ML19183A309

Hazard Modeling

ML19183A309
Person / Time
Issue date:	07/02/2019
From:	Office of Nuclear Regulatory Research
To:
K. Hamburger 415-2022
Shared Package
ML19183A307	List: ML19183A308 ML19183A309 ML19183A312
References
Download: ML19183A309 (31)

Text

High Energy Arcing Faults (HEAF)

Hazard Modeling Gabriel Taylor P.E.

Office of Nuclear Regulatory Research Division of Risk Analysis July 24, 2019 Public Meeting - July 24, 2019

Purpose

• Provide overview of modeling

- History

- Types

- Existing models

- Comparisons to measurement Public Meeting - July 24, 2019

Categories of Electrical Enclosure Failure Mode - Review Public Meeting - July 24, 2019

Operating Experience San Onofre Nuclear Generating Station, 2001

• Highlighted HEAF hazard

• NRC INFORMATION NOTICE 2002-27

- RECENT FIRES AT COMMERCIAL NUCLEAR POWER PLANTS IN THE UNITED STATES

- https://www.nrc.gov/docs/ML0226/ML022630147.pdf Public Meeting - July 24, 2019

Fire PRA Methodology NUREG/CR-6850 EPRI 1011989

• NUREG/CR 6850 forms the basis for nuclear power plant (NPP) Fire PRAs

- Published 2005

• This EPRI/NRC working group was the first to explicitly model HEAF events as part of a fire PRA

- The need was identified as part of accident investigation efforts for the development of 6850 & NRCs assessment of energetic faults from 1986-2001 https://www.nrc.gov/readin

• (ADAMS Accession No. ML021290364) g-rm/doc-collections/nuregs/contract

- Timely OpE- San Onofre 2/3/2001 /cr6850/

Public Meeting - July 24, 2019

Current Methodology Electrical Enclosures

• NUREG/CR-6850, Appendix M (2005)

• Zone of Influence (ZOI) Method largely based on one well documented fire event at San Onofre in 2001

• Methodology developed as an expert elicitation

- Observational data and OpE information only

- No test data available

- Currently this model has been used to support NFPA 805 transitions Public Meeting - July 24, 2019 6

HEAF OpE Electrical Enclosure SONGS, 2001 San Onofre; 2001 Onagawa; 2011 Public Meeting - July 24, 2019

Current Methodology Bus Ducts

• NUREG/CR-6850, Supplement 1

• Bus duct guidance for high energy arcing faults (FAQ 07-0035)

• Methodology developed as an expert elicitation

- Observational data and OpE information only

- No test data available

- Currently this model has been used to support NFPA 805 transitions Public Meeting - July 24, 2019

HEAF OpE Bus Duct Diablo Canyon Bus Duct (OpE) Bus Duct Testing Columbia Bus Duct (OpE) 2000 2016 2009 Public Meeting - July 24, 2019

Conceptual Modeling Approaches Public Meeting - July 24, 2019

As presented at 4/18/2018 public workshop Modeling Approach

• Bounding (Current models)

• Enclosure, bus ducts

• Bounding by Categories

• By power, energy, voltage, fault current, protection scheme, material, safety class

• Dynamic ZOI

• Scenario dependent source

• Target fragility 11 Public Meeting - July 24, 2019

As presented at 4/18/2018 public workshop Bounding ZOI (Current Model)

• Assumes worst case damage for all HEAF

- i.e., one size fits all

- Damage and ignition of components within ZOI

- Peak HRR

• Least amount of information needed to determine ZOI

• Least realistic for majority of cases

• Simple to apply

• Lowest cost 12 Public Meeting - July 24, 2019

As presented at 4/18/2018 public workshop Refined Bounding ZOI

• Subdivides equipment by HEAF damaged potential

- Equipment type

- Energy/Power potential

- Protection scheme

- Size, Material, Design, etc.

• More realistic

• Requires more information to apply

• More costly for development and application 13 Public Meeting - July 24, 2019

As presented at 4/18/2018 public workshop Dynamic ZOI

• Requires detailed information on power system

• Correlation from experiments and theory to model source term and incident flux as a function of distance

• Requires knowledge of fire PRA target fragility to high heat flux short duration.

• Potential to provide most realistic results

• Complex

• Most costly 14 Public Meeting - July 24, 2019

Modeling Approach Status

• No approach has been excluded

• Understand and evaluation existing and new hazard models

• Needs to consider development and application efficiencies along with level of realism in a holistically manner to make informed decision on appraoch

• NRC/EPRI working group advancing PRA modeling methodology Public Meeting - July 24, 2019

Overview of Existing Models Public Meeting - July 24, 2019

Theoretical Lee Model

• Simple geometric configuration

- arc modeled as sphere

• Heat transfer to predict distance where threshold is exceeded

• Used available research on human skin / clothing fragility (Stoll / Artz)

• Conservative due to maximum arc power assumption

• Used in IEEE 1584-2002 for > 15kV applications R. Lee, The Other Electrical Hazard: Electric Arc Blast Burns, 1982 Public Meeting - July 24, 2019

Theoretical Lees Method

• Output

- IE, incident energy (J/cm2) ASTM slug

• Inputs T-cap. slug

- V, system voltage (kV)

- t, arcing time (seconds) KEMA Daq

- Ibf, 3 phase bolted fault current

- D, distance from arc point Physical Measurement R. Lee, The Other Electrical Hazard: Electric Arc Blast Burns, 1982 Public Meeting - July 24, 2019

Semi-Empirical Wilkins-Allison-Lang Method

• Output ASTM slug

- IE, incident energy T-cap. slug

• Input

- VLL, line-line voltage KEMA Daq

- V, system voltage

- Varc, arc voltage

- Iarc, arc current

- t, arcing time

- D, distance from arc point

- a, enclosure dimension Physical Measurement

- g, gap

- Ve, electrode voltage Literature R. Wilkins, M. Allison, M. Lang, Improved Method for Arc Flash Hazard Analysis, 2004 Public Meeting - July 24, 2019

Semi-empirical Gammon Simplified

• Output ASTM slug

- IE, incident energy T-cap. slug

• Input KEMA Daq

- MVAsc, short-circuit MVA

- t, arcing time Physical Measurement

- D, distance from arc point

- X, configuration factor (IEEE)

Literature

- IEratioUB, Incident energy rate ratio upper bound (configuration based 0.758 - 2.098)

T. Gammon, J. Matthews, The IEEE 1584-2002 Arc Modeling Debate and Simple Incident Energy Equations for Low-Voltage Systems, 2006 Public Meeting - July 24, 2019

Empirical - Statistical Doughty - Neal - Floyd

• Output ASTM slug

- EMA,EMB, Max. Incident Energy T-cap. slug

• Input

- F, 3-phase short-circuit current KEMA Daq

- tA, tB, arc duration

- DA,DB, distance Physical Measurement R. Doughty, T. Neal, H Floyd, Predicting Incident Energy to Better Manage the Electric Arc Hazard on 600-V Power Distribution System, 2000 Public Meeting - July 24, 2019

Empirical - Statistical IEEE 1584 - 2002

• Guide for performing arc flash calculations

• Model for incident energy calculations

• Empirically derived model from 300 tests

• Methodology focused on personal protection

- Arc flash boundary is only applicable to human fragility

- Arc fault current and incident energy are independent of target IEEE 1584-2002, Guide for Performing Arc-Flash Hazard Calculation, 2002 Public Meeting - July 24, 2019

Empirical - Statistical IEEE 1584 - 2002

• Output ASTM slug

- IE, Incident Energy T-cap. slug

• Input

- V, system voltage KEMA Daq

- Ia, arc current

- t, arc duration

- G, conductor gap Physical Measurement

- D, distance

- x, distance exponent Literature

- Configuration (open / box)

IEEE 1584-2002, Guide for Performing Arc-Flash Hazard Calculation, 2002 Public Meeting - July 24, 2019

Empirical - Statistical IEEE 1584 - 2018

• Guide for performing arc flash calculations

• Significantly changed from 2002 edition

• Model for incident energy calculations

• Empirically derived model from 2,160 tests

• Five configurations IEEE 1584-2018, Guide for Performing Arc-Flash

- VCB, VCBB, HCB, VOA, HOA Hazard Calculation, 2018 Public Meeting - July 24, 2019

IEEE 1584 - 2018 Range of model

• System voltage: 208 to 15,000 Volts

• Frequency: 50 or 60 Hz

• Bolted fault current:

- Low Voltage: 500 to 106,000 A

- Med Voltage: 200 to 65,000 A

• Conductor Gaps:

- Low Voltage: 0.25 to 3 inches

- Med Voltage: 0.75 to 10 inches

• Target Distances: 12 inches

• Fault clearing time: no limit Public Meeting - July 24, 2019

Empirical - Statistical IEEE 1584-2018

• Output ASTM slug

- IE, Incident Energy T-cap. slug

• Input

- Ibf, Bolted fault current

- Voc, System voltage KEMA Daq

- T, Duration

- D, Distance Physical Measurement

- G, Conductor gap

- Enclosure Dimensions Literature

- Equip Configuration Public Meeting - July 24, 2019

Model Comparison IEEE 1584 - 2018 vs MV Alum

• ASTM slug calorimeter (copper)

- Model overpredict max measured incident energy

• Maximum overprediction : ~11x

- 550 kJ/m2 measured vs. 6,100 kJ/m2 calculated

• Minimum overprediction : ~2x

- 3.4MJ/m2 measured vs. 6.3MJ/m2

- Note: 2 instruments damaged due to HEAF damage likely higher heat flux at damaged sensors and better agreement with model Public Meeting - July 24, 2019

Model Comparison LEE vs MV Alum

• ASTM slug calorimeter (copper)

- Model overpredict max measured incident energy

• Maximum overprediction : ~17x

- 550 kJ/m2 measured vs. 9,100 kJ/m2 calculated

• Minimum overprediction : ~3x

- 3.4MJ/m2 measured vs. 9.4MJ/m2 calculated

- Note: 2 instruments damaged due to HEAF damage likely higher heat flux at damaged sensors and better agreement with model Public Meeting - July 24, 2019

Model Comparison IEEE 1584 - 2018 vs MV Alum

• T-cap slug calorimeter (tungsten)

- Model overpredict max measured incident energy

• Maximum overprediction : ~26x

- 236 kJ/m2 measured vs. 6,100 kJ/m2 calculated

• Minimum overprediction : agreement

- 6.0MJ/m2 measured vs. 6.3MJ/m2 Public Meeting - July 24, 2019

Model Comparison LEE vs MV Alum

• T-cap slug calorimeter (tungsten)

- Model overpredict max measured incident energy

• Maximum overprediction : ~39x

- 236 kJ/m2 measured vs. 9,100 kJ/m2 calculated

• Minimum overprediction : ~1.6x

- 6.0MJ/m2 measured vs. 9.4MJ/m2 Public Meeting - July 24, 2019

Wrap-up Existing Models

• Follow similar form

- Inverse power relationship with distance to target

• Supporting test configurations not directly applicable

- Open air or box w/opening

• Fragility different (human vs equipment)

• Existing models may be adapted to make representative and realistic.

Public Meeting - July 24, 2019