ML21230A357

LBB Studies V4

ML21230A357
Person / Time
Issue date:	08/18/2021
From:	NRC/RES/DE
To:
Homiack M
Shared Package
ML21230A354	List: ML21230A356 ML21230A357 ML21230A358 ML21230A360 ML21230A361 ML21230A363 ML21230A364 ML21230A366
References
Download: ML21230A357 (11)

Text

Leak-Before-Break Studies NRC Phase 1 LBB Report/ Plans for LBB Phase 2 and EPRI Report August 18, 2021 1

Introduction

- Application of xLPR to a real-life problem: Impact of primary water stress-corrosion cracking (PWSCC) on a Reactor Vessel Outlet Nozzle (RVON) for a Westinghouse 4-loop Nuclear Power Plant

- Largest number of same weld type with Alloy 82/182 within leak-before-break (LBB)-approved line segments in the U.S. nuclear fleet

- Coordinated effort by NRC/RES and EPRI via a memorandum of understanding, with the following goals:

1. Develop a defensible strategy of xLPR use for risk-informed decision making

2. Initiate the first phase of a system-level analysis considering a loop or a whole unit

3. Test different Quantities of Interest (QoIs) as criteria for decision-making

- Program to be followed by a second phase considering other weld locations and combination of results at the system level 2

Problem considered

• Reactor Vessel Outlet Nozzle (RVON) for a generic Westinghouse 4-loop pressurized-water reactor

- 34 inches diameter (0.869 m)

- 2.6 inches thickness (0.0663 m)

- Temperature : 320.5°C

- Pressure 15.41 MPa

- Focus on circumferential cracks (axial cracks considered in sensitivity run)

- Simulation time : 80 years (timestep : 1 month)

- Large number of uncertain inputs (e.g.,

welding residual stresses (WRS), crack initiation parameters, crack growth parameters) 3

Quantities of Interest

• Probability of rupture with leak rate detection

• Probability of Leak Rate Jump: going from less than 10 gpm to more than 50 gpm in one timestep

• Probabilistic LBB ratio (critical crack size over detectable leakage crack size)

- Similar to the deterministic ratio in NRC Standard Review Plan Section 3.6.3

- Estimated for each realization leading to a distribution of the LBB ratio

• Distribution of time from detectable through wall crack to rupture

• Outputs in blue are directly estimated in xLPR - other outcomes need postprocessing but can be done easily in Excel 4

Probabilistic Approach

- Considered acceptance threshold is 10-6 occurrence per reactor year.

- Equivalent to probability of occurrence of 8 x 10-5 over 80 years

- Sample size considered between 70,000 and 100,000 should lead to respectively 5 to 8 events on average, if the probability is higher than the threshold

- Analysis conditional on having a crack at time 0 use sample size of 5,000 since probability of having crack over 80 years is in the 10-3 range

- Large amount of data saved, so a single probabilistic run was not practical

- 7 replicates of 10,000 to 15,000 sample size, each with different random seeds, were run and combined

- Crack leak rate, inner length, outer length, and critical ratio extracted to estimate the composite QoIs 5

Reference results (1/2)

- Probability of leak rate jump and rupture with leak rate detection too low to occur over 100,000 realizations

- Good on a regulatory standpoint since it shows the risk is below the acceptable range. Not ideal on a scientific standpoint as there is no quantification of the risk

- Rather, alternate QoIs are presented (P(1st crack),

P(1st leak) and P(rupture) without inspection or leak rate detection) 6

Reference results (2/2)

• Composite QoIs give results for different criteria (for instance 1 gpm and 10 gpm leak rate detection limits)

• Results are conditional on having pipe rupture and do not reflect this probability 10 gpm 10 gpm 1 gpm 1 gpm Time between detectable LBB ratio leak rate and rupture 7

Sensitivity Case: Impact of WRS Profile (1/2)

• New WRS mean profile with similar uncertainty applied

• Larger mean at insider diameter leads to higher probability of 1st crack

• Lower minimum reduces probabilities of 1st leak and rupture relative to the reference value Still, no ruptures with leak rate detection or leak rate jump events 8

Sensitivity Case: Impact of WRS Profile (2/2)

• Stronger impact on LBB ratio and LBB time lapse than for the reference case

• Use of multiple metrics gives more insights on scenario differences and increases confidence in the results 9

Sensitivity Case: Impact of Temperature

• Change in operating Still, no ruptures with leak rate detection or leak rate jump events temperature (deterministic value) from 320.8°C to 330°C

• Considered large change since the same temperature is applied through the whole 80-year simulation time

• Temperature affects initiation and growth, thus impact in all three metrics presented 10

Conclusions and Perspectives

• This study helped identify bugs and enhancements for xLPR (they will be fixed in Version 2.2) but overall, the code is ready to be used for regulatory decisionmaking

• Beyond the already available outputs, composite QoIs can be defined and calculated in post-processing without too much effort

• This study helped identify advantages and drawbacks for each QoI considered and showed the benefit of using multiple metrics

• This first analysis serves as a good basis to extend the process to other weld types (pressurizer surge line, reactor coolant pump ) and estimate risk at the system level 11