ANS PSA 2008 Topical Meeting - Challenges to PSA during the nuclear renaissance, Knoxville, Tennessee, September 6–11, 2008 Implementation of a Concept for a Risk- informed Diagnosis/Prognosis of Plant S tates through the RISARD System Kwang-Il Ahn [email protected]Integrated Safety Assessment Korea Atomic Energy Research Institute
22
Embed
ANS PSA 2008 Topical Meeting - Challenges to PSA during the nuclear renaissance, Knoxville, Tennessee, September 6–11, 2008 Implementation of a Concept.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
ANS PSA 2008 Topical Meeting - Challenges to PSA during the nuclear renaissance, Knoxville, Tennessee, September 6–11, 2008ANS PSA 2008 Topical Meeting - Challenges to PSA during the nuclear renaissance, Knoxville, Tennessee, September 6–11, 2008
Implementation of a Concept for a Risk-informed Diagnosis/Prognosis of Plant States through the RISARD System
Implementation of a Concept for a Risk-informed Diagnosis/Prognosis of Plant States through the RISARD System
Integrated Safety AssessmentKorea Atomic Energy Research Institute
KAERIKAERI 2
Motivation & Objectives
The Concept of RI-SAM
Computerized Tool SARD
Demonstrative Application
Concluding Remarks
⊙ Outline
KAERIKAERI 3
Key Ways for a Successful Implementation of SAM
Develop a proper SAM strategy by answering the questions:
How to reduce uncertainties in implementing the established SAM strategies? especially when available resources are limited.
Which essential safety function was lost at the time of the accident? That is, the root cause of the accident; Which safety systems are currently available for SAM?
What will be important future events? and what will be their evolution?
What are potential ‘success paths’ for SAM?
Utilize the computer-based methods & tools for supporting SAM:
capable of ① diagnosing the functional states of plant safety systems and ② predicting the future trends of key plant parameters as possible as quickly:
The diagnostic capability for plant states at the time of the accident is required to reduce the uncertainties in the current plant system state and to have a good basis for estimating future plant states.
Based on the current damage states of the plant, the prognostic capability for the possible evolution of the accident gives time enough to take an action for mitigating the consequence of the accident.
⊙ Motivation
KAERIKAERI 4
Our Approach for supporting SAM:
Utilize a PSA-based and SA phenomenological trends-based database (DB) (e.g., plant-, code-, accident sequence-specific SA analysis results) => SAR DB
Systematic use of SAM-related information
Quick & fast retrieval of the SAM-related information (Quick view)
Provide a computerized platform for a comprehensive use of SAR DB in a simple, fast and risk-informing way => RI-SARD
Proper information about the plant damage states at the time of the accident: the root cause of the accident (Diagnosis)
Insights on the possible evolution of the accident (critical parameters), based on the current damage states of the plant (Prognosis)
Develop the best strategy for supporting SAM (especially when available plant information is limited) => RI-SAM
Helpful in finding success paths for intended SAM actions
Helpful in providing appropriate actions to mitigate the accident
⊙ Objectives: RI-SAM Strategy & Tool
KAERIKAERI 5
⊙ RI-SAM: Diagnosis & Prognosis of Plant States
Monitor Plant Data & Signal: Identify the IE & CD statesDetermine the status/availability of systems needed to mitigate the IE
Signal Validation Process
ReSet Plant symptoms& accident time
DiagnosisSymptom-based
SARD modulePlant
symptoms
Plant damage states
Determine Plant symptoms
More Symptoms
?
Quick view Future trend of symptom parameters
Performance1. Key plant safety parameters for SAM2. Performance of
key SSCs
Decision for Implementation of the relevant SAM strategies
DetermineRelevant PDS (1)
AS screening (iteration loop)(2)
Auto switch to module for Prognosis of Future Plant Status
PrognosisScenario-based SARD module
Plant Conditions
Prioritize Frequency-based potential accident
sequences
ModificationLink to SA Simulatorfor Interactive Action
Success PathsCountermeas
ures
Note(1): A prescribed accident sequences by which uncertainty can be reduced in taking an action for SAM.Note(2): A process by which the prediction can be updated based upon successive data from plant.
(dynamic loop)(2)
Yes
No
KAERIKAERI 6
RISARD Key Functions & Modules
SARDB
• Code-specific or user-supplied SA accident sequence information:
characterized as system functionality & frequency (probabilistic information)
• Plant-, sequence-, code-specific predictions of key plant parameters:
• Additional diagnosis for accident initiators (on-going)
⊙ A Computerized Tool: RI-SARD
KAERIKAERI 7
Data Set Spec. for SARD
- Plant/Code/User ID- Accident Sequence Inform(1)
- Sensitivity Information (Plant systems & Model parameters)- Severe Accident Code Analysis Results(2) (Code Responses)- Summary of Key Accident Progression Events (Code Result)- Accident Mitigation Options- Data Set & Databank Index- Commentary Parts
Accident Sequence Types- PSA code-specific plant damage event trees for graphical use- User-specified events sequence
Database Update &Modification
Data Search & Retrieval, Graphical Display
- Scenario-based Plant Responses & Behavior- Plant symptom-based Potential Accident Sequences- Status of Plant System & Containment Systems
SARDB: Databank(MS Access DB)
SAR-informed Decision-making
Data Allocation intoSARDB
- Level 1/2 PSA - Accident Analysis- SAM Information- The other SAR Inform
SAR Information
Formatted SAR Data Sets SARD System Operation
(1) Severe Accident Initiators: LOCA (Large, Medium, Small), Loss of Off-site Power (LOOP), Station Blackout (SBO), Loss of Feed Water (LOFW), Interfacing System LOCA, Steam Generate Tube Rupture (SGTR), Anticipated Transient w/o Scram (ATWS), Loss of AC Bus (125V, 4.16KV), Large Secondary Side Break, General Transient
(2) Number of Categorized MAAP Response Parameters (Total 883): RCS/SG/ESF Information (134); Behavior of Core and Fuel (152); Lower Plenum Debris Behavior (77); Lower Head Failure Information (85); Containment Information (196); Source Term Information (229); Hydrogen Generation (10)
⊙ SARD: Data Sets Operation
KAERIKAERI 8Allocation of Plant-, Code-specific SA Analysis Results into SARDB
SARDB
Typical Form of SARDB
Plant-specific PDS ET
⊙ SARD: SARDB Generation (1)
MELCORHISPLT
KAERI-IPLOT
MELCOR Run
SARDB Generation Module
SARDB:MELCOR/MAAP DB
Plant-specific
Accident Scenarios
Plot Data
SARD:Plant statePrognosis/Diagnosis
Parameter listfor comparison
MAAP Run
MELCORHISPLT
KAERI-IPLOT
MELCOR Run
SARDB Generation Module
SARDB:MELCOR/MAAP DB
Plant-specific
Accident Scenarios
Plot Data
SARD:Plant statePrognosis/Diagnosis
Parameter listfor comparison
MAAP Run
Identify the initiating event & the status and availability of systems and equipment needed to avoid or mitigate the severe accident
PDS sequence: plant damage state + frequency
SA phenomenological trends with timeDeveloping trends of key events during accident
KAERIKAERI 9
Key Role of the PDS ET-based Diagnosis & Prognosis Provide the status of plant and cont. systems at the time of core damage
All potential ASs for an IE can be shown at a glance with its graphical form
Occurrence probability (or frequency) be systematically derived from PSA
The graphical form of PDS ET can be very useful in specifying a particular AS during the data loading and information retrieval process
Probability can be utilized as a criterion for screening the risk-significant ASs
Frequency
OPR1000 IEs
Number of ASs for each IE
A frequency criterion for AS screening
1.0E-11/ry 1.0E-10/ry 1.0E-9/ry
All (16) Several hundreds - 95
LOOP 120 12 (99% of total ASs) -
LBLOCA 30 7 (99% of total ASs) -Risk-informed SA Analysis
⊙ SARD: SARDB Generation (2)
KAERIKAERI 10
I.Es (/ry) Sequences Contribution (%) to I.E
Functional States of Safety Systems(success state: ‘/sss’, failed state: ‘sss’)
LLOCA(1.05x10-6)
LLOCA-2 17.01 /SIT*/LPI*/HPR*/HPH*CSS
LLOCA-3 19.25 /SIT*/LPI*/HPR*HPH*/CSS
LLOCA-5 11.50 /SIT*/LPI*HPR*/LPR/*HPH*/CSS
LLOCA-8 5.9 /SIT*/LPI*HPR*LPR*HPH*CSS
LLOCA-9 45.36 /SIT*LPI*/HPI*/HPR*/CSS
LLOCA-15 0.46 /SIT*LPI*HPI*HPR*LPR*HPH*CSS
LLOCA-17 0.15 SIT*/LPI*/HPR*/CSS
Sub total 99.63
MLOCA(6.34x10-7)
MLOCA-2 28.15 /HPI*/HPR*/HPH*CSR
MLOCA-3 31.87 /HPI*/HPR*HPH*/CSR
MLOCA-5 10.07 /HPI*HPR*/LPR*/CSR
MLOCA-8 9.0 /HPI*HPR*LPR*CSR
MLOCA-9 19.1 HPI*/LPI*/LPR*/CSI*/CSR
MLOCA-19 0.76 HPI*LPI*CSI
Sub total 98.95
SLOCA(1.92x10-6)
SLOCA-11 0.05 /HPI*/AFW*/ADV*HPR*LPR*/CSS
SLOCA-12 57.79 /HPI*/AFW*/ADV*HPR*LPR*CSS
SLOCA-13 0.27 /HPI*/AFW*/ADV*HPR*/LPR*/CSS
SLOCA-21 0.08 /HPI*/AFW*ADV*/MSSV*/HPR*BDL*/CSS
SLOCA-26 0.14 /HPI*/AFW*ADV*/MSSV*HPR*LPR*CSS
SLOCA-45 0.19 /HPI*AFW*/LPR*BDE*/CSS
SLOCA-55 1.19 HPI*/AFW*ADV*/MSSV*/HPR*BD*/CSS
SLOCA-57 5.12 HPI*/DPI*LPI*CSI
SLOCA-59 32.6 HPI*DPI*/LPI*/CSI
Sub total 97.43
Dominant accident initiators:Frequency-based screening of PDS sequences
LLOCA
LargeLOCA
SIT
SITsInjection
LPI
LPSISInjection
HPI
HPSISInjection
HPR
HPSISRecir-
culation
LPR
LPSISRecir.
HPH
HPSISHot
Cold LegRecir.
CSI
ContainmentInjection
Spray
CSR
Recir.CoolingusingCSS
RFSI
CavityFloodingSystemInjection
LLOCA
CDSQ5
CDSQ2
CDSQ3
CSR
CSR
CDSQ4
LPR
CSR
CSR
HPI
HPR
CSR
CSI
CSR
RFSI
CDSQ6
CSR
HPR
CSR
LPR
CSR
LPI
CSR
HPR
CSR
HPI
CSR
CSI
RFSI
SEQ#
STATE
STATE
FREQ
1 OK
2 43
3 27
4 28
5 27
6 28
7 29
8 30
9 27
10 28
11 29
12 30
13 29
14 30
15 31
16 32
17 27
18 28
19 27
20 28
21 29
22 30
23 27
24 28
25 29
26 30
27 29
28 30
29 31
30 32
LLOCA PDS ET
⊙ SARD: SARDB Generation (3)
KAERIKAERI 11
SAMPLE: Key Events Summary for LF115 (MAAP) Time Events Code Functional Status 0.000 157:T MAIN FW OFF 0.000 224:T MOTOR-DRIVEN AUX FEED WATER FORCED OFF 0.000 232:T CHARGING PUMPS FORCED OFF 17.836 31:T PZR SPRAYS ON 37.431 13:T REACTOR SCRAM 37.431 156:T MSIV CLOSED 42.578 153:T SEC SV(S) FIRST OPEN BROKEN S/G 42.578 163:T SEC SV(S) FIRST OPEN UNBROKEN S/G'S 867.556 161:T UNBKN S/G DRY 870.375 151:T BROKEN S/G DRY 1109.376 3:TH VALVE FIRST OPENED 1109.376 4:TH VALVE FIRST OPENED 1109.376 5:TH VALVE FIRST OPENED 1113.182 35:T VOID FRACTION IN PZR < 0.1 1728.464 4:T MAIN COOLANT PUMPS OFF 2584.444 691:T TRUE: CORE HAS UNCOVERED 4888.784 509:T TRUE: MAX. CORE TEMP EXCEEDS 2200. F 5084.019 690:T TRUE: MAXIMUM CORE TEMPERATURE HAS EXCEEDED 2499 K 5188.688 508:T TRUE: MAX. CORE EXIT TEMP EXCEEDS 1200. F 5962.269 2:T RELOCATION OF CORE MATERIALS TO LOWER HEAD STARTED 5987.935 103:T UPPER COMPT. SPRAYS ON 6778.318 3:T RV FAILED 6794.329 5:T HPI ON 6794.329 6:T LPI ON 6857.934 188:T ACCUMULATOR WATER DEPLETED 8142.179 1003:T TRUE: 1 TH COMPT BURNING IN PROGRESS 8142.179 1048:T TRUE: 4 TH COMPT BURNING IN PROGRESS 8142.179 1063:T TRUE: 5 TH COMPT BURNING IN PROGRESS 8142.492 1033:T TRUE: 3 TH COMPT BURNING IN PROGRESS 9897.961 5:F HPI OFF 9897.961 181:T RECIRC SYSTEM IN OPERATION ….
PDS sequence-
specific SA code
analysis
Parameters history:SA code parameter behavior with time
Events history: Plant system
status with time
⊙ SARD: SARDB Generation (5)
KAERIKAERI 12
Specification of the Target Scenario
AMP & Summary Data
Plant Data
SAR Data
ASQ data
SA Code data Code results
Specification of Sensitivity Information
Specification of Code Data (Multiple)
Specification of Code Data (Single)
Check of the Allocated Information
Specification of Databank Index
SARDBMS Access DB
⊙ SARD: SARDB Generation Module
KAERIKAERI 13
Display - Plant states - Base response - Sensitivity case - SAMG parameters - SSC performance - Events History
End of Searc
h
Scenario Base (1) Symptom Base (2)
PSA Information:IE & Target Sequence
Data Search: Plant-/Code-/AS
sequence-specific Responses
Set target Plant ID & Code ID
Plant Symptoms: - Code Parameters - Time windows
Prioritize Accident Scenarios (i = 1, n), in a risk-informing way
Target Sequence
More symptoms? AS Screening
Auto Switc
h
(1) Retrieval of the specified- accident sequence-based -based plant/code behavior (Accident Diagnosis)(2) Retrieval of plant symptoms-based -based potential accident sequences (Accident Prognosis)
⊙ SARD: Two-way information Retrieval
KAERIKAERI 14
⊙ SARD: Plant Symptom-based Diagnosis
Switch to the Scenario-based Module
User-specified plant symptoms
The most probable plant system stateList of potential
plant damage states
Progression of key events Future trend of
plant parameters
User-specified plant symptoms
The most probable plant system stateList of potential
plant damage states
Progression of key events Future trend of
plant parameters
Switch to the Scenario-based Module
PDS ET Events Functional Status
Set Plant & Code information
KAERIKAERI 15
⊙ SARD: PDS sequence-based Prognosis
User- specified accident conditions
Future history key events &plant parameters
User- specified code/plant parameters
User- specified accident conditions
Future history key events &plant parameters
User- specified code/plant parameters
Display of the Corresponding PDS ET PDS ET Events Functional StatusSet Plant & Code information
KAERIKAERI 16
⊙ Demo Application: Diagnosis of PDS sequences
1. Initial Plant Symptom (1) TWCR (temperature of water in core, K) [580-600] for Time Window (Sec.) [110-130]
Matched PDS Sequences Freq. (/ry) Functional States of Safety Systems
2. Two Additional Plant Symptoms (2) PPS (pressure in primary system, MPa) [12.50-12.51] for Time Window (Sec.) [110-130] (3) TGUP (temperature of gas in upper plenum, K) [600-620] for Time Window (Sec.) [110-130]
3. Two Additional Plant Symptoms (4) TWCR (temperature of water in core, K) [550-600] for Time Window (Sec.) [1950-2050] (5) TWCR (temperature of water in core, K) [435-445] for Time Window (Sec.) [2950-3050]
The corresponding PDS Sequences SBLOCA_S012, SBLOCA_S013, SBLOCA_S011
Making predictions about future trend of the 7 plant safety parameters to trigger the relevant SAMG and their
entry times, based on the user-specified thresholds
Entry time: 6.85 sec.
Entry time:10.4 sec.
Entry time: 30.32 sec.
LBLOCA-S03
KAERIKAERI 19
Water level in RPV
RPV LH Creep at 37355 sec.
Core uncover at 19876 sec.
No induced creep failure
RCS HL: unbroken RCS HL: broken
S/G: unbroken S/G: broken
PRV LH creep P-tube ejection
P-tube heatup Debris jet impingement
Making predictions about when core damage, core support plate failure, induced RCS & SG creep failure, reactor vessel
failure, and containment failure will occur
⊙ SAM-SSC Performance-failure time & probability
SBLOCA-S26
KAERIKAERI 20
Summary
Based on a concept of a RI-SAM, the present RI-SARD system explores a symptom-based diagnosis of potential PDS sequences in a risk-
informing way & a plant damage sequence-based prognosis of key plant parameter
behavior, in a simple, fast, and efficient way.
The replicated use of both processes makes it possible to extract information required for taking the intended SAM actions, consequently leading to an answer about what is the best strategy for SAM.
An example application through the OPR1000- and MAAP code-specific SAR DB has shown that the present approach can
enhance a diagnostic capability for anticipated plant states, give the SAM practitioners more time to take actions for mitigating the accident, reduce the still relatively large uncertainty in the field of SAM, and consequently, help guide the TSC staffs through a severe accident.
⊙ Concluding Remarks
KAERIKAERI 21
Future Plan for Improvement
Will involve the ability to link decisions made by RISARD with the SAM procedure and SA simulator, so that the impact of the SAM actions on an accident progression can be feedback to in an interactive way to a user.
Will involve the use of a more structured approach capable of ① diagnosing the current plant system state, ② predicting the most probable accident pathway during the progress of an accident, and ③ taking the best strategy to terminate its progression into an undesirable consequence, including a linking with
a diagnostic logic tree to diagnose effectively potential plant damage states, a simplified APET capable of predicting the progress of accidents accurately,
and a more sophisticated logical rule capable of extracting appropriate SAM
strategies for a given plant damage state
In addition, we will explore increasing the number of accident types recognized by RI-SARD (e.g., various spectrum of break sizes for LOCA & SGTR)
⊙ Concluding Remarks
Thank you for your attention !!!Thank you for your attention !!!