Risk-Informed Changes to the Licensing Basis - II

22.39 Elements of Reactor Design, Operations, and Safety Lecture 14

Fall 2006

George E. Apostolakis Massachusetts Institute of Technology

Department of Nuclea r S c ien ce and Engineering 1

Current Situation

Current requirements are independent of the frequency of break size; PRAs have shown that LBDEGB contribution to CDF is very small.

LBDEGB

Break

Department of Nuclea r S c ien ce and Engineering Size 2

Draft Rule Structure

• Existing § 50.46 essentially uncha nged

• Voluntary alternative rule added (§ 50.46a, now 50.46b)

• 50.46a defines alternative acceptance criteria

• 50.46b addresses LOCA redefinition only

• LOCA break spectrum divided into 2 regions by “transition” break size (TBS)

– based upon frequency and other considerations

• Breaks in smaller break region continue to be DBAs; must meet current § 50.46 analysis require ments and acceptance criteria

• Breaks larger than TBS become beyond design-basis accidents, but mitigation capability must be demonstrated up to full DEGB

– less stringent analysis assumptions/acceptance criteria

– demonstrate for all at-power operating configu r ations

Department of Nuclea r S c ien ce and Engineering 3

Proposal for Risk-Informing 50.46

LBDEGB

10 -5

TBS

Department of Nuclea r S c ien ce and Engineering

Break Size 4

Transition Break Size (50.46a)

• A break of area equal to the cross-sectional area of the inside diameter of specified piping of a specific reactor.

• PWRs

 Expert judgment: 4 to 7 inches.

 The largest piping attached to the reactor coolant system.

• BWRS

 Expert judgment: 6 to 14 inches.

 The larger of the feedwater l ine in side containment or the residual heat removal line inside containment.

Department of Nuclea r S c ien ce and Engineering 5

Breaks at or below the TBS (50.46a)

(i) Peak cladding temperature . The calculated maximu m fuel element cladding temperature mus t not exceed 2200°F.

(ii) Maximum cladding oxidation . The calculated total oxidation of the cladd i ng must not at an y location exceed 0.17 ti mes the total cladding thickness before oxidation.

(iii) Maximum hydrogen generation . The calculate d total amount of hydrogen generated from the chemical reaction of the cladding with water or steam must not exceed 0.01 t i mes the hypothet ical am ount that would be generated if all of the metal in the cladding cylinders surro u nding the fuel, excluding the cladd i ng surroun ding the plenum volume, were to react.

(iv) Coolable g eometry . Calculated chan ges in core geometry must be such th at the core remains amenable to cooling.

(v) Long term cooling . After any calculated succes sf ul initial operation of the E CCS, t h e calculated core temperature must be maintained at an acceptably low value and decay heat must be removed for the extended p eriod of time req uired by the long-lived radioact ivity remaining in the core.

Department of Nuclea r S c ien ce and Engineering 6

Breaks larger than the TBS (50.46a)

• T he evaluation must be performed fo r a number of postulated LOCAs o f different sizes and locations suffici ent to provide assurance that the most severe postulated LOCAs l arger than the TBS up to the double-ended rupture of the largest pipe in the reactor coolant system are analyzed. These calculations may take credit for the availability of offsite power and do not require the assumption of a single failure. Realistic initial conditions and availability of safety -related or non safety-related equipment may be assumed if supported by plant-specific data or analysis .

• (i) Coolable g eometry . Calculated changes in core geometry must be such that the core remains amenable to cooling.

• (ii) Long term cooling . After any calculated successful initial operation of the ECCS, the calculated core temperature must be maintained at an acceptably low value and decay heat must be removed for the extended period of time required by the long-lived radioactivity remaining in the core.

Department of Nuclea r S c ien ce and Engineering 7

Plant Changes Under § 50.46b

• License amendment submittals must be risk-informed:

 Must meet criteria consistent wi th RG 1.174 (defense-in-depth, safety margins, monitorin g program, and acceptable risk )

 Must meet PRA quality and scope requirements

 The licensee shall periodically as sess the cumulative effect of changes to the plant, operational practices, equipment performance, and plant operational experience. The assessment must be based upon updated PR A and risk assessments. The assessment must be completed in a timely manner, but no less often than once every two refueling outages.

Department of Nuclea r S c ien ce and Engineering 8

Complications

LBDEGB

10 -5

TBS Break

Department of Nuclea r S c ien ce and Engineering Size 9

Summary of the Expert Opinion Elicitation Process

• Formal elicitation process used to estimate generic BWR and PWR passive-system LOCA frequencies associat ed with material degradation.

• Developed qu antitative estimates for pi ping and non-piping base cases for anchoring elicitation responses.

• Panelists provided quantitative es timates supported by qualitative rationale for underlying technical is sues in individual elicitations.

– Generally good agreement about LOCA contributing factors.

– Large individual uncertainty and panel variability in quantifying estimates.

• Quantitative resu l ts determined by aggregatin g ind i vidual panelists’ estimates.

• LOCA elicitation provides a suffi cient technical basis to support transition break size development.

NUREG-1829, Estimating L O CA Frequencies thro ugh the Eli c itation Pro c es s , Nucl ea r Regulatory Commission, Washington, DC, June 2005.

Department of Nuclea r S c ien ce and Engineering 10

Elicitation Objectives and Scope

• Develop generic BWR and PWR piping and non-piping passive system LOCA frequency distributions as func tion of break size and o p erating time.

 LOCAs which initiate in unisolable portion of reactor coolant system.

 LOCAs related to passive component aging, tempered by mitigation measures.

 Small, medium, and large-break LOCAs examined. Large break category further subdivided to consider LOCA sizes up to complete break of largest RCS piping.

 Time frames considered: 25 years (current day), 40 years (end of original license), and 60 years (end of life extensi o n).

• Primary focu s: frequencies associated with normal operating loads and expected transients.

• Assume that no sign ificant changes will occur in the plant operating profiles.

Department of Nuclea r S c ien ce and Engineering 11

Formal Elicitation Approach

 Co nduct preliminary elicitation.

 Select panel and facilitation team.

 Develop technical issues.

 Quantify base case estimates.

 Formulate eli c itation questions.

 Co nduct individual elicitations.

 Analyze quantitative results and qualitative rationale.

 Summarize and document results.

 Conduct internal and external review of process and results.

Department of Nuclea r S c ien ce and Engineering 12

LOCA Size Classification

• LOCA sizes based on leak rate to group into classes having similar mitigation measures.

 First three categories similar to NUREG-1150 and NUREG/CR-5750.

 Three additional LBLOCA categories used to determine larger break size frequencies.

• Correlations developed to relate flow rate to effective break area.

Department of Nuclea r S c ien ce and Engineering 13

Technical Issues Structure

LOCA C ontribution s

Top Down

Passive System LOCA s

Bottom Up

Active System LOCA s

Piping Contribution

Non-Piping Contribution

Service History

Plant Piping Systems

Componen t

 Elicitation foc u ses on passive

Geometr y

Lo ad in g Hi s t ory

Mit i gation & Maint.

Pum p s

Steam Gen.

Pressure Vessel

system LOCAs.

Materials

Ag i n g Mechs .

Press.

Val ves

 Important piping and non-piping variables identified.

 Elicitation stru cture supports top down and bottom up analyses.

Department of Nuclea r S c ien ce and Engineering 14

Piping Base Case Development

• All elicitation responses relative to the base cases.

• Base case conditions specify the piping system, piping size, material, loading, degradation mechanism(s), and mitigation procedures.

• Five Base Cases Defined.

– BWR

• Recirculation System (BWR-1)

• Feedwater System (BWR-2)

– PWR

• Hot Leg (PWR-1)

• Surge Line (PWR-2)

• High Pressure Injection makeup (PWR-3)

• The LOCA frequency for each base case condition is calculated as a function of flow rate and operating time.

• Four panel members individually estimated frequencies: two using operating experience and two using probabilistic fracture mechanics.

Department of Nuclea r S c ien ce and Engineering 15

Elicitation Questions

• Questions on the following topic areas.

– Base Case Evaluation.

– Regulatory and Utility Safety Culture pertaining to LOCA initiating events.

– LOCA frequencies of Piping Components.

– LOCA frequencies of Non-Piping Components.

• Quantitative Responses:

– Questions are relative to a set of chosen base case conditions.

– Each question asks for mid, low, and high values.

– Questions can be answered using a top-down or bottom-up approach.

• Qualitative Rationale:

– Rationale is provided and discussed for important issues to support quantitative values provided by each panelist.

– Possible inconsistencies between answers and rationales brought to panelists’ attention.

Department of Nuclea r S c ien ce and Engineering 16

Elicitation Insights: BWR & PWR Plants

• BWR Plants

– Thermal fatigue, intergranular stress corrosion cracking (IGSCC), mechanical fatigue, flow accelerated corrosion (FAC) identified as important degradation mechanisms.

– Increased operating transients (e.g., water hammer) compared to PWR plants.

– BWR community has more experience identifying and mitigating degradation due to IGSCC experience in the early 1980s.

– BWR service experience must be carefully evaluated due to preponderance of pre-mitigation IGSCC precursor events.

• PWR Plants

– Primary water stress corrosion cracking (PWSCC), thermal fatigue, and mechanical fatigue identified as important degradation mechanisms.

– PWSCC concerns paramount for panel.

• Near-term frequency increases due to PWSCC are likely before effective mitigation is developed.

• Most panelists believe that issue will be successfully resolved within the next several years.

Department of Nuclea r S c ien ce and Engineering 17

Total LOCA Frequencies: Current Day

10 -2

10 -3

10 -4

10 -5

10 -6

10 -7

10 -8

10 -9

10 -10

BWR: Basel ine Resu l t s

me dian and 95th per c ent i l e r es ults of fs et sl ight l y for c l ari t y

Medi an Mean

95th Per centile

10 -1

10 -2

10 -3

10 -4

10 -5

10 -6

10 -7

10 -8

10 -9

10 -10

PWR: Baseline Results

m edi an a nd 9 5th per c enti l e r esu lts of fse t for c l a r ity

Median Mean

95th P e rc en til e

0. 1 1 10 1 00

Threshold Break Diamet er (in)

0. 1 1 10 10 0

Threshold Break Diameter ( i n)

• Error bars represent 95% confidence bounds accounting for variability among panelist responses.

• Differences between median and 95 th percentile estimates reflect individual panelist uncertainty.

Department of Nuclea r S c ien ce and Engineering 18

Sensitivity Analyses: Overconfidence Adjustment

• Experts are generall y overconfident about their uncertainty.

– Demonstrated using almanac-type questions with kno wn a n swers.

– Rule of thumb: true coverage level is ap proximately half the nominal coverage level.

• Nominal elicitation coverage level: 90% (95 th – 5 th percentiles)

• Imp l ication is that true coverage level i s about 50% (75 th – 2 5 th percentile).

• Evaluate the effect of adjust ing the nominal coverage level.

– No change in the mid value responses

– Evaluate adjustmen t s of error factors associated with bottom line responses for each panelist.

– More ad hoc broad and targeted adjustment schemes evaluated and discussed in NUREG, but not as attractive.

Department of Nuclea r S c ien ce and Engineering 19

Sensitivity Analyses: Aggregating Expert Opinion

• Baseline method uses geometric mean of the individual panelist estimates to determine group estimates for all total LOCA frequency parameters: 5 th , 50 th , 95 th , mean.

– Based on assumed lognormal structure of individual estimates.

– Ensures estimates are not significantly influenced by outliers.

– Results using median or trimmed geometric mean are similar to baseline method.

• Alternative method is to use the arithmetic mean all the individual panelist distributions (mixture distribution).

– Assumes that individual results are obtained from equally credible models that are randomly sampled from population of models.

– Key regulatory parameters may be dominated by outliers.

– Difference between 5 th and 95 th percentiles is much wider.

Department of Nuclea r S c ien ce and Engineering 20

10 -1

10 -2

10 -3

10 -4

10 -5

10 -6

10 -7

10 -8

10 -9

Aggregating Expert Opinion: Arithmetic (AM) vs. Geometric (GM) Mean

B W R Current Day Estimates: PW R Cu r r ent Day Estimates: Mean Frequenc y Mean Frequenc y

10 -1

10 -2

10 -3

10 -4

10 -5

10 -6

10 -7

10 -8

10 -9

1 2 3 4 5 6

L O C A C a te g o r y

1 2 3 4 5 6

LO C A C a tegory

• Aggregated estimates can be sign ificantly affected by approach.

• Similar difference amon g 95 th percentile estimates.

Department of Nuclea r S c ien ce and Engineering 21

Selecting the TBS

• NRC staff adjusted results of expert elicitation process to account for: uncertainties in elicitation process, LOCAs caused by inadvertent actuation of active components, LOCAs caused by large loads (such as heavy loads, seismic, waterhammer), and other considerations (such as degradation in specific piping, specific pipe sizes).

Department of Nuclea r S c ien ce and Engineering 22

Department of Nuclea r S c ien ce and Engineering 23