basic level 1. psa course for analysts

Basic Level 1. PSA course for analystsBasic Level 1. PSA course for analysts

PSA quantificationPSA quantification

IAEA Training in level 1 PSA and PSA applications

PSA quantification

Slide 2.

ContentContent

Impacts of truncation on the numerical resultsImportance analysis

Roles and definitionsUncertainty analysis

Elements of the uncertainties

Impacts of truncation on the numerical resultsImpacts of truncation on the numerical resultsImportance analysisImportance analysis

Roles and definitionsRoles and definitionsUncertainty analysisUncertainty analysis

Elements of the uncertaintiesElements of the uncertainties

PSA quantification

Slide 3.

NUMERICAL TRUNCATIONNUMERICAL TRUNCATIONIMPACTS FROM TRUNCATIONIMPACTS FROM TRUNCATION

OPTIMISTIC RESULTS

DELETED CUTSETS

INCONSISTENT CONTRIBUTORS

INCORRECT IMPORTANCE MEASURES

LEVEL OF DETAIL FOR APPLICATIONS

LEVEL 1 / LEVEL 2 INTERFACE

OPTIMISTIC RESULTSOPTIMISTIC RESULTS

DELETED CUTSETSDELETED CUTSETS

INCONSISTENT CONTRIBUTORSINCONSISTENT CONTRIBUTORS

INCORRECT IMPORTANCE MEASURESINCORRECT IMPORTANCE MEASURES

LEVEL OF DETAIL FOR APPLICATIONSLEVEL OF DETAIL FOR APPLICATIONS

LEVEL 1 / LEVEL 2 INTERFACELEVEL 1 / LEVEL 2 INTERFACE

PSA quantification

Slide 4.

NUMERICAL TRUNCATIONNUMERICAL TRUNCATIONCONTROLLING TRUNCATIONCONTROLLING TRUNCATION

DO NOT ACCEPT RECOMMENDED CUTOFF VALUES

ADJUST TRUNCATION LIMITS UNTIL RESIDUALS ARE LESS THAN ~ 10 % OF TOTAL

PER INITIATING EVENTTOTAL CORE DAMAGE

BETTER RESOLUTION MAY BE REQUIRED FOR SPECIFIC PLANT DAMAGE STATES OR LEVEL 2 ISSUES

BETTER RESOLUTION MAY BE REQUIRED FOR SPECIFIC APPLICATIONS / SENSITIVITY STUDIES

DO NOT ACCEPT RECOMMENDED CUTOFF VALUESDO NOT ACCEPT RECOMMENDED CUTOFF VALUES

ADJUST TRUNCATION LIMITS UNTIL RESIDUALS ARE ADJUST TRUNCATION LIMITS UNTIL RESIDUALS ARE LESS THAN ~ 10 % OF TOTALLESS THAN ~ 10 % OF TOTAL

PER INITIATING EVENTPER INITIATING EVENTTOTAL CORE DAMAGETOTAL CORE DAMAGE

BETTER RESOLUTION MAY BE REQUIRED FOR BETTER RESOLUTION MAY BE REQUIRED FOR SPECIFIC PLANT DAMAGE STATES OR LEVEL 2 ISSUESSPECIFIC PLANT DAMAGE STATES OR LEVEL 2 ISSUES

BETTER RESOLUTION MAY BE REQUIRED FOR BETTER RESOLUTION MAY BE REQUIRED FOR SPECIFIC APPLICATIONS / SENSITIVITY STUDIESSPECIFIC APPLICATIONS / SENSITIVITY STUDIES

PSA quantification

Slide 5.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESROLE OF RISK IMPORTANCE MEASURESROLE OF RISK IMPORTANCE MEASURES

PROVIDE SIMPLE ESTIMATES OF SENSITIVITY TO EXTREME VALUES FOR MODEL PARAMETERS

PROVIDE A SIMPLE AND COARSE RANKING OF ITEMS WITH RESPECT TO RISK OR SAFETY IMPORTANCE

CAN BE USED TO FOCUS REVIEWS AND SENSITIVITY STUDIES

SHOULD BE VIEWED AS A SUPPLEMENT TO, NOT A REPLACEMENT FOR, CAREFUL “TOP-DOWN” ANALYSIS OF THE RISK CONTRIBUTORS

KNOWLEDGE OF THE MODEL IS ESSENTIAL TO AVOID MISLEADING CONCLUSIONS

PROVIDE SIMPLE ESTIMATES OF SENSITIVITY TO PROVIDE SIMPLE ESTIMATES OF SENSITIVITY TO EXTREME VALUES FOR MODEL PARAMETERSEXTREME VALUES FOR MODEL PARAMETERS

PROVIDE A SIMPLE AND COARSE RANKING OF ITEMS PROVIDE A SIMPLE AND COARSE RANKING OF ITEMS WITH RESPECT TO RISK OR SAFETY IMPORTANCEWITH RESPECT TO RISK OR SAFETY IMPORTANCE

CAN BE USED TO FOCUS REVIEWS AND SENSITIVITY CAN BE USED TO FOCUS REVIEWS AND SENSITIVITY STUDIESSTUDIES

SHOULD BE VIEWED AS A SUPPLEMENT TO, NOT A SHOULD BE VIEWED AS A SUPPLEMENT TO, NOT A REPLACEMENT FOR, CAREFUL REPLACEMENT FOR, CAREFUL ““TOPTOP--DOWNDOWN”” ANALYSIS ANALYSIS OF THE RISK CONTRIBUTORSOF THE RISK CONTRIBUTORS

KNOWLEDGE OF THE MODEL IS ESSENTIAL TO AVOID KNOWLEDGE OF THE MODEL IS ESSENTIAL TO AVOID MISLEADING CONCLUSIONSMISLEADING CONCLUSIONS

PSA quantification

Slide 6.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESRISK INDICES AND MODEL ELEMENTSRISK INDICES AND MODEL ELEMENTS

RISK INDICESCORE DAMAGE FREQUENCYLARGE, EARLY RELEASE FREQUENCYEARLY FATALITY FREQUENCYANY OTHER RISK MEASURE WITHIN THE PSA MODEL SCOPE

MODEL ELEMENTSINITIATING EVENTSSYSTEMSHUMAN ACTIONSCOMPONENTS / INDIVIDUAL BASIC EVENTSGROUPS OF BASIC EVENTS

RISK INDICESRISK INDICESCORE DAMAGE FREQUENCYCORE DAMAGE FREQUENCYLARGE, EARLY RELEASE FREQUENCYLARGE, EARLY RELEASE FREQUENCYEARLY FATALITY FREQUENCYEARLY FATALITY FREQUENCYANY OTHER RISK MEASURE WITHIN THE PSA MODEL SCOPEANY OTHER RISK MEASURE WITHIN THE PSA MODEL SCOPE

MODEL ELEMENTSMODEL ELEMENTSINITIATING EVENTSINITIATING EVENTSSYSTEMSSYSTEMSHUMAN ACTIONSHUMAN ACTIONSCOMPONENTS / INDIVIDUAL BASIC EVENTSCOMPONENTS / INDIVIDUAL BASIC EVENTSGROUPS OF BASIC EVENTSGROUPS OF BASIC EVENTS

PSA quantification

Slide 7.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESMOST COMMON IMPORTANCE MEASURESMOST COMMON IMPORTANCE MEASURES

FRACTIONAL IMPORTANCE

FUSSELL - VESELY IMPORTANCE

BIRNBAUM IMPORTANCE

RISK REDUCTION WORTH

RISK ACHIEVEMENT WORTH

FRACTIONAL IMPORTANCEFRACTIONAL IMPORTANCE

FUSSELL FUSSELL -- VESELY IMPORTANCEVESELY IMPORTANCE

BIRNBAUM IMPORTANCEBIRNBAUM IMPORTANCE

RISK REDUCTION WORTHRISK REDUCTION WORTH

RISK ACHIEVEMENT WORTHRISK ACHIEVEMENT WORTH

PSA quantification

Slide 8.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESVARIABLE DEFINITIONSVARIABLE DEFINITIONS

R(QX = N) Calculated value of Risk Index R with the value of Model Element X set equal to its nominal Mean Value N.

R(QX = 1) Calculated value of Risk Index R with the value of Model Element X set equal to 1.0.

R(QX = 0) Calculated value of Risk Index R with the value of Model Element X set equal to 0.

R(QR(QXX = N)= N) Calculated value of Risk Index R with the value Calculated value of Risk Index R with the value of Model Element X set equal to its nominal of Model Element X set equal to its nominal Mean Value N.Mean Value N.

R(QR(QXX = 1)= 1) Calculated value of Risk Index R with the value Calculated value of Risk Index R with the value of Model Element X set equal to 1.0.of Model Element X set equal to 1.0.

R(QR(QXX = 0)= 0) Calculated value of Risk Index R with the value Calculated value of Risk Index R with the value of Model Element X set equal to 0.of Model Element X set equal to 0.

PSA quantification

Slide 9.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESWARNINGWARNING

THE MATHEMATICAL FORMULAS FOR SOME NUMERICAL IMPORTANCE MEASURES ARE NOT DEFINED CONSISTENTLY IN THE LITERATURE

THE “BASIC PHILOSOPHY” IS CONSISTENT

DIFFERENCES PERTAIN PRIMARILY TO TREATMENT OF “SUCCESS STATES”

THE MATHEMATICAL FORMULAS FOR SOME THE MATHEMATICAL FORMULAS FOR SOME NUMERICAL IMPORTANCE MEASURES ARE NOT NUMERICAL IMPORTANCE MEASURES ARE NOT DEFINED CONSISTENTLY IN THE LITERATUREDEFINED CONSISTENTLY IN THE LITERATURE

THE THE ““BASIC PHILOSOPHYBASIC PHILOSOPHY”” IS CONSISTENTIS CONSISTENT

DIFFERENCES PERTAIN PRIMARILY TO TREATMENT OF DIFFERENCES PERTAIN PRIMARILY TO TREATMENT OF ““SUCCESS STATESSUCCESS STATES””

PSA quantification

Slide 10.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESFRACTIONAL IMPORTANCEFRACTIONAL IMPORTANCE

MEASURE OF THE FRACTION OF RISK INDEX R THAT IS CONTRIBUTED BY FAILURE OF MODEL ELEMENT X

GENERAL DEFINITION

FIX = SUM (All cutsets with X failed) / R(QX = N)

RISK SPECTRUM DEFINITION (SAME AS FUSSELL -VESELY IMPORTANCE)

FIX = [ R(QX = N) - R(QX = 0) ] / R(QX = N)

MEASURE OF THE FRACTION OF RISK INDEX R THAT IS MEASURE OF THE FRACTION OF RISK INDEX R THAT IS CONTRIBUTED BY FAILURE OF MODEL ELEMENT XCONTRIBUTED BY FAILURE OF MODEL ELEMENT X

GENERAL DEFINITIONGENERAL DEFINITION

FIFIXX = SUM (All = SUM (All cutsetscutsets with X failed) / R(Qwith X failed) / R(QXX = N)= N)

RISK SPECTRUM DEFINITION (SAME AS FUSSELL RISK SPECTRUM DEFINITION (SAME AS FUSSELL --VESELY IMPORTANCE)VESELY IMPORTANCE)

FIFIXX = [ R(Q= [ R(QXX = N) = N) -- R(QR(QXX = 0) ] / R(Q= 0) ] / R(QXX = N)= N)

PSA quantification

Slide 11.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESFUSSELL FUSSELL -- VESELY IMPORTANCEVESELY IMPORTANCE

MEASURE OF THE FRACTION OF RISK INDEX R THAT IS CONTRIBUTED BY FAILURE OF MODEL ELEMENT X

GENERAL DEFINITION

FVX = [ R(QX = N) - R(QX = 0) ] / R(QX = N)

RISK SPECTRUM DEFINITION

FVX = [ R(QX = N) - R(QX = 0) ] / R(QX = N)

MEASURE OF THE FRACTION OF RISK INDEX R THAT IS MEASURE OF THE FRACTION OF RISK INDEX R THAT IS CONTRIBUTED BY FAILURE OF MODEL ELEMENT XCONTRIBUTED BY FAILURE OF MODEL ELEMENT X


FVFVXX = [ R(Q= [ R(QXX = N) = N) -- R(QR(QXX = 0) ] / R(Q= 0) ] / R(QXX = N)= N)

RISK SPECTRUM DEFINITIONRISK SPECTRUM DEFINITION

FVFVXX = [ R(Q= [ R(QXX = N) = N) -- R(QR(QXX = 0) ] / R(Q= 0) ] / R(QXX = N)= N)

PSA quantification

Slide 12.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESBIRNBAUM IMPORTANCEBIRNBAUM IMPORTANCE

MEASURE OF THE MAXIMUM POSSIBLE FRACTIONAL CONTRIBUTION TO RISK INDEX R FROM FAILURE OF MODEL ELEMENT X. (SOMETIMES CALLED THE PARTIAL RISK DERIVATIVE FOR ELEMENT X.)

GENERAL DEFINITION

BIX = [ R(QX = 1) - R(QX = 0) ] / R(QX = N)

RISK SPECTRUM DEFINITION - NOT CALCULATED

MEASURE OF THE MAXIMUM POSSIBLE FRACTIONAL MEASURE OF THE MAXIMUM POSSIBLE FRACTIONAL CONTRIBUTION TO RISK INDEX R FROM FAILURE OF CONTRIBUTION TO RISK INDEX R FROM FAILURE OF MODEL ELEMENT X. (SOMETIMES CALLED THE MODEL ELEMENT X. (SOMETIMES CALLED THE PARTIAL RISK DERIVATIVE FOR ELEMENT X.)PARTIAL RISK DERIVATIVE FOR ELEMENT X.)


BIBIXX = [ R(Q= [ R(QXX = 1) = 1) -- R(QR(QXX = 0) ] / R(Q= 0) ] / R(QXX = N)= N)

RISK SPECTRUM DEFINITION RISK SPECTRUM DEFINITION -- NOT CALCULATEDNOT CALCULATED

PSA quantification

Slide 13.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESRISK REDUCTION WORTHRISK REDUCTION WORTH

MEASURE OF THE AMOUNT BY WHICH RISK INDEX R MAY BE REDUCED IF MODEL ELEMENT X IS PERFECT

GENERAL DEFINITION (INVERTED IN SOME REFERENCES)

RRWX = R(QX = N) / R(QX = 0)


RRWX = R(QX = N) / R(QX = 0)

MEASURE OF THE AMOUNT BY WHICH RISK INDEX R MEASURE OF THE AMOUNT BY WHICH RISK INDEX R MAY BE REDUCED IF MODEL ELEMENT X IS PERFECTMAY BE REDUCED IF MODEL ELEMENT X IS PERFECT

GENERAL DEFINITION (INVERTED IN SOME GENERAL DEFINITION (INVERTED IN SOME REFERENCES)REFERENCES)

RRWRRWXX = R(Q= R(QXX = N) / R(Q= N) / R(QXX = 0)= 0)


RRWRRWXX = R(Q= R(QXX = N) / R(Q= N) / R(QXX = 0)= 0)

PSA quantification

Slide 14.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESRISK ACHIEVEMENT WORTHRISK ACHIEVEMENT WORTH

MEASURE OF THE AMOUNT BY WHICH RISK INDEX R MAY INCREASE IF MODEL ELEMENT X IS ALWAYS FAILED

GENERAL DEFINITION (INVERTED IN SOME REFERENCES)

RAWX = R(QX = 1) / R(QX =N)


RAWX = R(QX = 1) / R(QX =N)

MEASURE OF THE AMOUNT BY WHICH RISK INDEX R MEASURE OF THE AMOUNT BY WHICH RISK INDEX R MAY INCREASE IF MODEL ELEMENT X IS ALWAYS MAY INCREASE IF MODEL ELEMENT X IS ALWAYS FAILEDFAILED

GENERAL DEFINITION (INVERTED IN SOME GENERAL DEFINITION (INVERTED IN SOME REFERENCES)REFERENCES)

RAWRAWXX = R(Q= R(QXX = 1) / R(Q= 1) / R(QXX =N) =N)


RAWRAWXX = R(Q= R(QXX = 1) / R(Q= 1) / R(QXX =N)=N)

PSA quantification

Slide 15.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESUSE OF RISK IMPORTANCE MEASURESUSE OF RISK IMPORTANCE MEASURES

USED TO IDENTIFY COMMON CONTRIBUTORS THAT APPEAR IN MANY SEQUENCES AND CUTSETS

USED FOR RANKING PLANT FEATURES BY RISK SIGNIFICANCE (E.G., FOR FOCUSED TESTING OR MAINTENANCE)

RISK ACHIEVEMENT WORTH IS USEFUL FOR ESTIMATING THE RISK SIGNIFICANCE OF EQUIPMENT THAT IS REMOVED FROM SERVICE

RISK REDUCTION WORTH IS USEFUL FOR BOUNDING THE RISK BENEFITS FROM PROPOSED IMPROVEMENTS

IMPACTS FROM SOME PRECURSOR EVENTS CAN BE EVALUATED BY EXAMINATION OF IMPORTANCE MEASURES

EXAMINATION OF GROUPS CAN PROVIDE INSIGHTS ABOUT COMPOUND IMPACTS AND DEPENDENCIES NOT EVIDENT FROM SINGLE-COMPONENT ANALYSES

USED TO IDENTIFY COMMON CONTRIBUTORS THAT APPEAR IN MANY USED TO IDENTIFY COMMON CONTRIBUTORS THAT APPEAR IN MANY SEQUENCES AND CUTSETSSEQUENCES AND CUTSETS

USED FOR RANKING PLANT FEATURES BY RISK SIGNIFICANCE (E.G., FOR USED FOR RANKING PLANT FEATURES BY RISK SIGNIFICANCE (E.G., FOR FOCUSED TESTING OR MAINTENANCE)FOCUSED TESTING OR MAINTENANCE)

RISK ACHIEVEMENT WORTH IS USEFUL FOR ESTIMATING THE RISK RISK ACHIEVEMENT WORTH IS USEFUL FOR ESTIMATING THE RISK SIGNIFICANCE OF EQUIPMENT THAT IS REMOVED FROM SERVICESIGNIFICANCE OF EQUIPMENT THAT IS REMOVED FROM SERVICE

RISK REDUCTION WORTH IS USEFUL FOR BOUNDING THE RISK BENEFITS RISK REDUCTION WORTH IS USEFUL FOR BOUNDING THE RISK BENEFITS FROM PROPOSED IMPROVEMENTSFROM PROPOSED IMPROVEMENTS

IMPACTS FROM SOME PRECURSOR EVENTS CAN BE EVALUATED BY IMPACTS FROM SOME PRECURSOR EVENTS CAN BE EVALUATED BY EXAMINATION OF IMPORTANCE MEASURESEXAMINATION OF IMPORTANCE MEASURES

EXAMINATION OF GROUPS CAN PROVIDE INSIGHTS ABOUT COMPOUND EXAMINATION OF GROUPS CAN PROVIDE INSIGHTS ABOUT COMPOUND IMPACTS AND DEPENDENCIES NOT EVIDENT FROM SINGLEIMPACTS AND DEPENDENCIES NOT EVIDENT FROM SINGLE--COMPONENT COMPONENT ANALYSESANALYSES

PSA quantification

Slide 16.

NUMERICAL IMPORTANCE MEASURESNUMERICAL IMPORTANCE MEASURESCOMMON PROBLEMS IN USE OF RISK IMPORTANCECOMMON PROBLEMS IN USE OF RISK IMPORTANCE

IMPACTS FROM NUMERICAL TRUNCATION

ARTIFICIAL ASYMMETRIES IN MODELS FOR NORMALLY RUNNING EQUIPMENT

LIMITATIONS IN SIMULATING EQUIPMENT OUT OF SERVICE (GUARANTEED FAILED)

LIMITATIONS IN SIMULATING “SUCCESS STATES”

NO CONSIDERATION OF UNCERTAINTIES; DETAILED NUMBERS IMPLY THAT ALL VALUES ARE KNOWN PRECISELY

FOCUS TOO MUCH ATTENTION ON NUMERICAL COMPARISONS OF FINE STRUCTURE, RATHER THAN “BIG PICTURE” UNDERSTANDING OF RISK CONTRIBUTORS

IMPACTS FROM NUMERICAL TRUNCATIONIMPACTS FROM NUMERICAL TRUNCATION

ARTIFICIAL ASYMMETRIES IN MODELS FOR NORMALLY RUNNING ARTIFICIAL ASYMMETRIES IN MODELS FOR NORMALLY RUNNING EQUIPMENTEQUIPMENT

LIMITATIONS IN SIMULATING EQUIPMENT OUT OF SERVICE LIMITATIONS IN SIMULATING EQUIPMENT OUT OF SERVICE (GUARANTEED FAILED)(GUARANTEED FAILED)

LIMITATIONS IN SIMULATING LIMITATIONS IN SIMULATING ““SUCCESS STATESSUCCESS STATES””

NO CONSIDERATION OF UNCERTAINTIES; DETAILED NUMBERS IMPLY NO CONSIDERATION OF UNCERTAINTIES; DETAILED NUMBERS IMPLY THAT ALL VALUES ARE KNOWN PRECISELYTHAT ALL VALUES ARE KNOWN PRECISELY

FOCUS TOO MUCH ATTENTION ON NUMERICAL COMPARISONS OF FINE FOCUS TOO MUCH ATTENTION ON NUMERICAL COMPARISONS OF FINE STRUCTURE, RATHER THAN STRUCTURE, RATHER THAN ““BIG PICTUREBIG PICTURE”” UNDERSTANDING OF RISK UNDERSTANDING OF RISK CONTRIBUTORSCONTRIBUTORS

PSA quantification

Slide 17.

UNCERTAINTYUNCERTAINTYWHY DO WE CARE ABOUT UNCERTAINTY?WHY DO WE CARE ABOUT UNCERTAINTY?

HAZARD VS. RISK

BOUNDING VS. REALISTIC

DECISION MAKING

ROLE OF COMMUNICATION IN RISK ASSESSMENT

HAZARD VS. RISKHAZARD VS. RISK

BOUNDING VS. REALISTICBOUNDING VS. REALISTIC

DECISION MAKINGDECISION MAKING

ROLE OF COMMUNICATION IN RISK ASSESSMENTROLE OF COMMUNICATION IN RISK ASSESSMENT

PSA quantification

Slide 18.

UNCERTAINTYUNCERTAINTYSOURCES OF UNCERTAINTYSOURCES OF UNCERTAINTY

DATA

MODEL

APPLICATION OF THE MODEL

DATADATA

MODELMODEL

APPLICATION OF THE MODELAPPLICATION OF THE MODEL

PSA quantification

Slide 19.

UNCERTAINTY UNCERTAINTY DATA UNCERTAINTYDATA UNCERTAINTY

INTERPRETATION AND CLASSIFICATION OF FAILURE EVENTS

DETERMINATION OF SUCCESS DATA (NUMBER OF DEMANDS, OPERATING HOURS, EXPOSURE TIME, ETC.)

SIZE OF DATA SAMPLE (STATISTICAL UNCERTAINTY)

APPLICABLE POPULATION

MATHEMATICAL MODELS FOR DATA ANALYSIS

INTERPRETATION AND CLASSIFICATION OF FAILURE INTERPRETATION AND CLASSIFICATION OF FAILURE EVENTSEVENTS

DETERMINATION OF SUCCESS DATA (NUMBER OF DETERMINATION OF SUCCESS DATA (NUMBER OF DEMANDS, OPERATING HOURS, EXPOSURE TIME, ETC.)DEMANDS, OPERATING HOURS, EXPOSURE TIME, ETC.)

SIZE OF DATA SAMPLE (STATISTICAL UNCERTAINTY)SIZE OF DATA SAMPLE (STATISTICAL UNCERTAINTY)

APPLICABLE POPULATIONAPPLICABLE POPULATION

MATHEMATICAL MODELS FOR DATA ANALYSISMATHEMATICAL MODELS FOR DATA ANALYSIS

PSA quantification

Slide 20.

UNCERTAINTYUNCERTAINTYCORRELATED UNCERTAINTIESCORRELATED UNCERTAINTIES

COMMON DATA FOR SEVERAL COMPONENTS / FAILURE MODES

INDEPENDENT SAMPLING REDUCES OVERALL UNCERTAINTY

ACCOUNT FOR CORRELATION TO MAINTAIN CORRECT UNCERTAINTY

MEAN VALUE OF (A)2 =/= (MEAN VALUE OF A)2

COMMON DATA FOR SEVERAL COMPONENTS / COMMON DATA FOR SEVERAL COMPONENTS / FAILURE MODESFAILURE MODES

INDEPENDENT SAMPLING REDUCES OVERALL INDEPENDENT SAMPLING REDUCES OVERALL UNCERTAINTYUNCERTAINTY

ACCOUNT FOR CORRELATION TO MAINTAIN CORRECT ACCOUNT FOR CORRELATION TO MAINTAIN CORRECT UNCERTAINTYUNCERTAINTY

MEAN VALUE OF (A)MEAN VALUE OF (A)22 =/= (MEAN VALUE OF A)=/= (MEAN VALUE OF A)22

PSA quantification

Slide 21.

UNCERTAINTYUNCERTAINTYMODEL UNCERTAINTYMODEL UNCERTAINTY

SCOPE

COMPLETENESS

SUPPORTING DOCUMENTATION AND ANALYSES

SUCCESS CRITERIA

ASSUMPTIONS

ERRORS

SCOPESCOPE

COMPLETENESSCOMPLETENESS

SUPPORTING DOCUMENTATION AND ANALYSESSUPPORTING DOCUMENTATION AND ANALYSES

SUCCESS CRITERIASUCCESS CRITERIA

ASSUMPTIONSASSUMPTIONS

ERRORSERRORS

PSA quantification

Slide 22.

UNCERTAINTYUNCERTAINTYMODEL SCOPEMODEL SCOPE

LEVEL 1, LEVEL 2, LEVEL 3

FULL-POWER OPERATION, LOW POWER, SHUTDOWN

INTERNAL EVENTS, EXTERNAL EVENTS

LEVEL 1, LEVEL 2, LEVEL 3LEVEL 1, LEVEL 2, LEVEL 3

FULLFULL--POWER OPERATION, LOW POWER, SHUTDOWNPOWER OPERATION, LOW POWER, SHUTDOWN

INTERNAL EVENTS, EXTERNAL EVENTSINTERNAL EVENTS, EXTERNAL EVENTS

PSA quantification

Slide 23.

UNCERTAINTYUNCERTAINTYMODEL SCOPE MODEL SCOPE (Cont.)(Cont.)

WE CAN MAKE VERY LIMITED STATEMENTS ABOUT OUR STATE OF KNOWLEDGE WITH RESPECT TO PUBLIC HEALTH RISK IF WE HAVE PERFORMED ONLY A LEVEL 1 PSA THAT ESTIMATES THE FREQUENCY OF CORE DAMAGE DUE TO INTERNAL INITIATING EVENTS DURING FULL-POWER OPERATION.

WE CAN MAKE VERY LIMITED STATEMENTS ABOUT WE CAN MAKE VERY LIMITED STATEMENTS ABOUT OUR STATE OF KNOWLEDGE WITH RESPECT TO OUR STATE OF KNOWLEDGE WITH RESPECT TO PUBLIC HEALTH RISK IF WE HAVE PERFORMED ONLY A PUBLIC HEALTH RISK IF WE HAVE PERFORMED ONLY A LEVEL 1 PSA THAT ESTIMATES THE FREQUENCY OF LEVEL 1 PSA THAT ESTIMATES THE FREQUENCY OF CORE DAMAGE DUE TO INTERNAL INITIATING EVENTS CORE DAMAGE DUE TO INTERNAL INITIATING EVENTS DURING FULLDURING FULL--POWER OPERATION.POWER OPERATION.

PSA quantification

Slide 24.

UNCERTAINTYUNCERTAINTYMODEL COMPLETENESSMODEL COMPLETENESS

INITIATING EVENTS

PHENOMENA

DEPENDENCIES

HUMAN PERFORMANCE

INITIATING EVENTSINITIATING EVENTS

PHENOMENAPHENOMENA

DEPENDENCIESDEPENDENCIES

HUMAN PERFORMANCEHUMAN PERFORMANCE

PSA quantification

Slide 25.

UNCERTAINTYUNCERTAINTYINITIATING EVENTSINITIATING EVENTS

SUPPORT SYSTEMS

“INSIGNIFICANT” INITIATORS

RECOMMENDED TREATMENTESTIMATE FREQUENCYDETERMINE FUNCTIONAL IMPACTSALLOW PSA MODEL TO QUANTIFY ITS SIGNIFICANCE

SUPPORT SYSTEMSSUPPORT SYSTEMS

““INSIGNIFICANTINSIGNIFICANT”” INITIATORSINITIATORS

RECOMMENDED TREATMENTRECOMMENDED TREATMENTESTIMATE FREQUENCYESTIMATE FREQUENCYDETERMINE FUNCTIONAL IMPACTSDETERMINE FUNCTIONAL IMPACTSALLOW PSA MODEL TO QUANTIFY ITS SIGNIFICANCEALLOW PSA MODEL TO QUANTIFY ITS SIGNIFICANCE

PSA quantification

Slide 26.

UNCERTAINTYUNCERTAINTYPHENOMENAPHENOMENA

FAILURE TO SCRAM (ATWS)

OVERCOOLING (PTS)

TRANSIENT-INDUCED EVENTS (LOCA, LOSP, ETC.)

CONTAINMENT PHENOMENA (HYDROGEN, STEAM, AEROSOLS, ETC.)

RECOMMENDED TREATMENTINCLUDE IN MODEL, IF POSSIBLEBOUND FUNCTIONAL IMPACTS IF NOT MODELED EXPLICITLYSENSITIVITY STUDIES CAN ESTIMATE NUMERICAL CONSERVATISM

FAILURE TO SCRAM (ATWS)FAILURE TO SCRAM (ATWS)

OVERCOOLING (PTS)OVERCOOLING (PTS)

TRANSIENTTRANSIENT--INDUCED EVENTS (LOCA, LOSP, ETC.)INDUCED EVENTS (LOCA, LOSP, ETC.)

CONTAINMENT PHENOMENA (HYDROGEN, STEAM, AEROSOLS, CONTAINMENT PHENOMENA (HYDROGEN, STEAM, AEROSOLS, ETC.)ETC.)

RECOMMENDED TREATMENTRECOMMENDED TREATMENTINCLUDE IN MODEL, IF POSSIBLEINCLUDE IN MODEL, IF POSSIBLEBOUND FUNCTIONAL IMPACTS IF NOT MODELED EXPLICITLYBOUND FUNCTIONAL IMPACTS IF NOT MODELED EXPLICITLYSENSITIVITY STUDIES CAN ESTIMATE NUMERICAL SENSITIVITY STUDIES CAN ESTIMATE NUMERICAL CONSERVATISMCONSERVATISM

PSA quantification

Slide 27.

UNCERTAINTYUNCERTAINTYDEPENDENCIESDEPENDENCIES

PHYSICAL

FUNCTIONAL

LOCATION / ENVIRONMENTAL

DATA-BASED

HUMAN

PHYSICALPHYSICAL

FUNCTIONALFUNCTIONAL

LOCATION / ENVIRONMENTALLOCATION / ENVIRONMENTAL

DATADATA--BASEDBASED

HUMANHUMAN

PSA quantification

Slide 28.

UNCERTAINTYUNCERTAINTYSUPPORTING DOCUMENTATION AND ANALYSESSUPPORTING DOCUMENTATION AND ANALYSES

PLANT DESIGN INFORMATION

OPERATING, TESTING, MAINTENANCE PROCEDURES

DESIGN-BASIS ACCIDENT ANALYSES

BEST-ESTIMATE THERMAL / HYDRAULIC ANALYSES

RECOMMENDED TREATMENT“FIRST PRINCIPLES” CALCULATIONSDOCUMENT AND QUANTIFY UNCERTAINTYBOUND FUNCTIONAL IMPACTS IF NOT MODELED EXPLICITLYSENSITIVITY STUDIES CAN ESTIMATE NUMERICAL CONSERVATISM

PLANT DESIGN INFORMATIONPLANT DESIGN INFORMATION

OPERATING, TESTING, MAINTENANCE PROCEDURESOPERATING, TESTING, MAINTENANCE PROCEDURES

DESIGNDESIGN--BASIS ACCIDENT ANALYSESBASIS ACCIDENT ANALYSES

BESTBEST--ESTIMATE THERMAL / HYDRAULIC ANALYSESESTIMATE THERMAL / HYDRAULIC ANALYSES

RECOMMENDED TREATMENTRECOMMENDED TREATMENT““FIRST PRINCIPLESFIRST PRINCIPLES”” CALCULATIONSCALCULATIONSDOCUMENT AND QUANTIFY UNCERTAINTYDOCUMENT AND QUANTIFY UNCERTAINTYBOUND FUNCTIONAL IMPACTS IF NOT MODELED EXPLICITLYBOUND FUNCTIONAL IMPACTS IF NOT MODELED EXPLICITLYSENSITIVITY STUDIES CAN ESTIMATE NUMERICAL SENSITIVITY STUDIES CAN ESTIMATE NUMERICAL CONSERVATISMCONSERVATISM

PSA quantification

Slide 29.

UNCERTAINTYUNCERTAINTYSUCCESS CRITERIASUCCESS CRITERIA

SYSTEMS

OPERATOR ACTIONS

RECOMMENDED TREATMENTBEST-ESTIMATE ANALYSESBOUNDING ASSUMPTIONSSENSITIVITY STUDIES CAN ESTIMATE NUMERICAL CONSERVATISM

SYSTEMSSYSTEMS

OPERATOR ACTIONSOPERATOR ACTIONS

RECOMMENDED TREATMENTRECOMMENDED TREATMENTBESTBEST--ESTIMATE ANALYSESESTIMATE ANALYSESBOUNDING ASSUMPTIONSBOUNDING ASSUMPTIONSSENSITIVITY STUDIES CAN ESTIMATE NUMERICAL SENSITIVITY STUDIES CAN ESTIMATE NUMERICAL CONSERVATISMCONSERVATISM

PSA quantification

Slide 30.

UNCERTAINTYUNCERTAINTYASSUMPTIONSASSUMPTIONS

DOCUMENT ALL ASSUMPTIONS

SENSITIVITY STUDIES CAN ESTIMATE NUMERICAL CONSERVATISM

DOCUMENT ALL ASSUMPTIONSDOCUMENT ALL ASSUMPTIONS

SENSITIVITY STUDIES CAN ESTIMATE NUMERICAL SENSITIVITY STUDIES CAN ESTIMATE NUMERICAL CONSERVATISMCONSERVATISM

PSA quantification

Slide 31.

UNCERTAINTYUNCERTAINTYERRORSERRORS

THERE ARE ERRORS IN YOUR STUDY

SOME ERRORS PRODUCE CONSERVATIVE RESULTS, AND SOME PRODUCE OPTIMISTIC RESULTS

TYPICAL REVIEWS TEND TO FIND ERRORS THAT PRODUCE CONSERVATIVE RESULTS

TYPICAL REVIEWS TEND TO MISS ERRORS THAT PRODUCE OPTIMISTIC RESULTS

REVIEWS SHOULD EXAMINE “WHAT IS NOT IMPORTANT” WITH EQUAL EMPHASIS AS “WHAT IS IMPORTANT”

DEVELOP CONFIDENCE AND UNDERSTANDING WHY SPECIFIC INITIATING EVENTS, SCENARIOS, AND CONDITIONS ARE NOT IMPORTANT

ANALYST KNOWLEDGE, SENSITIVITY STUDIES, AND IMPORTANCE ESTIMATES CAN HELP TO FOCUS EXAMINATION

THERE ARE ERRORS IN YOUR STUDYTHERE ARE ERRORS IN YOUR STUDY

SOME ERRORS PRODUCE CONSERVATIVE RESULTS, AND SOME SOME ERRORS PRODUCE CONSERVATIVE RESULTS, AND SOME PRODUCE OPTIMISTIC RESULTSPRODUCE OPTIMISTIC RESULTS

TYPICAL REVIEWS TEND TO FIND ERRORS THAT PRODUCE TYPICAL REVIEWS TEND TO FIND ERRORS THAT PRODUCE CONSERVATIVE RESULTSCONSERVATIVE RESULTS

TYPICAL REVIEWS TEND TO MISS ERRORS THAT PRODUCE OPTIMISTIC TYPICAL REVIEWS TEND TO MISS ERRORS THAT PRODUCE OPTIMISTIC RESULTSRESULTS

REVIEWS SHOULD EXAMINE REVIEWS SHOULD EXAMINE ““WHAT IS NOT IMPORTANTWHAT IS NOT IMPORTANT”” WITH EQUAL WITH EQUAL EMPHASIS AS EMPHASIS AS ““WHAT IS IMPORTANTWHAT IS IMPORTANT””

DEVELOP CONFIDENCE AND UNDERSTANDING DEVELOP CONFIDENCE AND UNDERSTANDING WHYWHY SPECIFIC INITIATING SPECIFIC INITIATING EVENTS, SCENARIOS, AND CONDITIONS ARE NOT IMPORTANTEVENTS, SCENARIOS, AND CONDITIONS ARE NOT IMPORTANT

ANALYST KNOWLEDGE, SENSITIVITY STUDIES, AND IMPORTANCE ANALYST KNOWLEDGE, SENSITIVITY STUDIES, AND IMPORTANCE ESTIMATES CAN HELP TO FOCUS EXAMINATIONESTIMATES CAN HELP TO FOCUS EXAMINATION

PSA quantification

Slide 32.

UNCERTAINTYUNCERTAINTYSUMMARYSUMMARY

COMMUNICATION IS A CENTRAL ELEMENT OF RISK ANALYSIS AND DECISION MAKING

WE HAVE NOT COMPLETED OUR JOB AS RISK ANALYSTS UNTIL WE HAVE ADDRESSED UNCERTAINTY

BETTER A SIMPLIFIED QUANTITATIVE, OR EVEN QUALITATIVE, UNCERTAINTY ANALYSIS THAN NOTHING

COMMUNICATION IS A CENTRAL ELEMENT OF RISK COMMUNICATION IS A CENTRAL ELEMENT OF RISK ANALYSIS AND DECISION MAKINGANALYSIS AND DECISION MAKING

WE HAVE NOT COMPLETED OUR JOB AS RISK WE HAVE NOT COMPLETED OUR JOB AS RISK ANALYSTS UNTIL WE HAVE ADDRESSED UNCERTAINTYANALYSTS UNTIL WE HAVE ADDRESSED UNCERTAINTY

BETTER A SIMPLIFIED QUANTITATIVE, OR EVEN BETTER A SIMPLIFIED QUANTITATIVE, OR EVEN QUALITATIVE, UNCERTAINTY ANALYSIS THAN NOTHINGQUALITATIVE, UNCERTAINTY ANALYSIS THAN NOTHING

basic level 1. psa course for analysts

Documents