ML18213A081
| ML18213A081 | |
| Person / Time | |
|---|---|
| Issue date: | 07/31/2018 |
| From: | Tammie Rivera NRC/RES/DRA/FRB, Electric Power Research Institute |
| To: | |
| Shared Package | |
| ML18213A072 | List: |
| References | |
| Download: ML18213A081 (89) | |
Text
Module III - Fire Analysis Task 6 - Fire Ignition Frequency Joint EPRI/NRC-RES Fire PRA Workshop August 6-10, 2018 A Collaboration of the Electric Power Research Institute (EPRI) & U.S. NRC Office of Nuclear Regulatory Research (RES)
Fire Ignition Frequencies Purpose of Task 6 per NUREG/CR-6850 / EPRI 1011989 Task 6 establishes fire source ignition frequencies
- How often do we expect to see a particular type of fire or a fire in a particular location?
Just for reference - what sort of numbers are we talking about?
- Plant-wide frequency for all fire sources is roughly*
0.2 fire events per year (or 2.0E-1/yr)
- That is roughly* one fire somewhere in the plant every 5 years
- Most of these fires are small, quickly suppressed, and cause no damage beyond the ignition source
- We are still interested in these events because well ask the risk question:
What would happen if a similar fire occurs that is not quickly suppressed and/or in a critical location more sensitive to fire damage?
- Emphasis on roughly - this is an illustrative discussion only. The actual plant wide fire frequency depends on which source you use. The original 6850/1011989 value was 2.89E-1/ry. Well talk about these topics as we continue 2
Corresponding PRA Standard Element Task 6 maps to element IGN - Ignition Frequency
- IGN Objectives (per the PRA Standard):
Establish the plant wide frequency of fires of various types on a generic basis for NPPs Tailor the generic fire frequency values to reflect a particular plant Apportion fire frequencies to specific physical analysis units, and/or fire scenarios 3
IGN HLRs (per the PRA Standard)
HLR IGN-A: The Fire PRA shall develop fire ignition frequencies for every physical analysis unit that has not been qualitatively screened (10 SRs)
HLR IGN-B: The fire PRA shall document the fire frequency estimation in a manner that facilitates Fire PRA applications, upgrades, and peer review (5 SRs) 4
Fire Ignition Frequency Evolution NUREG/CR-6850 and NUREG/CR-6850 S. 1 NUREG-2169 and EPRI 1011989 and EPRI 1019259 EPRI 3002002936 Original methodology Updated counting Updated methodology and frequencies guidance and updated with additional decade frequencies from FAQ of data event data.
process Most recent fire frequencies for at-power and LPSD.
EPRI 1022994 and EPRI 1025284 Improved fire frequency calculation methodology and fire event collection and classification through 2009.
5
Summary of Available Reference Material NUREG/CR-6850 / EPRI 1011989
- Overall methodology
- General bin descriptions Supplement 1
- Chapter 3: Ignition source counting guidance for electrical cabinets FAQ 06-0016
- Chapter 4: Ignition source counting guidance for high-energy arcing faults (HEAFS) FAQ 06-0017
- Chapter 5: Ignition source counting guidance for main control board (MCB) FAQ 06-0018
- Chapter 6: Miscellaneous fire ignition frequency binning issues FAQ 07-0031
- Chapter 7: Bus duct (counting) guidance for high energy arcing faults FAQ 07-0035
- Chapter 10: Fire ignition frequency FAQ 08-0048 Recent FAQs
- FAQ 12-0064 Hot work/transient fire frequency: influencing factors
- FAQ 14-0008 Main control board treatment (counting)
- Updated fire ignition frequencies through 2009 6
Fire Ignition Frequencies High level summary General approach
- Fire event data
- Sources of data
- Types of ignition sources Task 6 procedure
- Step by step
- Counting example 7
Fire Ignition Frequencies A note on terminology We have noted that different documents use different terms for the physical plant partitions used in fire PRA
- NUREG/CR-6850 / 1011989 refers to fire compartments
- The standard refers to physical analysis units or PAUs This makes no difference to Task 6 fire frequency analysis
- You are developing fire ignition frequencies for whatever set of fire locations you have defined
- Whether you call it a fire area, fire compartment or PAU does not really matter - it is what is in that location that counts The total frequency for any location is simply the sum of the frequencies for the ignition sources present in that location
- Once you get to the scenario level (individual fire sources or fire source groups) the differences are totally irrelevant You are estimating fire frequency for a very specific ignition source 8
Fire Ignition Frequencies General approach The generic fire frequencies are based on the collective experience of the US nuclear power industry
- EPRI Fire Event Database (FEDB) included data from 1968 through December 2000 including over 1400 records
- EPRI published the Updated Fire Events Database in 2013 extending the collective experience through 2009 (approximately 2000 records)
Although the database quality and supporting information has advanced in the Updated FEDB, there are still some limitations
- Inconsistent reporting practices - reporting has been largely voluntary
- Uneven data collection - different folks at different times have added data using different sources and bases
- Completeness of event descriptions - reports tend to focus more on plant response to event than the details of fire Industry data collection through INPO has enhanced the collection efforts through a standardized reporting process 9
Fire Ignition Frequencies Fire Event Data In the end each event is binned considering four primary factors
- Was it a risk-relevant fire (potentially challenging or greater)?
- What was the fire source?
- How was the fire put out?
- How long did the fire last?
The historical fire event set is used to estimate generic fire frequencies
- Generic frequencies are in events/reactor-year
- Numbers are for one unit (i.e., frequency plant-wide for a single plant unit)
Event sources (* indicated as primary source for Updated FEDB)
- Mandatory reporting to NRC*
Licensee event reports (LERs)
Event notifications (ENs)
- Comprehensive NRC records search
- Voluntary reporting to industry sources Nuclear Energy Insurance Limited (NEIL) and American Nuclear Insurers (ANI)
- Ad-hoc additions based on specific PRA studies IPEEE analysis during early to mid 1990s
- Plant fire reports*
Individual follow up for key events or for events requiring additional information Corresponding PRA Standard SRs: IGN-A1, IGN-A5, & IGN-A10 10
Fire Events Database Screening 1 Request CR List for Key Words n = input from utilities required (2,000 - 4,000 / unit)
= Audit (NRC)
Owners Groups Perform m Screening Review 1
Request CRs for Probable 2 Fires 102 units responded to #1 (75 - 125 / unit) 100 units responded to #2 EPRI Screen for Potential Significance 85 units responded to #3 Important 2 Existing FEDB Site Follow-up for 3
Fire Events Real Fires (0 - 5 / unit) (10 - 15 / unit)
EPRI Categorize Events in New Important 3 FEDB Fires Potentially Challenging and (0-3/unit)
Challenging Fires 11
Not all events in the database are risk-relevant Identification of the events that count to risk Challenging (CH) fires that had an observable and substantive effect on environment outside the ignition source Potentially challenging (PC) fires are those that could have grown or caused damage under foreseeable alternate circumstances
- May be potentially challenging even if no damage occurred
- Ask: what could happen given the same sort of fire in a different location, failures of fire protection defense in depth, or delays in successful intervention?
Non-challenging (NC) fires that did not cause or would not have caused adjacent objects or components to become damaged regardless or location for essentially any amount of time. These fires are not counted in frequency
- Not a location of interest to PRA (e.g., parking lot fires, off-site fires)
- Occurred during plant construction
- Case specific rules such as:
Hot work fires suppressed by a fire watch using a single extinguisher Smoked component reports (e.g., a burned out relay with no suppression needed, no signs of damage beyond that one failed component)
Some event records we simply could not tell (Undetermined (U PC-CH or U NC-PC) 15
Fire Ignition Frequencies Counting the fire events For EPRI 3002002936 / NUREG-2169 frequencies:
- Challenging, potentially challenging, and undetermined (PC-CH) fires count as one event each
- Undetermined (NC-PC) fire events count as 0.5 events For original and supplement 1 frequencies:
- Potentially challenging events count as one event each
- Weighted the unknowns to get final event count, so We assume that resolution of all unknowns would yield the same split between PC and NC as we got for those events we could resolve If we had 100 raw events in a bin We classify 30 as non-challenging We classify 40 as potentially challenging The other 30 were unknown Each unknown event would be weighed based on ratio of PC to total resolved: W = [40÷(40+30)] = 4/7 So for this example our total event count would be:
57.1 = [40 + 30xW)]
16
Fire Ignition Frequencies Assumptions The model developed for estimating fire ignition frequencies is based on the following assumptions:
- The generic frequencies are fire events per reactor year per operating unit
- Frequencies remain constant over time
- Each fire event is assigned to an Ignition Source Bin and frequencies are calculated for each bin See Table 6-1, Bins 1-37 (electrical cabinets, motors, pumps)
- Total unit-wide ignition frequency for each ignition source bin is the same for all units in the US fleet Unit A at Plant X has the same plant-wide frequency of electrical cabinet fires as Unit B Plant Y
- Within a plant, the ignition frequency is the same for each individual member of a given ignition source bin At Unit A of Plant X, each individual electrical cabinet is assigned the same fire frequency (frequency of cabinet A = frequency in cabinet B) 17
Fire Ignition Frequencies Available data There are now three sources of fire ignition frequency data:
- NUREG/CR-6850 and EPRI 1011989 (2005)
Original fire PRA data for fire ignition frequencies from 1965 through 2000 Event details are limited and typically uninformative on fire attributes Consider events prior to the implementation of Appendix R fire protection programs
- EPRI 1016737 / Supplement 1 / FAQ 08-0048 (2008)
Update of original fire ignition frequencies that considered potential industry trends (i.e., towards reduced fire frequencies)
- EPRI / industry proposed that some ignition source bin frequencies have decreased based on analysis of post 1990s data
- This set weights the more recent data (1991 forward) more heavily If using this set, review the NRC staff position on FAQ 08-0048 (ML092190457)
As of May 2015, NUREG-2169, EPRI 3002002936 supersedes FAQ 08-0048 (ML15134A046) 18
Fire Ignition Frequencies Available data (continued)
- EPRI 3002002936 / NUREG-2169 (2014)
Included additional decade of fire event data (through 2009)
Improved methodology / different calculations for sparse versus medium or dense event sets per EPRI 1022994 Most current and complete event set
- Split data into three time periods 1968-1989 - used to develop prior 1990-1999 - used as update period for sparse bins (20 year update) 2000-2009 - used as update period for medium and dense bins (10 year update)
- Fire event density determined by 2000-2009 time period Sparse bins (< 2.5 events)
Medium or dense bins (> 2.5 events) 19
Fire Ignition Frequencies Generic data sources - NUREG-2169 20
Fire Ignition Frequencies Comparison of generic data sources Total sum of all bins
- NUREG/CR-6850 EPRI 1011989 2.89E-01
- EPRI 1016735 1.50E-01
- Updated FEDB 2.10E-01 21
Fire Ignition Frequencies A note on nomenclature Fire frequency = = (fire ignition events) / (specified time period)
- Time period of interest is either a reactor year (ry) or calendar year (cy)
- We generally work on an ry basis (more on this a bit later)
You will see subscripts that designate location and/or source type
- MCR = fire frequency for the main control room
- HW,J = fire frequency for hot work in location J
- Be careful because subscripts are context driven We can estimate frequency at many levels of detail
- The entire plant (e.g., the generic tables in 6850/1011989)
- A building
- A PAU
- Ultimately we often want frequency for a specific fire ignition source in a specific location 22
Fire Ignition Frequency Ignition source binning differences Caution - fire ignition source bin numbers and frequency basis for electrical cabinet HEAFs vary among the frequency data sets Mapping of electrical cabinet HEAFs by source document NUREG-2169 / EPRI NUREG/CR-6850 / EPRI Supplement 1 / FAQ 08-3002002936 Fire source 1011989 0048 and EPRI 1016735 (sparse event set with 20 year binning basis: (based on event set (heavily weighted 1990s update and legacy data used for 1965-2000) data) prior)
One bin for all Bin 16 Bin 15.2 Not available cabinet HEAF Split bins for cabinet HEAF Bins 16.a and 16.b Not available Bins 16.a and 16.b based on (Supp.1 Chapter 4) voltage level 23
Fire Ignition Frequency Ignition source binning differences Caution - there are two sets of frequency values for bus ducts Mapping of bus duct HEAF frequency sets by source document NUREG-2169 / EPRI NUREG/CR-6850 / EPRI Supplement 1 / FAQ 08-3002002936 Fire source 1011989 0048 and EPRI 1016735 (sparse event set with 20 binning basis: (based on event set (heavily weighted 1990s year update and legacy 1965-2000) data) data used for prior)
FAQ 07-0035 (Supp. 1 Bus Duct HEAFs Ch. 7) Bins 16.1 and 16.2 Bins 16.1 and 16.2 No bin numbers provided 24
Fire Ignition Frequencies General approach Start with pre-calculated unit-level generic fire ignition frequencies (IS)
These are given for roughly 40 ignition source bins, for example:
Bin 21: general pump fires = 2.72E-02/ry About 1 fire every 37 years Bin 15: general fires in electrical cabinets = 3.0E-2/ry About 1 fire every 100 years Bin 37: transient fuel fires in the turbine building = 6.71E-03/ry About 1 fire every 150 years Bin 4: main control board (MCB) fires = 4.91E-3/ry About 1 fire every 205 years Bin 1: battery fires = 1.96E-4/ry About 1 fire every 5000 years 25
Fire Ignition Frequencies General approach We then distribute (partition or apportion) the unit-level frequency to suit needs of a scenario analysis Again, to a building, room, PAU, or to individual members of the ignition source bin Some bins are partitioned by population, others by location Fixed ignition sources we generally count and apportion based on local population versus plant-wide population How many are here versus total in the plant Exceptions: cables and junction boxes (more later)
Non-fixed ignition sources are apportioned by location Hot work and transients Qualitative weighting factor method We will cover details of both approaches 26
Fire Ignition Frequencies General approach The fire frequency for a location (e.g., a PAU) or for a scenario is the simple sum of the fire frequencies for each ignition source present in the location or that contributes to the scenario:
27
Fire Ignition Frequencies General approach There is a second weighting factor to consider for multi-unit sites:
WL is a location weighting factor used only for shared locations at multi-unit sites under specific conditions
- Common applications: turbine building and main control room There may be others
- For shared locations, you may need to double (or triple) the fire frequency to reflect contributions from multiple plant units We will cover this under Step 7 28
Fire Ignition Frequencies What are some examples of NPP fire ignition sources?
Photo from: http://newenergyandfuel.com/wp-content/uploads/2013/01/Nuclear-Plant-Block-Diagram.gif 29
Fixed Fire Ignition Sources Batteries Off-gas/H2 Recombiners Reactor Coolant Pumps Pumps Main Control Board RPS MG Sets Diesel Generators Transformers Air Compressors Ventilation Subsystems Battery Chargers Yard Transformers Dryers Boilers Electric Motors Main Feedwater Pumps Electrical Cabinets Turbine Generator Excitor High Energy Arcing Turbine Generator Hydrogen Faults Turbine Generator Oil Hydrogen Tanks Misc. Hydrogen Fires 30
Fire Ignition Frequency Non-countable sources PWR Containment
- Transients and Hotwork Reactor / Control / Auxiliary Building
- Cables fires caused by welding and cutting
- Transient fires caused by welding and cutting
- Transients Turbine Building
- Cables fires caused by welding and cutting
- Transient fires caused by welding and cutting
- Transients Plant-Wide Locations
- Cables fires caused by welding and cutting
- Transient fires caused by welding and cutting
- Transients All Locations
- Self ignited cables fires
- Junction boxes 31
Fire Ignition Frequencies Plant Level Ignition Source Bins (Table 6-1)
Table 6 -1 Fire Frequency Bins and Generic Frequencies Generic Split Fractions for Fire Type Ignition Source Freq ID Location Mode 1 (Equipment Type) (per rx yr) Electrical Oil Transient Hotwork Hydrogen HEAF 1 Battery Room Batteries All 7.5E-04 1.0 0 0 0 0 0 2 Containment (PWR) Reactor Coolant Pump Power 6.1E-03 0.14 0.86 0 0 0 0 4 Control Room Main Control Board All 2.5E-03 1.0 0 0 0 0 0 8 Diesel Generator Diesel Generators All 2.1E-02 0.16 0.84 0 0 0 0 Room 11 Plant-Wide Cable fires caused by Power 2.0E-03 0 0 0 1.0 0 0 Components welding and cutting Ignition Frequency Bin 14 Plant-Wide Electric Motors All 4.6E-03 1.0 0 0 0 0 0 Components 15 Plant-Wide Electrical Cabinets All 4.5E-02 1.0 0 0 0 0 0 Components 20 Plant-Wide Off-gas/H2 Recombiner Power 4.4E-02 0 0 0 0 1.0 0 Components (BWR) 27 Transformer Yard Transformer - Catastrophic2 Power 6.0E-03 1.0 0 0 0 0 32 Turbine Building Main Feedwater Pumps Power 1.3E-02 0.11 0.89 0 0 0 0
- 1. See Appendix M for a description of high-energy arcing fault (HEAF) fires.
- 2. See Section 6.5.6 below for a definition.
Note that these slides use the original 6850/1011989 frequency table, not the updated table from the supplement or EPRI 3002002936 / NUREG-2169.
32
Fire Ignition Frequencies Plant Level Ignition Source Bins Table 6 -1 Fire Frequency Bins and Generic Frequencies Generic Split Fractions for Fire Type Ignition Source Freq ID Location Mode 1 (Equipment Type) (per rx yr) Electrical Oil Transient Hotwork Hydrogen HEAF 1 Battery Room Batteries All 7.5E-04 1.0 0 0 0 0 0 2 Containment (PWR) Reactor Coolant Pump Power 6.1E-03 0.14 0.86 0 0 0 0 4 Control Room Main Control Board All 2.5E-03 1.0 0 0 0 0 0 8 Diesel Generator Diesel Generators All 2.1E-02 0.16 0.84 0 0 0 0 Room 11 Plant-Wide Cable fires caused by Power 2.0E-03 0 0 0 1.0 0 0 Components ID Location welding and cutting 14 Plant-Wide Electric Motors All 4.6E-03 1.0 0 0 0 0 0 Components 1 Battery Room 15 Plant-Wide Electrical Cabinets All 4.5E-02 1.0 0 0 0 0 0 Components 2 Containment (PWR) 20 Plant-Wide Off-gas/H2 Recombiner Power 4.4E-02 0 0 0 0 1.0 0 Components 4 Control (BWR) Room 27 Transformer Yard Transformer - Catastrophic2 Power 6.0E-03 1.0 0 0 0 0 8 Diesel Generator Room 32 Turbine Building Main Feedwater Pumps Power 1.3E-02 0.11 0.89 0 0 0 0
- 1. See Appendix M for a description of high-energy arcing fault (HEAF) fires.
- 2. See Section 6.5.6 below for a definition.
Note that these slides use the original 6850/1011989 frequency table, not the updated table from the supplement.
33
Fire Ignition Frequencies Plant Level Ignition Source Bins Table 6 -1 Fire Frequency Bins and Generic Frequencies Generic Split Fractions for Fire Type Ignition Source Freq ID Location Mode 1 (Equipment Type) (per rx yr) Electrical Oil Transient Hotwork Hydrogen HEAF 1 Battery Room Batteries All 7.5E-04 1.0 0 0 0 0 0 2 Containment (PWR) Reactor Coolant Pump Power 6.1E-03 0.14 0.86 0 0 0 0 4 Control Room Main Control Board All 2.5E-03 1.0 0 0 0 0 0 8 Diesel Generator Diesel Generators All 2.1E-02 0.16 0.84 0 0 0 0 Room Ignition Source 11 Plant-Wide ID Cable fires caused by Location Power 2.0E-03 0 0 0 1.0 0 0 Components welding and cutting (Equipment Type) 14 Plant-Wide Electric Motors All 4.6E-03 1.0 0 0 0 0 0 Components 1 Battery Room Batteries 15 Plant-Wide Electrical Cabinets All 4.5E-02 1.0 0 0 0 0 0 Components 2 Containment (PWR) Reactor Coolant Pumps 20 Plant-Wide Off-gas/H2 Recombiner Power 4.4E-02 0 0 0 0 1.0 0 Components 4 Control Room (BWR) Main Control Boards 27 Transformer Yard Transformer - Catastrophic2 Power 6.0E-03 1.0 0 0 0 0 8 Diesel Generator Room Diesel Generators 32 Turbine Building Main Feedwater Pumps Power 1.3E-02 0.11 0.89 0 0 0 0
- 1. See Appendix M for a description of high-energy arcing fault (HEAF) fires.
- 2. See Section 6.5.6 below for a definition.
Note that these slides use the original 6850/1011989 frequency table, not the updated table from the supplement.
34
Fire Ignition Frequencies Plant Level Ignition Source Bins Table 6 -1 Fire Frequency Bins and Generic Frequencies Generic Split Fractions for Fire Type Ignition Source Freq ID Location Mode 1 (Equipment Type) (per rx yr) Electrical Oil Transient Hotwork Hydrogen HEAF 1 Battery Room Batteries All 7.5E-04 1.0 0 0 0 0 0 2 Containment (PWR) Reactor Coolant Pump Power 6.1E-03 0.14 0.86 0 0 0 0 4 Control Room Main Control Board All 2.5E-03 1.0 0 0 0 0 0 8 Diesel Generator Diesel Generators All 2.1E-02 0.16 0.84 0 0 0 0 Room Generic Split Fractions for Fire Type Ignition 11 Plant-Wide Source fires caused Freq CableMode by Power 2.0E-03 0 0 0 1.0 0 0 (Equipment Components Type) welding and cutting (per rx yr) Electrical Oil Transient Hotwork Hydrogen HEAF1 14 Plant-Wide Electric Motors All 4.6E-03 1.0 0 0 0 0 0 Batteries Components All 7.5E-04 1.0 0 0 0 0 0 Reactor15Coolant Plant-Wide Components Pump Electrical Cabinets Power 6.1E-03 All 0.144.5E-02 0.861.0 0 0 0 0 0 0 0 0 0
Transients and Hotwork 20 Plant-Wide Power Off-gas/H2 2.0E-03 Recombiner Power 0 4.4E-02 0 0 0.44 0 0 0.560 0 1.0 0 0 Components (BWR)
Main Control Board All 2.5E-03 1.0 0 0 0 0 0 27 Transformer Yard Transformer - Catastrophic2 Power 6.0E-03 1.0 0 0 0 0 32 Turbine Building Main Feedwater Pumps Power 1.3E-02 0.11 0.89 0 0 0 0
- 1. See Appendix M for a description of high-energy arcing fault (HEAF) fires.
- 2. See Section 6.5.6 below for a definition.
Note that these slides use the original 6850/1011989 frequency table, not the updated table from the supplement.
35
Fire Ignition Frequencies Plant Level Ignition Source Bins 36
Fire Ignition Frequencies About the mode column For each plant, two time periods were established based on operating mode:
- Power operations (including low-power) - total years spent in power operation since initial commercial operation or reactor years (ry or rx yr)
- Shutdown operations - total time since initial commercial operation spent in non-power operating modes Some frequency bins cover all modes of operation, some only cover power operations (including low-power)
Both sets represent ignition frequencies per mode-year
- Some are simply power-modes-only and some are applicable to all modes Applied to at-power fire PRA the generic frequencies (in all references) are all per ry 37
Fire Ignition Frequencies And what about the standard IGN-A5 says (same for all CCs):
INCLUDE in the fire frequency calculation the plant availability, such that the frequencies are weighted by the fraction of time the plant is at-power.
This means that to meet IGN-A5, convert the generic frequencies:
FROM per reactor year (ry) TO per calendar year (cy)
You do that by multiplying the generic frequencies by the plant-specific average annual availability factor
- These are typically in excess of 90% for most plants today The risk results you get will then be per cy numbers (instead of per ry)
Why? Creates common basis for estimating total plant risk numbers that include the contribution from all sources including shutdown operations 38
Fire Ignition Frequencies Task 6 procedure Task 6 develops location and item specific fire frequency values for each fire frequency bin using an 8-step process:
Step 1. Mapping plant ignition sources to generic sources, Step 2. Plant fire event data collection and review, Step 3. Plant specific updates of generic ignition frequencies, Step 4. Mapping plant-specific locations to generic locations, Step 5. Location weighting factors, Step 6. Fixed fire ignition source counts, Step 7. Ignition source weighting factors, and Step 8. Ignition source and compartment (PAU) fire frequency evaluation.
Relevant PRA Standard Supporting Requirement: IGN-A7 39
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources In practice, Steps 1 and 6 are done together so well cover them together here
- Step 1 is ignition source mapping
- Step 6 is fixed ignition source counting Both are done by visual inspection In short, the process is:
- Perform a thorough walkdown of the plant (preferably with tablet in hand)
- Identify any and all potential fire ignitions sources
- Map each source to one of the 37 ignition source bins
- And, oh by the way, count them (Step 6)
- Keep a list with name, bin assignment, and location as you go
- If you want to get fancy, link the list to photos Now for the long version 40
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources In step 1 everything in the plant that is capable of starting a fire should be mapped to one of the 37 pre-defined ignition source bins.
Nominally covers all locations within the global analysis boundary, but EXCLUDE locations that screen out qualitatively
- Qualitative screening is Task 4 which is covered in Module 1
- Locations with no fire PRA equipment or cables and no plant trip initiators dont require quantitative analysis - they screen out on a qualitative basis Parking lots, office buildings, warehouses, security access buildings,
- The fire frequency classification excluded fire events in these types of locations Classified as non-challenging based on location not of interest to fire PRA
- For consistency do not apportion any fire frequency to such locations That means we dont count ignition sources in those locations Exclude them from the transient and hot work location sets as well (more later) 41
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources Steps 1 & 6 specifically focus on fixed equipment, but watch for transients and hot work activities as well
- There are several location based bins for transients and hot work
- You dont count these as sources per-se, but you do have to characterize likelihood and, for transients, the type expected in each plant location
- Make note of what you see and where you see it
- We will talk about the weighting factor approach later but you must use judgment to assign relative weighting factors to each PAU
- Insights gained during the plant walkdown can help this process
- Again, more on this later 42
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources If an ignition source does not map directly to one of the 37 available bins you have two options:
You may be able to match to an existing bin even if fit is not perfect
- Look for a bin with similar characteristics relative to fire likelihood e.g., a motor-driven widget may map to motors if the widget part is not significant
- Provide explanation for why you think fit is OK, again, relative to frequency Create a new bin for the item, but then you need a fire frequency
- A plant-specific history of fires may be enough to establish frequency Caution: a history of no fires at one plant probably wont be enough by itself
- Relevant experience at other plants may help
- Fire history in other industries may be used with caution
- You will be on the hook to quantify and justify your frequency assumptions One example we ran into: a gas-fired emergency generator unit 43
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources For each of the 37 ignition source bins there is corresponding description and counting guidance Well highlight a few bins, but for the others refer to the report:
- Bin 1 - Batteries (Battery Room): Each bank of interconnected sets of batteries located in one place should be counted as one battery set. Cells may not be counted individually.
- Bin 4 - Main Control Board (Control Room): A control room typically consists of one or two (depending on the number of units) main control boards as the central element of the room. The control room may also include plant computers, other electrical cabinets containing plant relays, and instrumentation circuits, a kitchen type area, desks, bookshelves, and etc. Aside from the main control board, the ignition source weighting factors of the remaining ignition sources of the control room should be based on the approach specific to each ignition source.
FAQ 06-018 (Supplement 1) - clarification of MCB definition (horseshoe or equivalent)
- There is a one-to-one correspondence between Appendix L and Bin 4
- All other electrical cabinets in MCR should be counted with Bin 15 FAQ 14-008 - updates the definition to include the rear side of the MCB 44
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources Counting guidance examples continued:
- Bin 15 - Electrical Cabinets (Plant-Wide Components): Electrical cabinets represent such items as switchgears, motor control centers, DC distribution panels, relay cabinets, control and switch panels (excluding panels that are part of machinery), fire protection panels, etc. The following rules should be used for counting electrical cabinets:
Simple wall-mounted panels housing less than four switches may be excluded from the counting process (these become junction boxes)
Well-sealed electrical cabinets that have robustly secured doors (and/or access panels) and that house only circuits below 440V should be excluded from the counting process Free-standing electrical cabinets should be counted by their vertical segments 45
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources Counting guidance examples continued:
- FAQ 06-0016 (Supplement 1) - Provides updated counting guidance for electrical cabinets
- Clarifies guidance on counting electrical cabinets and for treating outlier cabinets Cabinet counting guidance gets applied to a wide range of cabinet sizes Ignition frequency is more a function cabinet contents than cabinet size A basis is needed to address outlier conditions
- Each user should establish criteria for identifying outliers and a basis for counting them
- Examples of possible rule-set approaches:
Establish a nominal standard or reference cabinet size Consider cabinet internals relative to a defined standard or reference configuration 46
Fire Ignition Frequencies Steps 1&6 -FAQ 06-0016 Counting Example An analyst defines a standard cabinet as nominally 4 long and 3 deep and an outlier is any cabinet with any horizontal dimension greater than 8 6 long cabinet, Cabinet is not an outlier -
Count = 1 no partitions 4 long cabinet, Cabinet is same as standard -
Count = 1 no internal partitions Larger cabinet with Internal dividers are not solid -
Count = 6 non-solid internal partitions Larger cabinet with Internal dividers are solid -
Count = 6 solid internal partitions 47
Fire Ignition Frequencies Steps 1&6 -FAQ 06-0016 Counting Example (continued)
How to count using example rule set Three independent cabinets -
Count = 3 Panel is an outlier, using a 4 12 ft wide, 3 ft deep standard cabinet -
Count = 3 Cabinet is an outlier, no evaluation of 9 ft wide, 6 ft deep contents, based on reference cabinet -
Count = 3 due to variation from the standard length and depth The counts should depend on the 9 ft wide, 6 ft deep cable termination load and devices walk through cabinet in the panel by comparing it with a reference cabinet.
48
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources Some specific exclusions for certain bins
- Bin 14 - Electric motors: exclude small motors of 5 hp or less and totally enclosed motors.
- Bin 21 - Pumps: exclude small sampling pumps, and other pumps of 5 hp or less
- Bin 23 - Transformers: exclude dry transformers of 45 KVA or less
- Bin 26 - Ventilation subsystems: exclude small subsystems powered by motors of 5 hp or less (consistent with electric motors Bin 14)
FAQ 07-0031 (Supplement 1) provides clarification and extension beyond 6850 / 1011989 that are reflected in the bullets above 49
Fire Ignition Frequencies Steps 1&6 - FAQ 06-0017 Ignition source counting guidance for high energy arcing faults (HEAFs) in electrical switching equipment
- Issue: Originally all HEAF events were lumped in one ignition source bin (16) that were applied across all voltages 440V or greater.
However, cabinet voltage should impact fire frequency
- Resolution: FAQ 06-0017 Supplement 1 splits Bin 16 into 2 parts:
16.a Low-voltage panels (440 to 1,000V) - 1.52 E-04/ry (mean) 16.b Medium-voltage panels ( > 1000V) - 2.13E-03/ry (mean)
- This treatment carries forward for Supplement 1 & new fire ignition frequencies in EPRI 3002002936 / NUREG-2169
- Counting method remains unchanged (i.e., vertical sections) 50
Fire Ignition Frequencies Steps 1&6 - FAQ 07-0035 Guidance for counting HEAFs in bus ducts
- Issue: NUREG/CR-6850 / EPRI 1011989 was silent on this topic
- Resolution: FAQ 07-0035 Supplement 1 Acknowledge the potential for such events (e.g., Diablo Canyon 5/2000)
Provides plant wide frequency and counting/partitioning guidance Provides zone of influence and scenario development guidance Two categories of bus duct are defined:
- 16.1 Segmented Bus Duct
- 16.2 Iso-Phase Bus Duct Bins 16.1 and 16.2 in Supplement 1 Chapter 10 and new frequencies NUREG-2169 / EPRI 3002002936
- Beware the bin numbering issue - not 16a/16b
- FAQ 07-0035 did not list bin number with frequencies 51
Fire Ignition Frequencies Steps 1&6 - Mapping and Counting Ignition Sources Counting guidance examples continued:
- Some things we dont count at all, we use a relative weighting factor based on location characteristics (more on this later):
Hot work and transients
- Qualitative relative weighting factor by location Cables
- We will use a relative mass/volume weighting factor Junction boxes
- We apportion to a location using cable factors 52
Fire Ignition Frequencies FAQ 12-0064 This FAQ is unique and you should download a copy
- NRC ADAMS Accession Number: ML12346A488 Impacts some of the fire frequency mapping guidance
- Clarification of mapping of plant-specific locations to generic locations
- Modifies / expands the transient and hot work weighting factor methods Presented in the form of Chapter 6 redline / strikeout revisions
- This format is unique for Fire PRA FAQs, but necessary because the changes impact many parts of Chapter 6 We will cover this in more detail, but in general
- Clarification of where the Location = Plant Wide Components bin apply Short version: everywhere else not everywhere
- Revised (slightly expanded) ranking factors for hot work and transients 53
Fire Ignition Frequencies FAQ 12-0064 FAQ clarifies the intent of the location = plant-wide components bin
- These bins do not apportion to every location throughout the plant
- They are apportioned only to locations / components not explicitly covered by another corresponding ignition source bin Example 1:
- There are transient fire bins for the Turbine Building (Bin 37),
Containment (Bin 3), and for the Control/Aux/Reactor building complex (Bin 7)
- Bin 25 is the corresponding plant-wide components-transients bin
- Bin 25 applied to the rest of the plant, i.e. all locations except those mapped to bins 3,7 or 37 Example 2:
- Main feedwater pumps have their own bin (32), so
- Plant-wide components - pumps bin 21 excludes the MFW pumps
- Same goes for reactor coolant pumps (bin 2) 54
Fire Ignition Frequencies FAQ 12-0064 Other cases that also overlap:
- Main control board (bin 4) vs. electrical cabinets (bin 15)
- Battery chargers (bin 10) vs. electrical cabinets (bin 15)
- Specific yard transformer (bins 27-29) vs. general transformers bin (23a/b)
- The various location-based hot work fire bins Bottom line: In general only one bin contributes to the frequency of:
- Any given location for location-based bins (hot work and transients)
- Any given fixed fire ignition source
- Exceptions:
Some ignition source bins have multiple fire types that are reflected as split fractions in the table - one bin, multiple fire types Electrical cabinets have general thermal fires and high energy arc faults (HEAF)
- These were bins 15 and 16 originally,
- but were re-named in FAQ 48 as 15.1 and 15.2
- And finally in NUREG-2169 as 15 and 16.a and 16.b 55
Fire Ignition Frequencies Steps 1 & 6 - Concluding remarks At the end of the step 1/6 you will have:
- Been on several plant walkdowns
- Identified all fixed ignition sources and:
Mapped each to a location Mapped each to the generic fire ignition source bins Created a new bin if you found something truly unique Counted them
- Hopefully, you also made some observations on transients and hot work 56
Fire Ignition Frequencies Steps 1 & 6 - Concluding remarks Some hints:
- Every ignition source needs some sort of identifier so you can track them through the analysis
- Document so that you, or someone else, can tell what was counted and what name or identifier was assigned to each item - photos may be helpful
- Also document what was excluded from the count and why
- Counting is mainly about being consistent in application 57
Fire Ignition Frequencies Step 2 - Plant Fire Event Data Collection The generic fire frequencies are just that - generic If the plant under analysis has some unusual fire experience, then that should be reflected by updating the fire frequencies What constitutes unusual is a matter of judgment
- Every plant in the country has had some fires even though they may not have had reportable fires
- The question really is whether or not the plants experience is consistent with the rest of industry, example:
FAQ 07-0035 found roughly a dozen high energy arc faults in bus ducts, but Six had occurred at the same plant over the course of three years That constitutes unusual experience
- Note that it is not unusual for a plant to have experienced no fires in a given ignition source bin We have roughly 40 total bins and given plant-wide total fire frequency or 2E-1/ry, we would only expect about 6-10 fires over a nominal 40 year lifetime 58
Fire Ignition Frequencies Step 2 - Plant Fire Event Data Collection Common practice is to perform a Bayesian update of the generic fire frequencies to reflect plant-specific fire experience You need to gather plant-specific fire event data to establish plant-specific fire ignition frequencies
- Gather and review plant reports relating to fire events over some reasonable time period 10-15 years minimum, more if possible Look at the screening criteria and think about your event experience in the same context - are they risk relevant or not?
- Screening criteria NUREG/CR-6850 EPRI 1011989 Appendix C Potentially challenging screening criteria EPRI 1025284, updated challenging and potentially challenging criteria
- First question to ask is are plant specific fire ignition frequencies warranted?
Plant has experienced a repeated set of similar events Events that cannot be mapped to a bin Unusual fire occurrence patterns
- May be selective in which plant specific frequency bins are updated Not an all or nothing situation 59
Fire Ignition Frequencies Step 3 - Calculate Plant Specific Frequencies (IS)
The Bayesian update approach is the accepted method used to estimate plant-specific fire ignition frequencies
- PRA standard endorses/requires Bayesian methods in the SRs related to formal data analysis Youll find this in the Internal Events section (Part 2) rather than the fire section (Part 4)
Look for the DA technical element
- Generic frequency uncertainty distributions are used as the prior, plant specific data is used to do update Note that this approach does raise possible double-counting issue since same events identified in update may already be in the FEDB
- Generally not considered a significant issue, but be aware Corresponding PRA Standard SRs: IGN-A4, IGN-A6, and IGN-A10 60
Fire Ignition Frequencies Steps 2 & 3 - Illustrative example
- The following events have taken place at the unit under analysis over the past 10 years of plant operation:
Event 1: Fire in MCC-A because breakers were not properly engaging the bus bars.
Event 2: Fire in 125VAC-A panel. The fire was extinguished when 4kV bus-A was de-energized from the control room. Fire resulted from arcing of supply lead to one of the fittings connecting to a controller to the bus.
- Both fires can be included in the frequency analysis
- Both events would map to Electrical Cabinets - non HEAF Per NUREG/CR-6850 / 1011989 and EPRI 3002002936 / NUREG-2169 this is bin 15 EPRI 1019259 (Supplement 1 to NUREG/CR-6850) calls this bin 15.1
- 2 electrical cabinet fires in 10 years is high compared to generic frequency so an update would be appropriate
- Given 2 fires in 10 years Bayesian update would increase mean fire frequency from 0.024/ry to 0.084/ry 61
Fire Ignition Frequencies Step 4 - Mapping Plant-Specific Locations Not a major step, but plant-specific locations should be mapped to the locations defined by the ignition source bins
- Several ignition source bins are explicitly location based, especially hot work and transients
- You will need to define what constitutes the following locations for the plant being analyzed:
Battery rooms Turbine building Control/auxiliary/reactor buildings Control room Containment Transformer yard Everywhere else
- Names often dont match exactly - you have to match based on function 62
Fire Ignition Frequencies Step 5 - Location Weighting Factor (WL)
Recall our fire frequency equation from earlier:
WL is a weighting factor that only applied to multi-unit sites
- WL = number of units that share locations at the site
- If WL = 1, it has no effect at all and we simply drop it Takes a bit of common sense to apply, and how you use it depends a lot on what scope of the analysis is (i.e., one unit only versus all site units) and on the specific scenarios
- Dont just dive in and multiply all the frequencies by 2 because you are at a dual unit site
- Use of WL depends on how you are doing the analysis 63
Fire Ignition Frequencies Step 5 - Location Weighting Factor (WL)
Examples where WL > 1 applies even if we are analyzing one unit:
- Main control room abandonment:
Two units share one control room but we are only analyzing Unit 1 The likelihood of fire leading to MCR abandonment for Unit 1 must consider uncontrolled fires in Unit 2 MCR equipment It is all the same space, but each unit has its own stuff That means nominal fire frequency doubles (WL = 2)
- Catastrophic turbine generator fires in a shared turbine hall Same sort of issue - fires in sister unit may impact unit under analysis
- General shared equipment areas - e.g., pump room with pumps for both units This one is a bit tricky - done carefully you can use WL Alternative is to count only the one unit under analysis and use Step 7 approach - simply add the extra pumps to local count (but not plant-wide count) even though they are sister unit equipment (example coming up) 64
Fire Ignition Frequencies Step 5 - Location Weighting Factor (WL)
WL works best when we are analyzing all units of a multi-unit site Example: We are working a two-unit site and analyzing both units
- We can use (WL = 2) which effectively doubles the base frequencies for all ignition source bins
- Then we can simply identify and count ignition sources for both units and partition using Step 7 (next up)
- In terms of fire frequency, this eliminates the need to worry about what equipment goes with which unit A cabinet is a cabinet One caution:
- This works well when units are essentially the same and have shared areas
- Not good when units are very different styles of plant e.g., a BWR and PWR happen to share a site
- Not recommended when each unit is stand-alone (no shared locations) 65
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
Again, recall our frequency equation:
WJ,IS = weighting factor for ignition sources in a given scenario
- For fixed ignition sources it is calculated based on the local versus plant-wide population ratio for each type of fixed ignition source in the scenario
- For transient and hot work, there is a qualitative approach that weights locations relative to each other
- We calculate one or more weighting factors for each fire scenario we analyze Generally one per ignition source bin represented in the scenario
- Corresponding PRA standard SRs: IGN-A7 & IGN-A9 66
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
Scenarios can involve anything from a building to a single piece of equipment - whatever sources contribute to the fire scenario
- Example 1:
At an early stage we postulate room-wide burn-out screening scenarios We would include all of the ignition sources in the room We calculate an ignition source weighting factor for each unique ignition source bin represented in the room Scenario fire frequency is sum of all these contributing fire sources
- Example 2:
Scenario is a bank of electrical cabinets threatening overhead cables All the cabinets have the same fire characteristics and same potential to damage the cables Well only need one location weighting factor to represent the bank of cabinets since all are members of the same ignition source bin (15) 67
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
The general process for fixed ignition sources:
- We start from the total number of relevant items in each bin plant-wide Recall that we counted these in step 6 Caution: be consistent relative to WL if you are at a shared site
- We count the ignition sources that contribute to the scenario of interest
- The scenario-specific ignition source weighting factor is then the simple ratio of the local count to the plant-wide count
- Example: estimate the fire frequency for PAU J:
PAU J contains 2 pumps assigned to Bin 21 We counted 50 Bin 21 pumps plant-wide WJ,21 = 2/50 = 0.04 21 = 2.7E-2/ry (from 3002002936 / NUREG-2169)
J,21 = 0.04 x 2.7E-2/ry = 1.08E-3/ry Repeat for each ignition source bin represented in PAU J 68
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
Single unit plant Count Room 1A Room 1B Total Elec. Cab. 2 2 Pump 2 2 Room 1A Room 1B 1
pmp i = g Wis WL = g 1 2
room 1B = pmp i N pmp = pmp i 2 69
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
Two units, two units in scope Room 1A Room 1B Room 2B Room 2A Count 1A 1B 2A 2B Total Elec. Cab. 2 2 4 Pump 2 2 4 1
pmp i = g Wis WL = g 2 room 1B = pmp i N pmp 1B = pmp i 2 4
70
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
Two Units, Two Units in scope, shared room Room 1C Room 1A Room B Room 2A Room 2C Count 1A 1C 2A 2C B Total Elec. Cab. 2 2 4 Pump 3 3 4 10 1 room B = pmp i N pmp B = pmp i 4 pmp i = g Wis WL = g 2 10 71
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
Two units, two units in scope, shared room, swing pump Room 1C Room 1A Room B Room 2A Room 2C Count 1A 1C 2A 2C B Total Elec. Cab. 2 2 4 Pump 2 2 3 7 1
pmp i = g Wis WL = g 2 room B = pmp i N pmp B = pmp i 3 7
72
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
Two units, one unit in scope, shared room Room 1C Room 1A Room B Room 2A Room 2C Count 1A 1C 2A 2C B Total Elec. Cab. 2 2 Pump 2 2 4 1 room B = pmp i N pmp B = pmp i 3 pmp i = g Wis WL = g 1 4
73
Fire Ignition Frequencies Step 7 - Ignition source weighting factors (WJ,IS)
Two units, one in scope, shared room, swing pump Room 1C Room 1A Room B Room 2A Room 2C Count 1A 1C 2A 2C B Total Elec. Cab. 2 2 Pump 2 1.5 3.5 1 room B = pmp i N pmp B = pmp i 3 pmp i = g Wis WL = g 1 3 .5 74
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients and hot work fires There is a new FAQ out that modifies the approach for non-fixed ignition sources
- Transient fires
- Transient fires caused by hot work
- Cable fires caused by hot work Reference is FAQ 12-0064 (ML12346A488)
- Issued January 2013 We will present the methods as modified by this FAQ
- The set of ranking factor values has been expanded
- Fractional values now allow more credit for very strict administrative controls in a location Application of the fractional values generally requires a review of plant records to verify that controls have not been violated 75
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients and hot work fires All three fire types are handled in a similar manner, but factors will vary General approach:
- The PAUs that make up the location set (L) for each of the transient and hot work bins are defined (see slide 32)
- The PAUs that make up a location set are assigned a numerical ranking value based on certain characteristics (covered shortly)
- The PAUs within a set are ranked relative to each other
- The ranking values for all PAUs in the set are summed to get a total for the set (this becomes the normalizing factor for the set)
- The fraction of the fire frequency assigned to a given location (J) is calculated based on ratio of ranking for location to set total:
WIS,J,L = (rank for the location J) / (total for all locations in the set L) 76
Fire Ignition Frequencies Step 7: WIS,J,L - Transients Transient fire bins are 3, 7, 25 and 37 General transients covers all the fires that didnt fit in the other frequency bins and that were not associated with hot work
- Trash
- Liquids not actually inside a component (e.g. oil, solvents)
- Portable heaters
- Portable lighting
- Stored materials
- Staged materials and associated packing materials
- Temporary computers or instruments
- Rad protection dress-up areas
- Temporary structures inside plant 77
Fire Ignition Frequencies Step 7: WIS,J,L - Transients Transient fire frequencies are apportioned based on qualitatively estimated ranking level for three factors:
- General electro/mechanical maintenance Excludes hot work
- Occupancy level and traffic density Implication is that people bring stuff with them
- Storage (temporary and permanent)
Combustible and flammable materials Includes liquids Staging area 78
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients There are five ranking levels that applied to all three ranking factors:
- No (0) - Can be used only for those PAUs where transients are precluded by design Administrative restrictions do not qualify for a 0 ranking*
- Very low (0.3) - applies only to locations with the strictest levels of administrative control - strict prohibitions in force and verified effective (no violations)
- Low (1) - Reflects minimal level of the factor
- Medium (3) - Reflects average level of the factor
- High (10) - Reflects the higher-than-average level of the factor plus one rating level that applied to general maintenance factor only:
- Very high (50) - Reflects the significantly higher-than-average level of the factor (only for maintenance influencing factor).
- Corresponding PRA Standard SR: IGN-A9 79
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients - Table 6-3 (FAQ 12-0064) 80
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients - Table 6-3 (FAQ 12-0064)
81
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients A ranking value for each factor is assigned to each PAU RankingJ,L { ngenmaint,J,L , noccup,J,L , nstorage,J,L }
The total for the room is the simple sum of the three assigned values:
NGT,J,L = (ngenmaint,J,L + noccup,J,L + nstorage,J,L)
The total for the location set is the sum of the values for each PAU:
NGT,L = NGT,J,L (summed over J, for all PAUs included in location set L)
The PAUJ general transients weighting factor is the ratio of these two values:
WGT,J,L = NGT,J,L / NGT,L 82
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients caused by hot work This is for bins 3, 6, 24, and 36 Process is similar but there is only one ranking factor:
- Hot work maintenance activities - welding and cutting Ranking values for NWC,J,L are:
- No (0) - Can be used only for those PAUs where hot work is precluded by design Administrative restrictions do not qualify for a 0 ranking *
- Extremely low (0.1) - only allowed in MCR (requires verification)
- Very low (0.3) - only applies two places:
Cable spreading room (CSR) (required verification)
MCR if it did not qualify for 0.1 ranking
- Low (1) - Reflects minimal level hot work
- Medium (3) - Reflects average level of hot work
- High (10) - Reflects the higher-than-average level of hot work
- Very high (50) - Reflects the significantly higher-than-average level of hot work
- See IGN-A8 - assign an ignition frequency greater than zero to every PAU (CC-I/II) or PAU and risk relevant ignition source (CC-III) 83
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients caused by hot work 84
Fire Ignition Frequencies Step 7 - WIS,J,L - Transients caused by hot work Each PAU (J) within a location set (L) gets a hot work ranking factor:
PAU weighting factor -> NWC,J,L Normalizing factor for the location set (L) as a whole:
NWC,L = NWC,J,L (summed over J, for all PAUs included in location set L)
The PAUJ transients caused by welding and cutting weighting factor is the ratio of the PAU to the location set:
WGT,J,L = NWC,J,L / NWC,L 85
Fire Ignition Frequencies Step 7 - WIS,J,L - Cables fires caused by hot work Cable fires caused by hot work is similar, but adds a factor based on the relative cable mass/volume in each PAU (J) of the location set (L)
Wcable,J,L = Ratio of cable load in PAU (J) over the total cable load for all members of location set (L)
- Check your fire hazards analysis (FHA combustible fuel loads)
Use the hot work weighting factors you already have (NWC,J,L)
The normalizing factor is:
NHWCF,L = Wcable,J,L x NWC,J,L (sum over all PAUs in the location set)
The weighting factor for PAU J is:
WHWCF,J,L = [ Wcable,J,L x NWC,J,L ] / NHWCF,L 86
Fire Ignition Frequencies Step 7 - Final notes on weighting factors Weighting factors are always relative within each frequency bin / location set
- Each ignition source bin where we apply weighting factors defines a location set
- DO NOT weigh across bins, DO NOT weight across location sets:
For transients in the turbine building (Bin 37), weigh locations in the turbine building against each other For transients in the Aux/Control/Reactor building complex (Bin 7), weigh locations in that complex against each other Do NOT compare the turbine building to the control building. That comparison is built into the base frequencies 87
Fire Ignition Frequencies Step 7 - Final notes on weighting factors A ranking of 3 is considered Normal/Average/Typical
- Decide what is typical for the location set overall - just the one set
- Define that condition as 3 in your ranking scheme
- Everything else is up or down from the typical condition
- You do not need to average the rankings for a set and show the average is 3
The method is designed to reflect real differences in the likelihood of these kinds of fires in different locations
- You need to exercise the full range of ranking values to take full advantage of the method
- Otherwise, frequency for each bin will be distributed evenly to each PAU 88
Fire Ignition Frequencies Step 8 - Fire Frequency Evaluation The fire frequency (generic or plant-specific) for each ignition source, IS,J, can now be calculated using the data quantified in the preceding steps J,L = IS WL WIS,J,L summed over all ignition sources Where:
J,L : Fire frequency associated with PAU J at location L IS: Plant level fire ignition frequency associated with ignition source IS WL: Location weighting factor WIS,J,L: Ignition source weighting factor Corresponding PRA Standard: IGN-A7 89
Fire Ignition Frequencies Concluding remarks Fire ignition frequency evaluation (Task 6) uses a mix of plant specific and generic information to establish the ignition frequencies for specific fire compartments or PAUs and from that for specific fire scenarios.
- Generic fire ignition frequencies based on industry experience
- Elaborate data analysis method
- Frequencies binned by equipment type
- Methodology to apportion frequencies according to relative characteristics of each fire compartment or PAU 90
Mapping HLRs & SRs for the IGN Technical Element to NUREG/CR-6850 / EPRI 1011989 Technical HLR SR 6850 sections Comments element IGN A The Fire PRA shall develop fire ignition frequencies for every physical analysis unit that has not been qualitatively screened.
1 Appendix C The generic frequencies have been modified in EPRI 1019259 to reflect changes in fire event frequency trends.
The methodology used in that study is also consistent with this SR.
2 6.5.1 3 n/a Using engineering judgment to establish a frequency is not addressed in 6850/1011989.
4 6.5.2, 6.5.3 5 6.5.3 and Appendix C The generic frequencies of EPRI 1019259 are also consistent with this SR.
6 6.5.3 7 6.5.1, 6.5.4, 6.5.5, 6.5.6, 6.5.7 8 n/a Although it is effectively implied in Section 6.5.7.2, this SR is not explicitly discussed in 6850/1011989.
9 6.5.7 Inherent in transient weighting factor ranking approach 10 6.5.3, Appendix C Generic frequencies consistent with this SR B The Fire PRA shall document the fire frequency estimation in a manner that facilitates Fire PRA applications, upgrades, and peer review.
1 n/a Documentation is covered in minimal detail in 2 n/a 6850/1011989 3 n/a 4 n/a 5 n/a 91
TogetherShaping the Future of Electricity 92