expansion of the primary coolant superheated the steam in the pressurizer as it was compressed.
This resulted in a highly nonequilibrium situation with superheated steam above subcooled liquid in the pressurizer.
When the was nearly full at 36 s, TRAC-PF1 calculated a rapid drop in pressurizer pressure as condensation of the superheated steam took place at the top of the pressurizer and within the pressurizer relief piping.
In TRAC-PF1, the condensation rate is proportional to the relative velocity of the phases.
Water moving at a substantial velocity, as occurs when it exits the large flow area pressurizer into the small flow area relief piping, rapidly condenses the superheated steam with which it is in contact.
Equilibrium was thus produced at the top of the pressurizer.
As the water level approached the throat of the relief valves, the pressure once again increased.
Flow through the relief 57
0.015 1
0.010 -
/
/
NET l'
COOLANT TEMP
FUEL TEMP a'no3_
j VOID FRACTION l
-DISSOLVED BORON s
F l
C 0.000- m o
s 3
.f \\ ~ -
{.\\ 7
-0.005-
\\v!
- 0.010 -
- 0.015 0
200 400 600 800 1000 1200 1400 1600 1800 2000 TiuE (s)
Fig. 49.
Reactivity feedbacks during base T KMU transient.
2 30 Power decreased by coolant heating
=
25 20-m 3:
v my 15 -
on.
mo 10 -
H o
quilibriurn attained ta" 5-0-
Cooling results in power increase
-5 0
200 400 600 800 1000 1200 1400 1600 1800 2000 TIME (s)
Fig. 50.
Reactor power during base T KMU transient.
y l
58 i
l
1 TABLE XIV EVENT SEQUENCE FOR T KMU BASE TRANSIENT (NO IIPI) 2 Time (s)
Event and Comments O
MFW flow to zero in 1 s 1.0 TSV closed 4.6 Overpressurereactortripsignalfailstotfp, reactor; p > 15.96 MPa (2315 psia) 9.9 PORV opens; p > 16.99 MPa (2465 psia)
~10 Primary SVs open; p > 17.34 MPa (2515 psia) 27.5 AFW begins with full flow of 128 kg/s
-50 Peak primary pressure.(20.50 MPa, 2972 psia)
-110 Boiling in core region begins 220 Primary SVs close 247.5 PORY closes; p < 16.65 MPa (2415 psia) 2000 End of calculation; quasi-steady condition with 11% voided and power of 316 MW removed by AFW.
valves, initiation of AFW flow at 27.5 s and a reduction of power produced a peak pressure of 20.5 MPa (2972 psia) at approximately 50 s.
At 95 s, sufficient steam had flowed into the containment building to cause an overpressure ECC s i gna l.
For the base transient, it was assumed that the HPI system failed.
The pressure decrease along with continued heating of the primary coolant resulted in boiling at approximately 110 s.
Increased expansion when boiling began produced another increase in pressure.
The combination of reduced power and energy removal by the AFW lowered the systen pressure, and the primary SVs closed at approximately 220 s, followed by PORY closure at 247.5 s.
Following closure of the relief valves, the system temperatures and void fractions changed so that a power level was achieved (316 MW) that could be removed by the AFW flow.
A quasi-equilibrium was thus obtained and the calculation was ended at 2000 s.
At this time 59
i l
l i
1.
AFW boiled near the inlet to the tube bundle region, was superheated and then exited through the secondary SVs; 2.
two-phase primary coolant in forced convection transferred the core power to the SGs; and 3.
the average vapor fraction in the core was approximately 11%.
B.
Normal Plant Response This section discusses the transient behavior when the system functioned as designed ard the llPI system began to deliver water to the cold legs at 125 s.
The event sequence is given in Table XV.
As in the base case, sufficient steam flowed through the relief valves to produce a containment overpressure signal at approximately 95 s.
In this case. it was assumed that the IIPI began to deliver water 30 s after the ECC signal at 95 s.
Iligh primary pressures prevented large l
llPI flows until the relief valve flow and energy removal by the AFW at approximately 175 s resulted in a pressure reduction.
The primary pressure and the liquid and saturation temperatures for the transient are given in Figs. 51 and 52, respectively.
After the initial pressure peaks, which are the same as those of the base case, the pressure and temperature decreased and the primary SVs and then the PORV closed.
At approximately 300 s, the pressure increased slightly.
This increase was a result of increasing power before 300 s, which is shown in Fig. 53.
The components of the reactivity that control the power are given in Fig. 54.
Power increased largely as a result of increasing coolant temperature reactivity.
Cooling continued to increase the reactivity as can be seen in Fig. 34.
Additional boron introduced with the llPI water was eventually sufficient to override. the reactivity increases f rom cooling the moderator and reduce the power.
The system initial boron concent rat ion was 1200 ppm.
The boron concentration in the llPI water source was 2200 ppm. so that by 700 s the a
boron concentration in the primary increased to approximately 1300 ppm, and this I
Ak produced a negative reactivity of approximately 0.01
-, as shown in Fig. 54.
k j
Boiling ceased in the core region at 700 s, changing the shape of the depressurization curve (Fig. 51).
Cooling proceeded, as shown in Fig. 52, and at about 1100 s, the subcooling increased substantially as water reentered the pressurizer.
At about the same time, the reactor became permanently subcritical from the increased boron concentration.
At approximately 1460 s, the SG levels approached 509 of the normal operating range and the AFW flow was throttled.
This resulted in increase in the liquid temperature (see Fig. 52) and an 60
-._m e-
,...__,_.__,--...__._...__.___-_-.=m.-
22 AFW cooling sufficient to remove F*'#
20-7 18 -
I v
Pressure maintained as power y
increased briefly m
3 16 -
W W
wm a.
g4 AFW throttled to Core refilled maintain 507. level 12 -
- 4 10 -
ou 8-System subcooled reactor shutdown 6-0 200 400 600 800 1000 1200 1400 1600 1800 TIME (s)
Fig. 51.
Core average pressure during T KMU transient with normal 2
system response.
eeo n
-720 m
b b
W w
640-. [.,
--690 H
n 4
4m g
w wc.
Q-
-880 2
2 620-w w
H L10Vl0 TEMPERATURE
/
H z
,.......... SATURATION TEMPERATURE /
z
,/
-630 O
9
,/
y Q
600-
/
?
-600 W
/
4 m
m 580-y w
o o
-570 4
m m
w y
4 560-w
-s40 m
m x
O o
u w
$40 o
zoo 400 soo soo ioco 12oo i4oo
$soo isoo TIME (s)
Fig. 52.
Core average liquid and saturation temperatures during T KMU 2
transient with normal system response.
61 l
l 30 25-20-m%
s.-
m W
g$ _
3 o
n.
m O
M-Wu Increased boron decreases power b
m 5~
~
Reactor shut down s'
N
\\ Moderator cooling, decreasing void
~
U~
increases power
-5 0
200 400 600 600 1000 1200 1400 1600 1800 TIME (s)
Fig. 53.
Reactor power during T KMU transient with normal system 2
response.
0.020
~ ~ " " " * ' " " ' " " ' " " " -
0.015 -
0.00-
/
l l
0.003
/
- j 0.000-m.
r s
{
-0.005-k 7.'N[./y 5
D
\\/
m - 0.010 -
-y s\\
NET
~ " ~
-COOLANT TEMP \\
~
FUEL TEMP l
-0.020-
. VOID FRACTION
~
- DISSOLVED BORON N
-0.02 5 --
~
0 200 400
-0.030-600 800 1000 1200 1400 1600 1800 i
i TIME (s)
Fig. 54.
Reactivity feedbacks during T KMU transient with normal 2
system response.
62
increase in the rate of system repressurization.
The calculation was terminated at 1734 s.
The primary system heating will continue until the secondary temperature reaches the saturation temperature corresponding to the secondary SV setpoint pressure.
Boiling will then resume on the secondary side, the secondary SVs will open, and a steady state will be achieved.
With no further operator action. the primary system pressure will increase and the PORY will open.
Control of the primary system pressure and level car be maintained by throttling the IIPI flow.
Ilowever, if we compare the boron reactivity feedback (-0.027) to the cooling reactivity feedback (the sum of fuel and coolant (+0.018)) at the we can see that the reactor will r erta in end of the calculation (Fig. 54),
subcritical because a nearly constant rate of negative reactivity insertion has been established.
A stable cooling mode by means of the AFW can thus be achieved with the reactor shut down because of increased boron in the coolant.
TABLE XV EVENT SEQUENCE FOR T KMU TRANSIENT WITil NORMAL SYSTEM RESPONSE 2
Time Event and Comments (s)
(min)
< 125.0
< 2.1 Same as Base Case (Table XIV) 125.0 2.1 IIPI begins but pumps dead head at 20.1 MPa
~220
~3. 7 Primary SVs close 239.7
- 4. 0 PORY closes 700 11.7 Core full 880 14.7 Pressurizer empties as system shrinks from cooling 1100 18.3 Reactor shut down by boron in IIPI water 1450 24.2 AFW throttled when 504 level attained in SGs 1734 28.9 End of calculation, 44 MW decay power, system full and cooled by AFV 63
C.
Equipment Studies In this section equipment failures in addition to those postulated for the base case are considered.
We will discuss the effects of two such equipment failures.
The first failure considered is a 50% reduction of AFW flow at 1500 s in the base case.
By 1500 s in the base case, a quasi-equilibrium situation existed in which reactivity feedback resulted in a reduced-power. but still critical, system.
The power was being removed by AFW flow.
The immediate effect of a 50% reduction in AFW flow is an increase in primary system pressure and temp rature.
This can be seen in Figs. 55 and 56.
The resultant reactivity feedbacks and reactor power are shown in Figs. 57 and 58.
The fuel temperature I
biiefly increased. then decreased as negative reactivity feedback from increased toolant temperature reduced the power.
A new quasi-equilibrium state was produced by 2000 s (33.3 min), in which feedback reduced the power from 317 MW to 151 MW.
In this new state, higher system pressure slightly reduced the void fraction in the core from 0.104 to 0.096.
The fuel temperature decreased from 683 K to 651.5 K (769.7 F to 713 F) and the coolant temperature increased from 616.7 K to 619.7 K (650.4 F to 655.8 F).
Thus, a 50% reduction in AFW flow did j
not produce a vastly different situation than had existed with full AFW flow.
Further reductions in AFW flow would produce a system pressure in excess of the PORV setpoint, PORV opening, and further loss of primary coolant inventory with the additional possible complication of the PORV sticking open after first opening.
In the next transient, we consider another branch f rom the base case that is a result of an equipment failure.
In the base case, energy removal by the l
AFW resulted in depressurizing and cooling, and the PORV closed at 247.5 s.
We now consider a transient in which the t9RV failed to close at 247.5 s.
The event sequence for this transient is given in Table XVI.
The primary continued to depressurize and cool following the PORY failure as shown in Figs. 59 and 60 i
respectively, and this produced an increase in power at approximately 350 s (5.8 min).
The power increase is shown i n F i g. 61.
The reactivity components for the transient are given in Fig. 62.
It can be seen in Fig. 62 that coolant temperature feedback (a positive effect) roughly balanced voiding reactivity (a negative effect), and the power remained approximately constant until about 1500 s (25.0 min).
At 1500 s, voiding react ivi t y was sufficient to shut down the reactor.
Energy flow through the PORY was a sufficient decay heat removal mechanism to allow the AFW flow to fill the SGs and at 2000 s. AFW was throttled 64
21 i
a e
i 20-
~
T n.A l
is -
W 5!
c m-m G.
w
~
Q 17 -
b 4
~
E-g u:o O
15 -
AFW flow reduced 507.
16' 0 tobo 2000 12' 0 14'0 0 10' 0 0
2bo 4bo 6bo 8bo 0
0 0
TIME (s)
Fig. 55.
Core average pressure during T KMU transient with a 509e 2
reduction in AFW at 1500 s.
To 1500 s this transient is identical to the base case.
650 i
^
n b
6
-700 640-
[i, w
m a
5 O
-680 g
g l..
w 630-Q-
a2 3
W AFW flow reduced 50%
-660 2
z 620-U o
D Q
4*
m
-640 n
S 610 -
H LIQUID TEMEPRATURE w
600-
.......... SATURATION TEMPERATURE
--620 b
=W g
4
+
590-y
-500 w"
8 ou 580-i i
0 200 400 600 800 1000 1200 1400 1600 1800 2000 TIME (s)
Fig. 56.
Core average liquid and saturation temperatures during T KMU 2
transient with a 50% reeduction in AFW at 1500 s.
65
(
W 0.015
,e 0.010 -
/
i
\\
./
NET i
COOLANT TEMP 0.00s.
FUEL TEMP
/
-VOlD FRACTION U
- DlSSOLVED BORON 5
e E
0.000-3 U
\\
g
\\
=
/ k. - ~ -.-...
I
-0.005-
)
f/
\\
\\
}
N.
',~
k!
- 0.010 -
v
- 0.015 -
600 800 1000 1200 1400 1600 1800 2000 0
200 400 Tiut (s)
Fig. 57.
Reactivity feedbacks during T KMU transient with a 50%
2 reduction in AFW at 1500 s.
30 25-20-m v
m W
15-M On.
m i
O 10 -
W U<w" s.
0-AFW flow reduced 507.,
0 200 400 600 800 1000 1200 1400 1600 1800 2000 TIME (s) l'i g. 58.
power during T KMU transient with a 50% reduction in Reactor 2
l AfW at 1500 5.
1
{
66
22 i
i i
-3000 20-n g
2700
.D 18 E
.y
'j PORV should close but retroins open 2400 w
S d
in m
M
-2100 a.
o.
g.
Power increases as system cools u
w o
o g
1800
- f 12-d
- g 3{
4
-1500 y
y
~
AFW hrottled rt:
o:
10 -
O oo u
8-1200
-900 G.
i i
i i
o 500 1000 1500 2000 2500 3000 TIME (s)
Fig. 59.
Core average pressure during T KMU base transient with y
failure of the PORV to reclose at 247.5 s.
eso Q
-70
[
v v
.4a-3 W
W b
--87s g
gg_
p 5
{
-450 3
s20-0-
g>
sio-6 Incmsed power slows cooling y-
-s2s m-Q 5
s Q
m-
--o Q
m W
LIQUl0 TZ wERATURC o
- ~
- ---- SATURAfl0N TEMPERATURE '
~
0 g
-sis y
u
~
w s70-g W suldown by voeding
.g g
\\
--sso u
AFW throttled u
see o
soo icoo stoo 200o 2500
- 00 TIME (s)
Fig. 60.
Core average liquid and saturation temperatures during T KMU 2
base transient with failure of the PORY to reclose at 247.5 s.
67 i
30 25-20-m 3:
v xy 15 -
on.
m O
10 -
U<
Void reactivity shuts down reactor w"
5-o-
Power increases as system cools
-5 o
500 1000 1500 2000 2500 3000 TIME (s)
Fig. 61.
Reactor power during T KMU base transient with failure of the 2
PORY to reclose at 247.5 s.
0.03 0 02*
' w.
..........................;,----(..................................
0.01-
/./
/:
0.00-
/
\\
t
/
(.,./
A..
-0.0s-O N.
%.s s 4w E
-0.02 -
NET
' - ~....
~ ~~
. N
~
-0 03 COOLANT TEMP
.......... FUEL TE MP VOID FRACTlON
.o.o 4
- DISSOLVED BORON
-0.05 -
o too 1000 tboo 2000 2500 Moo TIME (s) i l
l'ig. 62.
Reactivity feedbacks during T KMU base t ransient with failure 2
of the IURV to reclose at 247.5 s.
68 t
I i
.___.-r-
to mr.intain 50% or' normal operating level.
This resulted in an increase in pressure and temperature and established an equilibrium condition with the pr ima ry. sys t enif a t 7. 5 MPa (1088 psia) and $64 K (556 F).
This condition was maintained to the end of the calculation at 3000 s (50 min), as can be seen in Figs. 59 and 60.
Akthe end of the calculation, the decay power was 37 MW and thecorewdsvoidingat'a rate of 4.5 x 10~3 9/s.
The voiding rate is shown in Fig. 63.
This voiding, rate would result in complete uncovery of the core at 16000 s (4. 4 h). ' Several things would happen, however, to accelerate the voiding rate. s At some point, the void f raction would be too large to operate 4
the RCPs, and they would have to be turned off.
Phase separation that would follow would result in vastly reduced heat transfer to the SGs.
The only means of; energy removal would then be flow through the PORV; thus, inventory depletion wou!J 'be accelerated.
To prevent core uncovery, the operator would have to close either the PORY or the block valve downstream of the PORV and/or initiate llPI cooling.
TAllLE XVI EVENT SEQUENCE FOR T KMU 11ASE TRANSIENT 2
j WITil FAILURE OF Tile PORY TO RECLOSE AT 247.5 S Time Event and Comments (s)
(min)
< 247.5
< 4.1 Same as base case 247.5 4.1 p < 16. 65 MPa ( 2415 ps i a ). PORY shou ld close but remains open 350
- 5. 8 Power increased by cooling 1500 25.0 Reactor shut down by void fraction of 0.30 2000 33.3 AFW throttled as 50% level is achieved in SGs t
3000 50.0 Calculation ended, 37 MW decay power 16000 267 Core completely voided (linear extrapolation of void rate) 69 y
_-_+-y.--,,
-,_w
i D.
Conclusions for ATVS The ATVS studies have provided insights about the course of the event, event timing, and the importance of equipment availabilities.
The conclusions for the T KMU transient follow.
2 1.
An ATVS event will result in high primary system pressure that could induce a small break LOCA by means of a pump seal failure, SG tube rupture or the failure of some other system component.
2.
If there is no failure, reactivity feedback reduces the power to a level that can be removed by AFW flow through the SGs.
3.
If the system responds as designed to a containruent overpressure signal or the operator intervenes at 95 s (1.6 min) and actuates the ilPI system, boron in the ilPI water will shut down the reactor and energy removal by AFW is assured.
4.
If the PORY fails to close and no llPI is provided, reactor uncovery is assured by 16000 s and will probably occur sooner because of loss of the SGs as an energy sink.
~
VII.
INTERFACING SYSTEMS LOCA A.
Introduction The primary system of a PWR operates at a relatively high pressure (15.5 MPa, 2250 psia) and consists of piping and components designed to withstand these pressures.
The I.PIS connects to the primary system but possesses low-pressure piping passing outside the containment.
Therefore, a potential exists for a LOCA out s ide the containment and concurrent damage to systems needed to cope with this problem.
Such a LOCA sequence assumes failure of a series of two check' valves in one of the LPIS liu
- and the opening of a normally closed isolation motor-opera ted valve (MOV) which is also in series with the check valves.
This would allow high-pressure coolant water to enter the low-pressure piping outside the containment and rupture the pipe.
The containment-engineered safety systems would be ineffective for this accident, the LPIS would partially fail (half the capacity), and the CFT (accumulators) on the damaged side would be available but would not be effective.
As a result, RSSMAP studies show that core meltdown is possible.
ECC injection is assumed to l
be available for this event.
l l
l 70 l
l 1
i A set of calculations was performed using the TRAC-PF1 code and a model of i
the Oconee-1 PWR to investigate the consequences of. and possible operator i
actions for. such an accident scenario.
Because the source of once-through 2
cooling is limited to the size of the BWST. which has a capacity of 1468.7 m (388000 gal).
a limited amount of time is available to the operator to discover l
alternate sources of borated water for core cooling.
Both the llPIS and the LPIS 1
l use the BWST as their source of borated water.
Both safety-injection systems (SIS) are actuated by a low primary-system pressure signal.
The HPI is 4
l initiated at 11.1 MPa (1615 psia) and the LPI at 1.45 MPa (210 psia).
The CFTs automatically flood the_ core when the RCS pressure drops below 4.2 MPa (612 psia).
The fundamental objective defined for the operator in this i
investigation was to extend the time interval that the SIS water supply is available while maintaining core cooling.
Therefore, a base case without the i
safety. injection (SI) was calculated to establish event timing.
Other cases involved operation of all SISs (both HPIS
- lines, one LPIS line and accumulators), of both HPISs and no LPIS, and of only one HPIS line.
The accumulators were used in all of the calculations except for the base transient.
This accident sequence is designated the "V"
sequence.
More details can be found in Ref. 1.
A diagram of the noding scheme used to mcdel the plant, i
plus other modeling details, can be found in Appendix D.IV.
B.
Base Case In this calculation, SI and the accumulators were unavailable.
Therefore, l
11PIS, LPIS. and the accumulators were inactive.
The transient was calculated to establish event timing.
The sequence of events is tabulated in Table XVII.
[
This transient was initiated by a 0.254-m (10-in. ) diameter break in the LPIS piping.
The check valves that normally protect this line were assumed to fail.
The primary system rapidly depressurized as shown in Fig. 64.
The initial flow through the break was 4000 kg/s (3.175 x 10' lb,/hr) (Fig. 65).
Primary-system water inventory was depleted by 150 s and core heatup started at a rate of 1.90 K/s (3.42 F/s) at 129 s as shown in Fig. 66.
The temperature at which core damage begins (1000 K, 1340 F) was reached at 385 s (6.4 min).
The hot-rod maximum temperature of 1000 K (1340 F) was reached at 253 s (4.2 min).
Therefore. if no safety injection (including the accumulators) is provided, core damage is certain by about 400 s (7 min).
If the accumulators are used, but no SI. core damage will occur about 480 s (8 min) af ter the break.
i i
71 I
1
-m__.-._. _., _ _...,.. _ _. _
1.03 l
1-.
z 9
0.9 5 --
eo<
{
0.90-La 3
0.85-
_a O>
0.80-5o Voiding shuts down reactor o,73 s
LaJ 0.70-o 0.65 --
Voiding continues on
.I decay power I
0.60 g
0 500 1000 1500 2000 2600 3cco TIME (s)
Fig. 63.
Core liquid volume fraction during T KMU base transient with 2
failure of the PORV to reclose at 247.5 s.
to sAsJ CASE 14-Doni HPl$a
" bat 00
.......... una + ina
^
- oneuns m
g
S uxn x
,2 v
v 6
2
- tuo
!g
- s 60-51
[
o c.
s.-
. -i200 g
u i.ri
(*.
initiated Q
e.
y e-
. m 3
g,,
g tom.ted g
i 4
More steam producuan
. -600 sn 31 case thaa h.
3-the base case Qn 2-Wore condensation
- -300 6
g in Si case 5=
4 b.ecause of cold A.
ater entertas 0-.
the core
- -0
-a O
SO 100 150 200 250 000 350 400 4',0 Timn (s)
Fig. 64.
Primary-systerr. pressure during V sequence base and paramet ric cases.
72
l TABLE XVII INTERFACING SYSTEMS LOCA EVENT SEQUENCE (BASE CASE)
Time Event (s)
(min)
- 0. 0 LPIS break (0.305 m (12 in) diameter double-ended break)
- 3. 6 Reactor scram caused by low primary-system pressure (0.5 s delay for control rod insertion) 4.1 TSV closed
- 7. 8 0.1 Auxiliary feedwater initiated 129.4
- 2. 2 Start of core heatup 200
- 3. 3 End of blowdown 253 4.2 Maximum hot-rod temperature exceeded temperature for fuel pin damage (1000 K, 1340 F) 385
- 6. 4 Start of core damage 400
- 6. 7 End of calculation C.
Normal Plant Response The SIS was active in this calculation.
Both IIPISs. both accumulators, and one LPIS (the other LPIS line was damaged by the break) were active.
The sequence of events is listed in Table XVIll.
The primary rapidly depressurized to the llPIS setpoint (11.14 MPa, 1615 psia) at 35 s.
As seen in Fig. 64, the pressure history for the transient is dominated by the break size and not the SI characteristics.
Therefore, the depressurization history is similar to the base case.
The accumulators began flooding the core at 120 s when their springloaded check valves opened because of low pressure in the pr imary (4. 22 MPa, 612 ps ia ).
The accumulators dumped a total of 5.95 x 10' kg (1.31 x 10' lb )
f borated m
water into the vessel.
The break flow was sufficiently large to depressurize the primary system to the LPIS setpoint and it was actuated at 155.0 s.
The 73
combined ilPIS and LPIS supplied 278 kg/s (2.21 x 10' lb /h) to the vessel.
The n
llWST was used to supply SI water.
This tank has a volume of 1468.7 m (388000 gal).
When both IIPIS and LPIS are active, the 51 inventory will last 5230 s (1.45 h) before the operator must have other sources of borated water available for core cooling.
The maximum cladding temperature of the average rod is illustrated in Fig. 66.
The core temperature decreased to 433 K (320 F) at 198 s and remained at this level.
Core cooling can be maintained until the llWST is empty.
D.
Operator Intervention (No LPIS)
For this case, the operator deactivated tne LPIS to conserve on BWST inventory.
Iloth IIPISs and both accumulators were active.
The sequence of events is given i n Table XIX.
The first 150 s of the transient were the same as for the previous case.
Accumulators were emptied by 176.2 s and 188.4 s.
The combined ilPI flow was 69.2 kg/s (5.49 x 10' l b,/ h ).
At this flow, the llWST supply would last for 21000 s (5.84 h).
The core temperature history illustrated in Fig. 66 is also similar to the normal response case.
The maximum cladding temperature of the average rod was reduced to 433 K (320 F) by depressurization overcooling and SI flow.
Stable core cooling was maintained to the end of this calculation.
Therefore. operation of both IIPISs p rovi.:ed adequate core cooling and increased the length of time for which SI was provided.
ii.
Operator Intervention (No LPIS and One IIPIS)
For this case. the LPIS and one liPIS were turned off by the operator to
- onserve IlWST supply.
The sequence of events is given in Table XX.
The first 35 s of the transient were identical to those of the previous case.
IIPIS was actuated at 37.8 s because of low primary-system pressure.
Primary-system inventory was lost rapidly, and at 129 s blowdown terminated and the reflood phase began.
The accumulators were emptied by 169 s and 175 s.
The maximum cladding temperature of the average rod (Fig. 66) decreased to 437 K (327 F) at 200 s and continued to decrease.
Therefore, flow from one llPIS pump was sufficient to maintain core cooling.
The cooling was more unstable in this case because of the low SI flow and boiling heat transfer with a large void fraction in the core.
It is estimated that the llWST inventory would last for about 4200 s (11.7 h) for this case.
74
c SOG) e BASE CASE 4500-
[ 7 '.* ~.I.~.~.* ~.
35000000 m
- m ms
~
4000- i 3'X)D -
m
-25000000.C m
N N
3000-
'.[
E m
04
.3
, M Jt.
si initiated 2000-1PIS initiated g
5
~
- 15(XXXX)0 2G30-Accumulator m
=
tnjectaan j
M as 13(X)-
began g
-10000000.4 1000~
Accumulators empty ir annnnnn
~
g-f,
~
Ii L
_2 -.~.,,
.a s
2n..
..g 0
00 lb0 gh M
N 300 b
4b Time (E) rig. 65.
fireak flow during V sequence base and parametric cases.
11 3 )
i i
BASE CASE
-1400
-.......... uris + tras 1000-D0DI HPian
- ouE aires
^
^N b
v 1200 v
M-e s
bU U
a "N
~
- -1000 g
B00-E E
E E
h 733-- HPI Accumulator
- -800 0b tnatiated injection e
bagna go u
LPI intilated
~ -600 E'
Accumulatore empty g
4i lt 500-i g
-400 g
M
~
RCPs off U
*...a
_ w w.kjl,,,
-(,,
O
~
End of blowdown,
~
start of rerkas 200 300 O
SO 100 150 20 250 000 350 400 450 Time (s)
Fig. 66.
Maximum cladding temperature of average rod during V sequence base and parametric cases.
75
TABLE XVIII INTERFACING SYSTEMS LOCA EVENT SEQUENCE (NORMAL PLANT RESPONSE CASE)
Time Event (s)
(min)
- 0. 0 LPIS break
[0.305-m (12-in.) diameter double-ended break) 3.6 Reactor scram caused by low primary-system pressure (0.5-s delay for control rod insertion) 4.1 TSV closed
- 7. 8 0.1 AFW initiated 37.8
- 0. 6 IIPIS actuated 97.8
- 1. 6 RCP off (60 s af ter IIPIS actuation) 116.5 1.9 Accumulator A check valve first opens 128.0 2.1 Accumulator B check valve first opens 138.3
- 2. 3 End of blowdown, start of reflood 155.0
- 2. 6 LPIS actuated 176.2
- 2. 9 Accumulator B empty 188.4 3.1 Accumulator A empty 412.0
- 6. 9 End of calculation F.
Conclusions for Interfacing Systems LOCA The interfacing systems LOCA is an extremely severe accident sequence in the Oconee-1 PWR.
Failure of the check valves. separating the high pressure primary system from the low pressure ECC system, causes failure of the low pressure ECC pi pi ng in the penetration room.
If failure of this piping is 76
TABLE XIX INTERFACING SYSTEMS LOCA EVENT SEQUENCE (OPERATOR INTERVENTION CASE--NO LPIS)
Time Event (s)
(min) 0.0 LPIS break
[0.305-m (12-in.) diameter double-ended break]
- 3. 6 Reactor scram caused by low primary-system pressure (0.5-s delay for control rod) insertion) 4.1 TSV closed
- 7. 8 0.1 AFV initiated 37.8 0.6 IIPIS actuated 97.8
- 1. 6 RCP off (60 s after llPIS actuation) 116.5 1.9 Accumulator A check valve first opens 128.0 2.1 Accumulator B check valve first opens 138.3
- 2. 3 End of blowdown, start of reflood 170.3 2.8 Accumulator A empty 180.2
- 3. 0 Accumulator B empty 277.5
- 4. 6 End of calculation severe enough to preclude pumped ECC injection, core meltdown is certain.
The penetration room will fail almost immediately after check valve failure because it is not designed to withstand blowdown pressures.
Therefore, the containment building is bypassed during blowdown, and there is a direct leakage path to the atmosphere for the release of fission products.
The check valve failure was assumed to produce a large LOCA [0.254-m-diam (10-in.) break) in the penetration 77
TABLE XX INTERFACIlXi SYSTEMS LOCA EVENT SEQUENCE (OPERATOR INTERVENTION CASE--No LPIS AND ONE IIPI)
Time Event (s)
(min)
- 0. 0 LPIS break
[0. 305-m (12-in. ) diame t e r double-ended break)
- 3. 6 Reactor scram caused by low primary-system pressure (0.5-s delay for control rod insertion) 4.1 TSV closed
- 7. 8 0.1 AFW initiated 37.8
- 0. 6 One llPIS actuated 97.8 1.6 RCP off (60 s af ter llPIS actuation) 117.6
- 2. 0 Accumulator A check valve first opens 127.9 2.1 Accumulator B check valve first opens 129.4
- 2. 2 End of blowdown, start of reflood 169.2
- 2. 8 Accumulat'or A empty 174.8
- 2. 9 Accumulator B empty 366.3 6.1 End of calculation room.
The event tree for calculations performed here is illust rated in Fig. 67.
Four cases were studied.
The duration of effective core cooling for each of the cases is summarized in Fig. 68.
The CFTs or accumulators are passive safety systems and were available to mitigate the transient.
The flow out of the accumulator on the damaged side was leaving through the break as it was opposing the flow out of the core.
Thus, only one of the two accumulators was effective 78
TIME (s) 0 100 200 300 400
-REACTOR SCRAM (3..)
-TSV C10 SED (4.9.)
A ur ko wu o og S.)
QM.)
(MS.)
- AfW listil&TED DA.)
De sAK TRANSIENT
.c, 3., mo.D Cf EM oo
(=
on.)
{ ACOAAAATOR-8 MTV (l7..)
gog wig 00fM De ynd!S,,,,
f inirEt gl L m...
.A,,,, (
.)
-na
<=
A.
A ACCUMLLA704 yg
- LP11 ihlfiAIED (ISS.)
oi, in.)
- Accatance.A twtv ore.)
Ec"ufdYa'as
_ pe L cemwon-e un peo.)
A on Mf*'
E =.)
.c,. o.),,
(se As EnEl*ltons Es
- pe 5'4"'
'cs'^d'#
w d;P'. T."!)
Fig. 67.
Oconee-1 V sequence event tree.
100000 (L43 h) pty in 6230 s DWST em
-21E DA$E CASE e',
.......... IIPts + LPl3
....... DODI HPIS.
W empty in 21000 s
,,p N~
g V
LP! inluated g'.~"'"'
y o
-t40000 C
~
"g DWST empty in 42000 s C
r,,,,
g, c
(lL7 h) c,
..o.
. tmnnn 3 c
40000-
,O c
Accumulator dump
.70000
>i s
d 20000-
<.a.
n
~
M HPI imtlated frh
'3 u e....-
oF
- -0 C
0 b
~"
-20000-O SO 100 150 200 250 000 3$0 4b0 450 Time (s)
Fig. 68.
Duration of effective core cooling during V sequence base and parametric cases.
79
in core cooling.
The parametric cases were calculated to determine the minimum flow necessary to cool the reactor core.
It was determined that one llPIS line is sufficient to provide stable core cooling until the BWST inventory is completely depleted.
In this configuration, t he BWST ca pa c i t y i s sufficient for 42000 s (11.7 h) of once-through cooling.
A parametric case that was not calculated is the case in which no safety injection is available, but the accumulators are available.
By inspection of the other parametric cases, we can predict that core damage in this case is possible within 8 min after the break has occurred.
Long-term core cooling is a concern because of the limited BWST capacity.
The containment sump recirculation system is also unavailable.
If the LPIS MOV can be closed in time, a feed-and-bleed operation may be successful.
Iloweve r, the probability of this valve closing after a two-phase blowdown is small.
The core-damage sequence and fission product transport and release could not be predicted with TRAC.
Other codes, such as MIMAS, are being developed at Los Alamos that can predict degraded core behavior.
The Oconee calculation can be continued with MIMAS to predict the core-degradation sequence and fission product transport.
This study is recommended for future work.
VIII.
SUMMARY
CONCLUSIONS The work reported herein is of importance in two areas: first, as part of an integral effort to identify and verify the dominant accident sequences for a particular reactor, Oconee-1, and second, by its focus on and delineation of a specific set of important transients.
Probabilistic risk assessments are conducted to identify those transients, equipment, and procedures that are the greatest contributors to the risk associated with specific nuclear power plants.
The quality of the PRA is directly related to the quality of the data base available to the analyst.
In addition to a data base identifying failure rates for equipment and human factors information, the course of the transient must be understood to determine what equipment will be challenged, the timing, and the decisions the operator will face.
Therefore, deterministic analyses of dominant accident sequences are an integral part of upgrading the data base for accurate risk assessments.
An initial PRA may be performed to identify the r i s k -dom i.na n t sequences.
If deterministic analyses are lacking, assumptions about the course of the accident are made.
Subsequent deterministic analyses that focus on the dominant 80
sequences then improve the data base and permit the PRA analyst to reassess risk.
We have completed deterministic analyses of four of the dominant accident sequences identified for the Oconee-1 nuclear power plant.
As reported in the conclusions for each sequence evaluation, insights about the course of the transient, event timing, the importance of equipment availabilities and the impact of selected operator actions have been examined.
This information will be useful the vendor, owners of plants similar to Oconee-1, and the NRC.
In addition to the plant, equipment, and procedural insights obtained, the calculated transients should be useful for refining or confirming the assumptions made in performing the Oconee-1 PRA.
We have examined four of the Oconee-1 dominant accident sequences.
These are the T (B )MLU, LOFW induced LOSP; T KMU, LOFW without scram:
S II' 3
3 2
3 small-small break LOCA; and V interfacing systems LOCA.
The reader is referred to Secs. IV.E. V.E VI.D, and VII.F. respectively, for conclusions specific to these transients.
REFERENCES 1.
Gregory J. Kolb, Steven W.
Ila t c h,
Peter Cybulskis, Roger O. Wooton,
" Reactor Safety Study Methodology Applications Program:
Oconee #3 PWR Power Plant," NUREG/CR-lo59 (2 out of 4), Sandia National Laboratories (May 1981).
2.
M. S. Sahota and J. F. Lime,
" TRAC-PF1 Choked Flow Model,"
Thermal liyd rau l ic s of Nuclear Reactors, 1, 480.
Papers presented at the Second International Meeting of Nuclear Reactor Thermal-llydraulics at Santa Barbara, California, US (January 11-14, 1983).
3.
John R.
Ireland and Rudolph J. llenninger, " Analyses of B&W Small-Break LOCA TRAC Ca l c u la t i ons, " Los Alamos National Laboratory document LA-UR-82-3294 (June 1976).
4.
Babcock and Wilcox, " Abnormal Transient Oper ating Guidelines," BWP-20004 (June 1976).
5.
NRC Staff, " Anticipated Transients Without Scram for Light Water Reactors,"
NUREG-0460 (Apri1 1978).
6.
R. F. Barry.
" LEOPARD - A Spectrum-Dependent Non-Spatial Depletion Code for the IBM-7094," West inghouse Report WCAP-3296-26 (September 1963).
81
APPENDIX A TRAC VERSION I.
INTRODUCTION The Transient Reactor Analysis Code (TRAC), an advanced best-estimate systems code for analyzing light water reactor accidents. is being developed at the Los Alamos National Laboratory under the sponsorship of the Reactor Safety Research Division of the US NRC.
A preliminary TRAC version consisting of only one-dimensional components was completed in December 1976.
Although this version was not released publicly nor documented formally, it was used in TRAC-Pf1 development and formed the basis for the one-dimensional loop-component modules.
The first publicly released version, TRAC-PF1, was completed in December 1977.
It is described in Los Alamos National Laboratory report LA-7279-MS (lune 1978).
The TRAC-P1 program was designed primarily for the analysis of large-break LOCAs in PWRs.
Ilowe ve r, because of its versatility, it could be applied directly to many analyses ranging from blowdowns in simple pipes to integral LOCA tests in multiloop facilities.
A refined version, TRAC-PIA, was released to the National Energy Sof tware Center in March 1979.
It is described in Los Alamos National Laboratory report LA-7777-MS (May 1979).
Although it still treats the same class of problems. TRAC-PIA is more efficient than TRAC-P1 and ir.corporates improved hydrodynamic and heat-transfer models.
It also is easier to implement on various computers.
TRAC-PD2 contains improvements in reflood, heat-transfer models, and numerical rolution methods.
Although a large-break LOCA code, it has been applied successfully to small-break problems and to the TMI incident.
TRAC-PF1 was designed to improve the ability of TRAC-PD2 to handle small-break LOCAs and other transients.
TRAC-PF1 has all of the major improvements of TRAC-PD2; in addition, it uses a full two-fluid model with two-step numerics in
{
the one-dimensional components.
The two-fluid model, in conjunction with a stratified-flow regime, handles countercurrent flow better than the drift-flux model used previously.
The two-step numerics allow large time steps for slow transients.
A one-dimensional core component permits calculations with reduced dimensionality although the three-dimensional vessel option has been retained.
A noncondensable gas field has been added to the one-and three-dimensional 82
hydrodynamics.
Significant improvements also have been made to the trip logic and the input.
TRAC-PF1 was released publicly in July 1981.
It is described in Los Alamos National Laboratory report LA-9944-MS (February 1984).
TRAC-PF1/ MOD 1 provides full balance-of-plant modeling through the addition general capability to model plant control systems.
The SG model was of a replaced to allow a wider variety of feedwater connections and better modeling of SG tube ruptures.
New components were not required to model condensers, heaters, and pumps in the secondary system; however, a special turbine component was added.
The TRAC-PF1/ MODI physical models also have been modified; the condensation, model contains the most significant changes.
During condensation the liquid-side interfacial heat-transfer coefficient (IITC), which is sensitive special model for thermally stratified to the flow regime, includes a
configurations.
Wall heat transfer in the condensation and film-boiling regimes has been improved.
The motion equations include momentum transport caused by phase change, and the momentum flux terms in the three-dimensional flow equations have been modified.
This last modification can change substantially the computed pressure drop across a vessel from that calculated by previous codes.
These model changes make TRAC-PF1/ MOD 1 not only a superior code for small-break and operational transients, but also for large-break analyses.
A.
TRAC Characteristics Some distinguishing characteristics of TRAC-PF1/ MODI are summarized below.
Within restrictions imposed by computer running times, we are incorporating state-of-the-art technology in two-phase thermal hydraulics into the code.
1.
Variable-Dimensional Fluid Dynamics.
A three-dimensional (r,6,z) flow calculation can be used within the reactor vessel; the flow within the loop components is treated one dimensionally to allow an accurate calculation of the complex multidimensional flow patterns inside the reactor vessel that are important in determining accident behavior.
For example, phenomena such as ECC downcomer penetration during blowdown, multidimensional plenum and core flow effects, and upper-plenum pool formation and core penetration during reflood can be treated directly, llowever, a one-dimensional vessel model may be constructed that allows fast transient calculations because the usual time-step restrictions are removed by the special stabilizing numerical treatment.
2.
Nonhomogeneous. Nonequilibrium Modeling.
A full two-fluid (six-equation) hydrodynamics model describes the steam-water flow, thereby allowing important phenomena such as countercurrent flow to be treated explicitly.
A 83
stratified-flow regime has been added to the one-dimensional hydrodynamics, a seventh field equation (mass balance) describes a noncondensable gas field, and an eighth field equation tracks the solutes in the liquid.
3.
Ilow-Regime-Dependent Constitutive Equation Packape.
The thermal-hydraulic equations describe the transfer of mass, energy, and momentum between the steam-water phases, and the interaction of these phases with the heat flow from the system structures.
lle c a u s e these interactions are dependent on the flow topology, a flow-regime-dependent constitutive equation package has been incorporated into the code.
Although this package undoubtedly will be improved in future code versions, assessment ca,1culations performed to date indicate that many flow conditions can be handled adequately with the current package.
4.
Comprehens ive llea t -Trans f e r Ca pabi l i t y.
TRAC-PF1/ MODI incorporates detailed heat-transfer analyses of the vessel and the loop components.
Included is a two-dimensional (r,z) treatment of fuel rod heat conduction with dynamic fine-mesh rezoning to resolve both bot tom-flood and falling-film quench fronts.
The heat transfer from the fuel rods and other system structures is calculated using flow-regime-dependent ilTCs obtained from a generalized boiling curve based on local conditions.
5.
Consistent Analysis of Entire Accident Sequences.
An important TRAC feature is its ability to address entire accident sequences, including computation of initial conditions, with a consistent and continuous calculation.
For example, the code models the blowdown, refill, and reflood phases of a LOCA.
This modeling eliminates the need to perform calculations using different codes to analyze a given accident.
In addition, a steady-state solution capability provides self-consistent initial conditions for subsequent transient calculations.
Both a steady-state ard a transient calculation can be performed in the same run, if desired.
6.
Component and Functional Modularity.
The TRAC program is completely modular by component.
The components in a calculation are specified through input data; available components allow the user to model virtually any PWR design or experimental configuration.
Thus, TRAC has great versatility in its range of applications.
This feature also allows component modules to be improved, modified. or added without disturbing the remainder of the code.
TRAC component modules currently include accumulators, breaks and fills, cores, pipes, pressurizers, pumps,
- SGs, tees, turbines, valves, and vessels with associated internals (downcomer, lower plenum, core, upper plenum, etc.).
84
The TRAC program is modular by function; that is, the major aspects of the calculations are performed in separate modules.
For example, the basic one-dimensional hydrodynamics solution algorithm, the wall-temperature field solution algorithm, llTC selection, and other functions are performed in separate sets of routines that are accessed by all component modules.
This modularity allows the code to be upgraded readily as improved correlations and test information become available, 11.
Physical Phenomena Treated Because of the detailed modeling in TRAC, the code can simulate physical phenomena important in large-and small-break LOCA analysis, such as 1.
ECC downcomer penetration and bypass, including the effects of countercurrent flow and hot walls; 2.
lower-plenum refill with entrainment and phase-separation effects; 3.
bottom-flood and falling-film reflood quench fronts; 4.
multidimensional flow patterns in the core and plenum regions 5.
pool formation and countercurrent flow at the upper-core support-plate region:
6.
pool formation in the upper plenum; 7.
steam binding; 8.
average-rod and hot-rod cladding-temperature histories; 9.
alternate ECC injection systems, including hot-leg and upper-head injection;
- 10. direct injection of subcooled ECC wa t e r, without artificial mixing zones;
- 11. critical flow (choking);
- 12. liquid carryover during reflood;
- 13. metal-water reaction;
- 14. water-hammer effects;
- 15. wall friction losses; and
- 16. horizontally stratified flow, including reflux cooling.
C.
Significant Chances from Previous TRAC Versions Substantial changes have been made to all input sections that cover the general control capability.
use of trips and signal variables to provide a more The initial definition cards for almost all components were modified.
The definitions of the wall heat sources and the direct power to the fluid were 85
changed.
The capabilities of the input loss coefficients and hydraulic diameters were expanded.
Four significant model changes should be considered when the TRAC-PF1/ MODI results are compared with those from previous TRAC versions.
Many details that affect the core reflood were changed;
- thus, the TRAC-PF1/ MOD 1 refl.od calculations are significantly better than those from TRAC-PF1.
The condensat ion mode l improvements include a more detailed flow-regime dependence and a special model for condensation on a stagnant interface in a vertical one-dimensional component.
Also, the choked-flow model was improved.
Finally, the axial-momentum t ransport terms in the vessel were modified so that internally generated losses at orifices were eliminated.
Loss coefficients probably will be required at core support plates to match results from previous TRAC versions.
D.
Planned Improvements TRAC-PF1/ MODI combines all of the PWR accident analysis capabilities thus far requested by the NRC into a single code.
This code re; resents such an important milestone in the TRAC series that future versions will be designed to accept any TRAC-PF1/ MOD 1 input deck.
Most of our short-term plans involve increasing the speed af the code end making it more flexible and convenient for the user.
Recent studies indicate that a change in the data base combined with recoding some key subroutines will cut run times by at least 301.
Most user-convenience features are being developed under the Nuclear Plant Analyzer project.
Iloweve r, we expect to 6
simplify the standa rd TRAC input by adding an option that allows input of the trips and controllers in FORTRAN-like code.
II.
TRAC VERSIONS USED IN Tills STUDY Because the transients studied in this report were calculated throughout a fairly long time interval (~1 year), different TRAC versions were used for each study.
The following table summarizes the TRAC versions used.
Number of Computational Cells Calculation TRAC Version System Vessel LOFW TRAC-PF1 Version 8.3 260 80 LOFW initiated ATWS TRAC-PF1/ MOD 1 Version 11.0 209 40 Small-Small Break LOCA TRAC-PF1 Version 9.7 317 120 Interfacing Systems LOCA TRAC-PF1/h0D1 Version 11.0 235 40 86
APPENDIX 11 SENSITIVITY OF RESULTS TO UNCERTAINTIES The numerical simulation of severe accidents requires models of plant characteristics and thermal-hydraulic phenomena.
Information on the plant condition at the start of the transient and on the performance capabilities of equipment actuated during the transient is used with these models to predict the accident progression.
Uncertainties in models, which approximate the physical characteristics and thermal-hydraulic
- behavior, and information on plant features and operating conditions can lead to uncertainties in the predicted results.
For these results to be meaningful, the effects of these uncertainties must be recognized.
Uncertainties can be grouped into three categories:
(1) plant state (input information) uncertainties, (2) model uncertainties, and (3) phe-nomenological uncertainties.
Uncertainties in the plant state include both the initial conditions for the transient (for example, operating power level, power history, control-rod reactivity, makeup and letdown flows, and SG inventories) and the performance characteristics of equipment actuated by the plant response (for
- example, llPI
- flows, pressurizer heater / sprayer capacities, and P/T setpoints).
Plant state uncertaint.es can be reduced or bo'inded by probabilistic treatment of the initial conditions and by developing b er input information on equipment performance.
Model uncertainties result from inaccuracies in the mathematical representation of physical features (for example, operation of valves and pumps) and from approximations or simplifying assumptions made by the analyst to improve calculational efficiency.
The reduction of modeling u., certainties requires more detailed experimental data on which the development of more accurate models can be based and more detailed representations of plant systems using available models.
Phenomenological uncertainties result from incomplete information or mathematical formulations describing the phenomena (for example, flow. of steam / water mixtures through steam relief va l ve s ).
Reductions in phenomenological uncertainties would require both new experimental data and development of models based on improved mathematical formulations.
87
APPENDIX C TIMING STATISTICS A summary of computer time used, number of time steps taken, and ratios of computer time to reactor time are given in the following table.
Ratio of Computer Computer Reactor Time to Number of Time Time Reactor Calculation Time Steps (s)
(s)
Time LOFW Base Case 11317 10861 3759 2.9 Normal Plant Response 31107 27084 5000 5.4 Both IIPl at 120 s 18276 15692 3563 4.4 Early Feed and Bleed 29540 30966 16052 1.9 Subtotal 90249 84603 33374
- 2. 5 Small-Small Break 102 mm Break 45218 28800 2364 12.2 31 mm Break 34295 22320 10000 2.2 22 mm Break 26685 15120 10000 1.5 11 mm Break 43077 20880 10000 2.1 Subtotal 149275 87120 32364
- 2. 7 LOFW-i n i t i a t ed AWS Base Case 36185 20967 1993 10.5 ECC at 125 s 27380 14717 1609 9.1 Stuck-Open PORV 40503 23004 2775
- 8. 3 1/2 AFW at 1500 s 8617 4841 500 9.7 Subtotal 112685 63529 6877 9.2 Interfacing Systems LOCA Base Case 9541 2722 398 6.8 Normal Plant Response 22750 20200 414 48.8 No LPI 18040 16630 277 60.0 No LPI and One IIPI 21690 18800 368 51.1 Subtotal 72021 58352 1457 40.0 Total 424221 293604 69072 4.3 (81.6 h)
(19.2 h)
Average Time Step Size = 0.16 s 88
APPENDIX D INPUT MODELS The TRAC mode l s used were slightly dif ferent for each calculation.
The primary difference was in the number of azimuthal nodes used in the three-dimensional vessel.
This was done for computational efficiency and economy.
The individual models are discussed next.
1.
TRAC-PF1 OCONEE MODEL FOR T (B )MLU SEQUENCE TRANSIENTS j
3 Oconee-1 is a Babcock and Wilcox lowered-loop PWR and consists of a vessel, two once-through SGs, and two hot legs and four cold legs, all of which are included in the TRAC model.
Reactivity feedback from fuel and moderator temperature is included in the vessel model.
The TRAC noding diagram for this model is shown i n Fi g. D-1.
Information for this model was obtained f rom the Final Safety Analysis Reports.
The two cold legs on the "B" side are modeled as one combined cold leg for computational efficiency.
Also modeled on the primary side are the main coolant pumps, loop seals. surge line, pressurizer, emergency core-cooling injection
[ including high pressure injection and accumulators), hot-leg candy canes, and upper plenum vent salves.
The secondary side of the model includes the steam lines, main feedwater, AFW v.ith SG-level control.
The atmospheric relief
- valves, TSVs and turbine-bypass valves are modeled by a
mass-flow-versus-pressure boundary condition.
The primary-system relief train was modeled as a TEE component that was connec ted to the PORV and SVs.
The code version used for the calculation was 8.3 + 18 updates and it is stored on the Los Alamos National Laboratory common file system (CFS) with the descriptor /085785/NRRFW/GOP.
The updates are stored on /095785/0CA45/8.30P/ UPDATES.
The files for the input models used in this study are as follows:
Steady-state input
/085785/(X'A45/ INSS Base transient, first restart
/085785/NNRFW/INPFWNE 89
th
@j e
- i s -i. i., i.9 %
gl-~H H-l~
I I
RC" T *l9l-l-H -lal-H ~l-
@ s - i - i~;el:leH-H-H-H~H B"" t-
@_ a@ie@
on m g
Esq~ - ps
,,~
1-i~i-o L
@am@ac:
~1-i-ig g gai-i-g s--
e 90s
=os
-e s
w,%, A; G,
g g
g ai - i - i -
a a
e e e
@g u
1 u
g N
e
~
,a, m
a
,a, e
. ~.
c n
3 s
,i s 's
~
v g
7 ai-i~g-q$1 V
j jg
~,
g g
}
u--
eE a os g
s.,,,,, 5 [
r.i_i 6
~
,. i-lel:leH-H-H*laH-laE}-
~
=c a -19 H - H - H la H -
1 I
4-~H-lal-l~
~
.j i.i.
Gle 2
I
@a 90 e
Subsequent restarts
/085785/0CA45/INNE
" Feed and Bleed" transient
/085785/NRRFW/INPFWE II.
TRAC-PF1 OCONEE MODEL FOR S 11 SEQUENCE TRANSIENTS 3
A schematic of the TRAC input model used in this analysis is shown in Fig. D-2.
Important features of this model are as follows:
1.
Three-dimensional vessel with 10 axial levels, 6 circumferential nodes, and 2 radial rings; 2.
Split cold legs are modeled, with break locatior. as shown; and 3.
Fine noding of downcomer-boiler boundary and addition of downcomer orifice plate in SGs reduce computational instabilities.
The TRAC version used in this study was TRAC-PF1 version 9.7 plus the following additional updates:
RGS39 RGS40 RGS41 ENTLIM1 IIGAMFX FIXIM ER FIXMEM3 STRPD16 All updates are described in an internal Los Alamos National Laboratory memo, 4
J. R. Netuschil, "PF1 VERSION 9.8," May 4, 1983.
Files necessary for recreating the TRAC runs in this report are stored on the following Los Alamos CFS nodes.
ASSEMBLED CODE:
/093978/ TRAC /CRAY/GO+
STEADY STATE INPUT:
/093978/ TRAC /CRAY/SSOC1 91
g a
b
-......vA y
[i -FH-H-FE
- f. 4 i e
i i
-,a
,:l.l.H-FE gg g
G s -. 'cleblelel-bH i-HI-I:( P---- I-ancs:=(
.)=s i-i y
....-,s g
g 8
~
-v)g GE 20 1
i@
i - i - i - #
g g
g 4
A
@ - :0
==
n
,N,
/
O-
.L 1
s e
s
=
5
- =
70 os a
g 6-
~
~
c ;;
z I
~
s g
a
~
I@
i. i i
- f y
2 2
eE O
@ E@
.4 0E ab!Of W
b h
.f s.
1 - i -
C A
~ '
~'
6
~
n
,.i.lepjepl:lel.H-HI-I-6.....r_
." t C U ms
\\t-FH H-Fd
.., -,.g e @@
I 3
92
TRANSIENT INPUT:
102 mm: /093978/ TRAC /CRAY/0CSB4in/TRANS.
TRANS2.TRANS3.TRANS4,TRANS5 31 mm: /093978/ TRAC /CRAY/0CSB8/TRANSI.
TRANS2.TRANS3.TRANS4,TRANS5 22 mm: /093978/ TRAC /CRAY/0CSB4/TRANS1, TRANS2,TRANS3 11 mm: /093978/ TRAC /CRAY/0CSB1/TRANS, TRANS1,TRANS2,TRANS3 III.
TRAC-PF1 OCONEE MODEL FOR T KMU SEQUENCE TRANSIENTS 2
See Appendix D.I for a description of this model.
The TRAC noding diagram for this model is shown in Fig. D-3.
The code version used for the calculations was 11.1 + 4 updates and it is stored on the Los Alamos National Laboratory common file system with the desc riptor /085785/OCAWS/G0ll.
The updates are stored on /085785/OCATWS/ UPDATES.
The files for the input models used in this study are as follows:
Steady-state input
/085785/0 CAWS /INSS Base transient, first restart
/085785/NRRFW/INRSO Subsequent restarts
/085785/0CA45/RS1 IV.
TRAC-PF1 OCONEE MODEL FOR V SEQUENCE TRANSIENTS A system schematic of the model for Oconee PWR that was used in the interfacing systems LOCA cciculations is shown in Fig. D-4.
The low-pressure piping break is modeled on the RilRS line with a valve component.
A two-theta vessel model is used because the main point of interest was event timing and not the detailed in-core physics which is limited by lack of degraded core models in TRAC.
The core uncovery and melt sequences should be modeled with a more appropriate code such as MIMAS.
The cold legs are combined for computational efficiency.
Both SGs, pressurizer, two RCPs. both CFTs (accumulators).
the reactor vessel, and 93
@Qe e
a e i si 5
- e s -,. i...V,/e A-H,H.a S
l l
- ei.i.H.s
@gQg @
0 ; - ' - 'clelslaleleleH=l-HIH eer
-::Q. - D::a
~
i-i~i-a: ~
s -
6
@" g @ ~
]
-i.i-,
@[
- g 0i
/A Di e e e
t e
fD b.fffY 4 $
/
e VVM N.d l
$9}
i
~
/\\
=
e n
e o
Y
!e E
g 65 S OE 5
sei.i-ib h
x F.
O
~
- i - i-tl:lelel:lel.H-HiH "EI-
- i ~ - el.l.H-Hl s
1
=
E.' H 'l*I a
- i. i...y i m.
@ g e
e y
@a 94
s: = 2
- s j l
@l E E
-,.....vA,.i.a s.H.
e i
i
,H-H-4 cs -
O -.. :=;e; ;=;=;=;si.:. -s-1,y f
g g g rgq-l - p<s 8
1-1-i-l L
a o
9
@, f4
.i..1e g
$ r
~
~ 3 - ' ~ ' f jj
% y_
g lV l
(@
)*
E c
n 5?fgzg:-
h
8
.t
.T 1
a;
.5 8 It k. "%.~f'S/
"3
- e M
O W Y Y Y_
-_,._. '!?'A. _. __@
l
.,s
/
\\
8 8
s m
e 5 e t
6 S OE W
s....,e n
cn
=
.i-,
i 6
c:af.
ga s m=lc; l=i=!:!si.H-4-j"El O
~'
k gy Ct'
- hah *4 g
3;, _,..g q,,.'g.i.i.g u
1.,-
_.......y el as E e
ES I l 95
associated piping are modeled.
The plant model consists of 234 computational cells, of which 40 are in the 3-dimensional vessel.
The code version used for the calculations was 11.2 and is stored on the Los Alamos National Laboratory CFS under
/SASA/OCONEE/V-SEQ / CODE /GO root directory.
The files for the input models are stored as:
Steady-State Input
/SASA/0CONEE/V-SEQ /SS/TRACIN Base-case transient
/SASA 000 NEE /V-SEQ / BASE /RST1 to RST3/TRACIN Normal-plant-response transient
/SASA/000 NEE /V-SEQ /AUT01/RST1 to RST6/TRACIN Operator intervention transient (No LPI)
/SASA/000 NEE /V-SEQ /AUT02/RST1 to RST6/TRACIN Operator intervention transient (No LPI, 1 HPI)
/SASA/000 NEE /V-SEQ /MINSI/RST1 to RST5/TRACIN P
96
DISTRIBUTION' Nuclear Regulatory Commission, R4, Bethesda, Maryland 298 Technical Information Center, Oak Ridge, Tennessee 2
Los Alamos National Laboratory 33 i
333
- u.s.oovsa usarm mmeomesam-warm 97
sunC PO 88 338 U $ NUCLi%8 6 I t'ULAf oav Co48 lS56oN 16 troxT Nuweed faw,W 87 FsOC ### ved No, af ear, 4
I 12 Sei I
- %"3j3 '
BIBLIOGRAPHIC DATA SHEET LA-10351-MS NUREG/CR-4140
.. =stroCrioN o= T i nive se 3 Taite &No SueTITL 3 LeavEOLANE Dominant Accident Sequences in Oconee-1 j
Press rized Water Reactor
,,,,,,,,,,;co,,,,,,,
fl
.oN T,.
e aviaoaisi January #
1985 J.F. Dea ng, R.J. Henninger, B. Nassersharif A 8^ at'oa' '55'to l
(Compiled B.E. Boyack and B. Nassersharif)
April /
1985 7..oa *No o a.Niz.7,oN N*e o.iLiNo.ooamis,,
i.C
.,aoaC
.sa.oa whiiNuu ea Los Alamos Nati l Laboratory
[7228 Los Alamos, NM 8
45
/
10 StoN50AiNu omGANi2ATsom NAWE ANo MAtLING Ao SS (f sele Cores Ila TYPE oF REPoMT Division of Accident Ev luation Informal Office of Nuclear Regula ry Research
..noocove so,,
...e.,0, U.S. Nuclear Regulatory C ission
/
Washington, D.C.
20555 y
42 SUP/LeM N T AR T NCT T,8 g
1
/
,,,...T.C,_.,,
j A set of dominant accident sequences in the Oconee-1 pressurized water reactor was selected using probabi'li's Because some accident scenarios pere, tic risk analysis methods.
cccident sequences was selected #to be a(milar, a subset of four d
nalyzed with the Transient Reactor Analysis Code (TRAC) to# further hpr insights into similar types of accidents.
The sequences s' elected werdgloss-of-feedwater, small-l gmall break loss-of-coolant,f oss-of-feedwater-initiated transient withoutscram,andinterfacingsystemsloss-Q-coolantaccidents.
The normal plant response ayid the impact of eggipment availability and potential operator actions were also examined.
Strategies were developed for operator acpions not covered in existing emergency operator guidelines and w$re tested using TRAC sitimlations to evaluate their effectiveness in preventing core uncovery andAmaintaining core cooling.
f v
i k
Tv vagas ts re0ert, Unclassified.
o ~uu...o...oes to PR*C
f A
Q' w
sg 7r sW.n
- 4 C
4_-M4',
Y
'd
/
- D1"
'I J'
A E
I 0
7 1
L
.s t
+
I
.I J
4
+
w 4
b.
1
,.-%.y 2
.(
1 i
.(
f 3
' 8 i
iy'.
t..
e a
'l k.
t i
1 s
5-S V
t
,~$.'.
5 k.
4 T Y.',
4 Y
4 b
b r
P a
g 7
t P
..g
.)
,ed Je s ' w. a - -._
v; 6'eh g
- pmO--"
A t --$. 9 9
4
.'9.) $-
d
[
.L m_%
f.-
b.
._p n.
_^s w
sm,w4.ss 4'
+.s J.
a
.W--
g 5
.C 4
,IP i
~
ft-t e"
a(J\\"
e
~
] b( L{) e-
._j.
'.{
L
. 4 0
i n-4 1
- 6' -
m,l\\ '-, >
p
,l V
Ag h"
g l
a
' l, '_
J
~,, - -..-
-. a j
1 %
g r
' i r,
. %, t. l '
1
., O'. ',,
II I-'.la
,s h j I
','l i
, 1,
- ',g s-^
f i
1
,1
---.1..-
~, _. -
.,.,,,.._o.-.m._... -._..
.e
,.,--,,..,m--
- -. - -,..... -. -. - _, _ -