Overview of R7 / Risk-Informed Safety Margin Characterization

Bob Youngblood, Robert Nourgaliev, Nam DinhIdaho National Laboratoryy

Second Workshop on

U.S. Nuclear Power Plant Life Extension Research and Development

22 – 24 February 2011, Washington, DC

Light Water Reactor Sustainability R&D Program

22 24 February 2011, Washington, DC

Nuclear EnergyNuclear Energy

Outline • Background Comments:g

– Dynamic PRA, Passive System Functional Reliability, Last 10 years’ work on “margins”

• What we mean by “Margin”• What we mean by “Margin”

• R7 Development Within the Risk-Informed Safety Margin Characterization (RISMC) Pathway of the Light Water Reactor Characterization (RISMC) Pathway of the Light Water Reactor Sustainability Program

• The Collaboration– Lab / Academic Partners– EPRI

P th F d• Path Forward

2

Background Comments

• Th h b l t f l t k i th l t 30 i t ti • There has been a lot of relevant work in the last 30 years, integrating Human Reliability Analysis, epistemic and aleatory uncertainty, phenomenology, and so on

– Dynamic PRA– Dynamic PRA– Passive System Functional Reliability– Last 10 years’ work on “margins”: USNRC-sponsored work, CSNI, SM2A, …

• While trying to break new ground methodologically, much of this work tried to use legacy codes

• The work described here is intended to support sustainability pp ydecision-making by:

– Integrating the above-mentioned elements into the analysis, along with:• Passive component behavior, effects of off-normal but not necessarily

i di t d i immediate core damage scenarios, …– Developing a modern tool (R7) to better support this kind of analysis

3

Margin“Margin” is about P(Load<Capacity). “The Logo:”

Therefore, – It’s not enough to try to characterize

margin JUST by a distance between designated points (e g mean values

0.5

0.6

y

Basis for quantifying P(L>C)

designated points (e.g., mean values, designated percentiles) of load and capacity pdf’s

– One must also say something about 0 1

0.2

0.3

0.4

Prob

ability Den

sity

PDF(Load)

PDF(Capacity)

uncertainties in L and C…– … and try to quantify P(L>C)– Develop engineering insight into

diti d hi h f il ill

0

0.1

0 2 4 6 8 10 12

Applied Load (SSC Capacity)

conditions under which failure will occur: i.e., L>C

I i

Abundant Margin

Less-Than-Abundant Margin

L: “Load”C: “Capacity”

Increasing “Margin”

MargingP(L>C |

Hardware Success)

<< P(Hardware Failure)

P(L>C | Hardware Success)

P(Hardware Failure)

P(L>C | Hardware Success)

~ P(Hardware Failure)

??<

Should Quantify

4

Should Quantify

R7

5

6

The State of RISMC / R7103

104

105

106

107

108

CF

Lm

at

50

60

10-1

100

101

102

W

660 665 670 675 680 685 690 695 700 705

10

20

30

40

50

Co

re p

ow

er,

MW

95

9899

99.5

99.9

99.99660 665 670 675 680 685 690 695 700 705

102 MC samples

103 MC samples

104 MC samples

105 MC samples

%

0 1 2 3 4 5

0

Years

107

108

40506070

80

90

95

12

CD

F,

ap

ac

ity

limit

100103

104

105

106

0.1

102 MC samples

PD

F

Ca

103 MC samples

104 MC samples

105 MC samples

100,000 runs in ~16hrson 512 512 CPUs ofCPUs of

7x10-1

8x10-1

9x10-1

10

Reference: 104 LHC

CD

F

97 LHC samples PCE ith 97 l ti

1.00 1.05 1.10 1.15

7660 665 670 675 680 685 690 695 700 705

0.01

PCT, K

ICESTORMICESTORM

640 650 660 670 680 690 700 7105x10-1

6x10-1

PCT, K

PCE with 97 evaluations

CPU time:CPU time:<2 min<2 min

8

Abderrafi Ougouag

Nam Dinh, INLRobert NourgalievINL

INLJoshua CogliatiINLBui Viet Anh

INL

Brandon BlackburnINL

Frederick Gleicher INL

Samet KadiogluINL

Pavel MedvedevINL

Hong Luo NCSU

Steve ShkollerUC Davis

Brian WilliamsLANL

Laura SwilerSNL

Mike PerniceINL

Hany S. Abdel-KhalikNCSUNCSU

Dana KellyINL

Bob YoungbloodINL

Tunc AldemirOSU

9

Ralph NelsonLANL

Barton SmithUSU

Qiao WuOSU

Tami Grimmet INL

Active Collaboration with Electric Power Research Institute (EPRI)• Key input into formulation of “use cases” to steer R7

• Collaboration on case studies to sharpen formulation of / ill t t th “l ” f killustrate the “logo” framework

• Examples: 1 Selection of “feed and bleed” as an intermediate term objective for R7 1. Selection of “feed and bleed” as an intermediate-term objective for R7

development2. EPRI 1021085 – Desired Characteristics for Next Generation Integrated

Nuclear Safety Analysis Methods and Software / Stephen M Hess / Senior Nuclear Safety Analysis Methods and Software / Stephen M. Hess / Senior Project Manager / Risk and Safety Management

3. Pending R7 availability for realistic feed-and-bleed analysis: refine and illustrate the “logo” framework using MAAP as a surrogate for R7 (EPRI-led)illustrate the logo framework using MAAP as a surrogate for R7 (EPRI led)

4. ANS panels & papers (e.g., Hess, Youngblood, and Vasseur: “Recent Trends in Risk Informed Safety Margin Characterization”)

10

Time Line

11

Summary

• R7 is being developed as a modern systems code– Improved modeling and numerics

Address sustainability issues – Address sustainability issues • Including DBA type scenarios• Scenarios that threaten the plant, if not the public• Address other plant safety issues• Address other plant safety issues

– Address uncertainty from the beginning: “Do the Logo”• In-line modeling of aleatory uncertainty (reliability issues)• Close coupling to tools like Dakota for epistemic uncertainty• Close coupling to tools like Dakota for epistemic uncertainty

– Develop engineering insight into the conditions under which failure will occur (i.e., L>C)

• The development is “use-case-driven” and closely coordinated with EPRI and EPRI contractors

12

13

“A decision made without taking uncertainty into account is barely worth calling a decision”, R.E. Wilson (1985)

PDF

LoadLoadCapacityCapacityp yp y

a number

a number

Safety Figure of Merit

14

Damage Mechanisms – 1/2Other disciplines       Time scales

Embrittlement

p(beside materials science)

Core neutronics Long (fluence)Fluid (mixing) Short (thermal shock) Heat transferStructural mechanics

Thermal fatigue (striping)

Fluid flow Long (accumulation)Sh ( ill i )H f

Structural mechanics

(striping) Short (oscillation) 

Structural fatigue Fluid flow Long (accumulation)

Heat transfer

Structural fatigue (vibration)

Fluid flow Long (accumulation)Short (oscillation) Structural mechanics (FSI)

15

Damage Mechanisms – 2/2Other disciplines      (beside materials science)

Time scales(beside materials science)

Flow‐Accelerated Corrosion

Coolant chemistry Long (accumulation)Fluid flowFluid flow

Stress Corrosion Cracking (SCC)

Coolant chemistry Long (accumulation)Fluid flowCracking (SCC) Fluid flow

Irradiation‐Assisted Stress Corrosion 

Neutronics Long (accumulation)Radiation chemistry (+ CC)

Other damage mechanisms: 

Cracking (IASCC)y ( )

Fluid flow

gErosion Corrosion, Liquid Droplet Impingement‐Induced Corrosion, Cavitation Corrosion, Environmentally Assisted SCC

16

Modeling of Plant Aging

Objective: Capture vulnerability hidden under four modes of negative synergy (non-intuitive / no-”prior” nonlinearity).

1. Synergy of physics: neutronics, fluid flow, heat transfer, structural mechanics, coolant / radiation chemistry

2 S f lti l d h i ( l t d d d ti )2. Synergy of multiple damage mechanisms (accelerated degradation)

3. Synergy of safety threats (i.e. addition of fire, flooding, earthquake challenge) on already weakened SSCschallenge) on already weakened SSCs

4. Synergy between aging / degraded materials and aging / obsolete procedures, training, I&C, surveillance and diagnostics strategiesp , g, , g g

17

Approach to Modeling of Plant Aging

Task: Compute Loading and Compute Capacity on sensitive SSCs

1. Multi-physics (needed for each damage mechanism)1. Multi physics (needed for each damage mechanism)

2. Comprehensive coverage of different damage mechanisms

3. Amenable for (future) integrated treatment of safety threats

4. Address plant complexity in its totality (e.g. by including effect f I&C ti d t ti S&M)of I&C, operating procedures, operator actions, S&M)

18

Computational Challenges

Multi‐Physics Integration:‐ CN, TH … + SM, CC + damage / degradation models   

Multi Scale Integration:Multi‐Scale Integration:‐ Need for “gap‐tooth” scheme …. to zoom in to capture

‐‐ rates of damage growth under high‐frequency stressors;‐‐ accelerated damage growth in “shock” events.accelerated damage growth in  shock  events.

‐ Need high fidelity (e.g. 3D CFD in the vicinity of sensitive SSCs) to capture the delicate stressors   0D/1D/3D transitionsCode Calibration:

‐ Incomplete historical data to reconstruct “plant health” Vulnerability Search:

‐ Need effective sampling strategies to overcome local minima.

19

Elements of the Strategy

Ri k I f d S f R S fN G iRisk-Informed Safety Margin Characterization(RISMC) Methodology

Reactor Safety Research & Testing(V&V) Infrastructure

Next GenerationSystem Safety

(R7) Code

R7 R&DR7 R&D

Broad Stakeholders and Support Base

I-NEST CORE “Safety & Licensing”

“Modeling, Experimentation, and Validation” (MeV)

Schools

20

21

22

Obj tObj t i t d (C )i t d (C )

Recovery DG (1D)Recovery DG (1D)(up to 12(up to 12thth--order)order)

ObjectObject--oriented (C++)oriented (C++)Parallel (MPICH)Parallel (MPICH)HighHigh--order in spaceorder in space

Reconstructed DG Reconstructed DG (2D & 3D unstructured(2D & 3D unstructured--grid)grid)

(3(3rdrd--order and higher)order and higher)F llF ll i li it Li li it L t bl Hi ht bl Hi h d d

(3(3 order and higher)order and higher)FullyFully--implicit, Limplicit, L--stable, Highstable, High--order order

in timein timeStateState--ofof--thethe--art linear algebra art linear algebra

((KrylovKrylov based JFNK based JFNK PETScPETSc))

LL--stablestable

AllAll--speed capabilities, t/d consistent speed capabilities, t/d consistent real fluid EOSsreal fluid EOSs

((KrylovKrylov--based, JFNK, based, JFNK, PETScPETSc))

Advanced 2Advanced 2--phase flow modelingphase flow modeling LL--stablestable

Uncertainty Quantification enabled Uncertainty Quantification enabled

Advanced 2Advanced 2 phase flow modelingphase flow modelingCFD capabilitiesCFD capabilities

Homogeneous Equilibrium (+Interfacial Slip) ModelHomogeneous Equilibrium (+Interfacial Slip) Model

TwoTwo--fluid fluid 66--equation hyperbolic equation hyperbolic 11--pressure pressure (+interfacial pressure) model(+interfacial pressure) model

NeutronicsNeutronics: nodal diffusion: nodal diffusion

“Born“Born--assessed”, with assessed”, with subsub--version version control (control (TracTrac webweb--based), based), extensive verification and documentation (“ANT” Manual extensive verification and documentation (“ANT” Manual

Uncertainty Quantification enabled Uncertainty Quantification enabled (integration with DAKOTA)(integration with DAKOTA)

TwoTwo--fluid fluid 77--equation hyperbolic equation hyperbolic 22--pressure pressure model (DEM)model (DEM)

23

extensive verification and documentation ( ANT Manual, extensive verification and documentation ( ANT Manual, DoxygenDoxygen –– htmlhtml--based Manual), 3D visualization, GUI, regression based Manual), 3D visualization, GUI, regression testingtesting

I/O, GUI, code control utilities Component Factory

Computational engineComponent & Interface

Pre/Post-Processors

Computational engineTime discretization,

Transient control

Component & Interface Library

1-phasepump

2-phasepump

1-phasepipe 2-phase

pipe2-phaseelbow 1 phase Li Al b p p

Turbine

1-phaseelbow

elbow 1-phaseT-junc.

Pressurizer

Linear Algebra, Nonlinear Solvers

PETScPETScTrilinosTrilinosUQ & SA

DakotaDakota

Component utilitiesand Material

Library ClosureLibrary

DakotaDakotaADIC (FSA & ADIC (FSA & AdjointAdjoint))

24

Library of templates Library Library

(in progress)(in progress)

Expansion/contractions, 3Expansion/contractions, 3--pipe junctionspipe junctions

FREESTEAMFREESTEAM water property package water property package

new realistic model is developednew realistic model is developed

in the process of testingin the process of testing

25