ML022410004
| ML022410004 | |
| Person / Time | |
|---|---|
| Issue date: | 08/28/2002 |
| From: | Beckner W NRC/NRR/DRIP/RORP |
| To: | |
| Sanders S | |
| References | |
| OMB 3150-0195 RIS-02-014 | |
| Download: ML022410004 (37) | |
See also: RIS 2002-14
Text
Attachment 1
Attachment 1, Section 2.2, Mitigating Systems Cornerstone, of NEI 99-02, Regulatory
Assessment Performance Indicator Guideline (Draft)
1
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
1
MITIGATING SYSTEM PERFORMANCE INDEX
2
Purpose
3
The purpose of the mitigating system performance index is to monitor the performance of
4
selected systems based on their ability to perform risk-significant functions as defined herein. It
5
is comprised of two elements - system unavailability and system unreliability. The index is used
6
to determine the significance of performance issues for single demand failures and accumulated
7
unavailability. Due to the limitations of the index, the following conditions will rely upon the
8
inspection process for determining the significance of performance issues:
9
10
1. Multiple concurrent failures of components
11
2. Common cause failures
12
3. Conditions not capable of being discovered during normal surveillance tests
13
4. Failures of non-active components
14
15
Indicator Definition
16
Mitigating System Performance Index (MSPI) is the sum of changes in a simplified core damage
17
frequency evaluation resulting from changes in unavailability and unreliability relative to
18
baseline values.
19
20
Unavailability is the ratio of the hours the train/system was unavailable to perform its risk-
21
significant functions due to planned and unplanned maintenance or test on active and non-active
22
components during the previous 12 quarters while critical to the number of critical hours during
23
the previous 12 quarters. (Fault exposure hours are not included; unavailable hours are counted
24
only for the time required to recover the trains risk-significant functions.)
25
26
Unreliability is the probability that the system would not perform its risk-significant functions
27
when called upon during the previous 12 quarters.
28
29
Baseline values are the values for unavailability and unreliability against which current changes
30
in unavailability and unreliability are measured. See Appendix F for further details.
31
32
The MSPI is calculated separately for each of the following five systems for each reactor type.
33
34
35
emergency AC power system
36
high pressure injection systems (high pressure coolant injection, high pressure core spray, or
37
feedwater coolant injection)
38
heat removal systems (reactor core isolation cooling)
39
residual heat removal system (or their equivalent function as described in the Additional
40
Guidance for Specific Systems section.)
41
2
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
cooling water support system (includes risk significant direct cooling functions provided by
1
service water and component cooling water or their cooling water equivalents for the above
2
four monitored systems)
3
4
5
emergency AC power system
6
high pressure safety injection system
7
auxiliary feedwater system
8
residual heat removal system (or their equivalent function as described in the Additional
9
Guidance for Specific Systems section.)
10
cooling water support system (includes risk significant direct cooling functions provided by
11
service water and component cooling water or their cooling water equivalents for the above
12
four monitored systems)
13
14
Data Reporting Elements
15
The following data elements are reported for each system
16
17
Unavailability Index (UAI) due to unavailability for each monitored system
18
Unreliability Index (URI) due to unreliability for each monitored system
19
20
During the pilot, the additional data elements necessary to calculate UAI and URI will be
21
reported monthly for each system on an Excel spreadsheet. See Appendix F.
22
23
24
Calculation
25
The MSPI for each system is the sum of the UAI due to unavailability for the system plus URI
26
due to unreliability for the system during the previous twelve quarters.
27
28
29
30
See Appendix F for the calculational methodology for UAI due to system unavailability and URI
31
due to system unreliability.
32
33
Definition of Terms
34
A train consists of a group of components that together provide the risk significant functions of
35
the system as explained in the additional guidance for specific mitigating systems. Fulfilling the
36
risk-significant function of the system may require one or more trains of a system to operate
37
simultaneously. The number of trains in a system is generally determined as follows:
38
39
for systems that provide cooling of fluids, the number of trains is determined by the number
40
of parallel heat exchangers, or the number of parallel pumps, or the minimum number of
41
parallel flow paths, whichever is fewer.
42
43
3
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
for emergency AC power systems the number of trains is the number of class 1E emergency
1
(diesel, gas turbine, or hydroelectric) generators at the station that are installed to power
2
shutdown loads in the event of a loss of off-site power. (This does not include the diesel
3
generator dedicated to the BWR HPCS system, which is included in the scope of the HPCS
4
system.)
5
6
Risk Significant Functions: those at power functions, described in the Additional Guidance for
7
Specific Systems, that were determined to be risk-significant in accordance with NUMARC 93-
8
01, or NRC approved equivalents (e.g., the STP exemption request.) The system functions
9
described in the Additional Guidance for Specific Systems must be modeled in the plants
10
PRA/PSA. of risk-significant SSCs as modeled in the plant-specific PRA. Risk metrics for
11
identifying risk-significant functions are:
12
13
Risk Achievement Worth > 2.0, or
14
Risk Reduction Worth >0.005, or
15
PRA cutsets that account for 90% of core damage frequency90% of core damage
16
frequency accounted for.
17
18
Risk-Significant Mission Times: The mission time modeled in the PRA for satisfying the risk-
19
significant function of reaching a stable plant condition where normal shutdown cooling is
20
sufficient. Note that PRA models typically analyze an event for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, which may exceed the
21
time needed for the risk-significant function captured in the MSPI. However, other intervals as
22
justified by analyses and modeled in the PRA may be used.
23
24
Success criteria are the plant specific values of parameters the train/system is required to achieve
25
to perform its risk-significant function. Default values of those parameters are the plants design
26
bases values unless other values are modeled in the PRA.
27
28
Clarifying Notes
29
Documentation
30
31
Each licensee will have the system boundaries, active components, risk-significant functions and
32
success criteria readily available for NRC inspection on site. Additionally, plant-specific
33
information used in Appendix F should also be readily available for inspection.
34
35
Success Criteria
36
37
Individual component capability must be evaluated against train/system level success criteria
38
(e.g., a valve stroke time may exceed an ASME requirement, but if the valve still strokes in time
39
to meet the PRA success criteria for the train/system, the component has not failed for the
40
purposes of this indicator because the risk-significant train/system function is still satisfied).
41
Important plant specific performance factors that can be used to identify the required capability
42
of the train/system to meet the risk-significant functions include, but are not limited to:
43
Actuation
44
o Time
45
4
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
o Auto/manual
1
o Multiple or sequential
2
Success requirements
3
o Numbers of components or trains
4
o Flows
5
o Pressures
6
o Heat exchange rates
7
o Temperatures
8
o Tank water level
9
Other mission requirements
10
o Run time
11
o State/configuration changes during mission
12
Accident environment from internal events
13
o Pressure, temperature, humidity
14
Operational factors
15
o Procedures
16
o Human actions
17
o Training
18
o Available externalities (e.g., power supplies, special equipment, etc.)
19
20
21
22
System/Component Interface Boundaries
23
24
For active components that are supported by other components from both monitored and
25
unmonitored systems, the following general rules apply:
26
27
For control and motive power, only the last relay, breaker or contactor necessary to
28
power or control the component is included in the active component boundary. For
29
example, if an ESFAS signal actuates a MOV, only the relay that receives the ESFAS
30
signal in the control circuitry for the MOV is in the MOV boundary. No other portions
31
of the ESFAS are included.
32
33
For water connections from systems that provide cooling water to an active component,
34
only the final active connecting valve is included in the boundary. For example, for
35
service water that provides cooling to support an AFW pump, only the final active valve
36
in the service water system that supplies the cooling water to the AFW system is
37
included in the AFW system scope. This same valve is not included in the cooling water
38
support system scope.
39
40
Water Sources and Inventory
41
42
Water tanks are not considered to be active components. As such, they do not contribute to URI.
43
However, periods of insufficient water inventory contribute to UAI if they result in loss of the
44
risk-significant train function for the required mission time. Water inventory can include
45
operator recovery actions for water make-up provided the actions can be taken in time to meet
46
5
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
the mission times and are modeled in the PRA. If additional water sources are required to satisfy
1
train mission times, only the connecting active valve from the additional water source is
2
considered as an active component for calculating URI. If there are valves in the primary water
3
source that must change state to permit use of the additional water source, these valves are
4
considered active and should be included in URI for the system.
5
6
Monitored Systems
7
8
Systems have been generically selected for this indicator based on their importance in preventing
9
reactor core damage. The systems include the principal systems needed for maintaining reactor
10
coolant inventory following a loss of coolant accident, for decay heat removal following a
11
reactor trip or loss of main feedwater, and for providing emergency AC power following a loss
12
of plant off-site power. One risk-significant support function (cooling water support system) is
13
also monitored. The cooling water support system monitors the risk significant cooling functions
14
provided by service water and component cooling water, or their direct cooling water
15
equivalents, for the four front-line monitored systems. No support systems are to be cascaded
16
onto the monitored systems, e.g., HVAC room coolers, DC power, instrument air, etc.
17
18
Diverse Systems
19
20
Except as specifically stated in the indicator definition and reporting guidance, no credit is given
21
for the achievement of a risk-significant function by an unmonitored system in determining
22
unavailability or unreliability of the monitored systems.
23
24
Common Components
25
26
Some components in a system may be common to more than one train or system, in which case
27
the unavailability/unreliability of a common component is included in all affected trains or
28
systems. (However, see Additional Guidance for Specific Systems for exceptions; for example,
29
the PWR High Pressure Safety Injection System.)
30
31
Short Duration Unavailability
32
33
Trains are generally considered to be available during periodic system or equipment
34
realignments to swap components or flow paths as part of normal operations. Evolutions or
35
surveillance tests that result in less than 15 minutes of unavailable hours per train at a time need
36
not be counted as unavailable hours. Licensees should compile a list of surveillances/evolutions
37
that meet this criterion and have it available for inspector review. In addition, equipment
38
misalignment or mispositioning which is corrected in less than 15 minutes need not be counted
39
as unavailable hours. The intent is to minimize unnecessary burden of data collection,
40
documentation, and verification because these short durations have insignificant risk impact.
41
42
If a licensee is required to take a component out of service for evaluation and corrective actions
43
for greater than 15 minutes (for example, related to a Part 21 Notification), the unavailable hours
44
must be included.
45
46
6
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
Treatment of Demand /Run Failures and Degraded Conditions
1
2
1. Treatment of Demand and Run Failures
3
Failures of active components (see Appendix F) on demand or failures to run, either
4
actual or test, while critical, are included in unreliability. Failures on demand or failures
5
to run at any other timewith the reactor shutdown must be evaluated to determine if the
6
failure would have resulted in the train not being able to perform its risk-significant at
7
power functions, and must therefore be included in unreliability. Unavailable hours are
8
included only for the time required to recover the trains risk-significant functions and
9
only when the reactor is critical.
10
11
2. Treatment of Degraded Conditions
12
13
a) Capable of Being Discovered By Normal Surveillance Tests
14
Normal surveillance tests are those tests that are performed at a frequency of a
15
refueling cycle or more frequently.
16
17
Degraded conditions, even ifwhere no actual demand existed, that render an
18
active component incapable of performing its risk-significant functions are
19
included in unreliability as a demand and a failure. The appropriate failure mode
20
must be accounted for. For example, for valves, a demand and a demand failure
21
would be assumed and included in URI. For pumps and diesels, if the degraded
22
condition would have prevented a successful start demand, a demand and a failure
23
is included in URI, but there would be no run time hours or run failures. If it was
24
determined that the pump/diesel would start and load run, but would fail
25
sometime during the 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> run test or its surveillance test equivalent, the
26
evaluated failure time would be included in run hours and a run failure would be
27
assumed. A start demand and start failure would not be included. If a running
28
component is secured from operation due to observed degraded performance, but
29
prior to failure, then a run failure shall be counted unless evaluation of the
30
condition shows that the component would have continued to operate for the risk-
31
significant mission time starting from the time the component was secured.
32
Unavailable hours are included for the time required to recover the risk-
33
significant function(s).
34
35
Degraded conditions, or actual unavailability due to mispositioning of non-active
36
components that render a train incapable of performing its risk-significant
37
functions are only included in unavailability for the time required to recover the
38
risk-significant function(s).
39
40
Loss of risk significant function(s) is assumed to have occurred if the established
41
success criteria has not been met. If subsequent analysis identifies additional
42
margin for the success criterion, future impacts on URI or UAI for degraded
43
conditions may be determined based on the new criterion. However, URI and
44
UAI must be based on the success criteria of record at the time the degraded
45
condition is discovered. If the degraded condition is not addressed by any of the
46
7
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
pre-defined success criteria, an engineering evaluation to determine the impact of
1
the degraded condition on the risk-significant function(s) should be completed
2
and documented. The use of component failure analysis, circuit analysis, or event
3
investigations is acceptable. Engineering judgment may be used in conjunction
4
with analytical techniques to determine the impact of the degraded condition on
5
the risk-significant function. The engineering evaluation should be completed as
6
soon as practicable. If it cannot be completed in time to support submission of the
7
PI report for the current quarter, the comment field shall note that an evaluation is
8
pending. The evaluation must be completed in time to accurately account for
9
unavailability/unreliability in the next quarterly report. Exceptions to this
10
guidance are expected to be rare and will be treated on a case-by-case basis.
11
Licensees should identify these situations to the resident inspector.
12
13
b) Not Capable of Being Discovered by Normal Surveillance Tests
14
These failures or conditions are usually of longer exposure time. Since these
15
failure modes have not been tested on a regular basis, it is inappropriate to include
16
them in the performance index statistics. These failures or conditions are subject
17
to evaluation through the inspection process. Examples of this type are failures
18
due to pressure locking/thermal binding of isolation valves, blockages in lines not
19
regularly tested, or inadequate component sizing/settings under accident
20
conditions (not under normal test conditions). While not included in the
21
calculation of the index, they should be reported in the comment field of the PI
22
data submittal.
23
24
25
Credit for Operator Recovery Actions to Restore the Risk-Significant Function
26
27
1. During testing or operational alignment:
28
Unavailability of a risk-significant function during testing or operational alignment need not
29
be included if the test configuration is automatically overridden by a valid starting signal, or
30
the function can be promptly restored either by an operator in the control room or by a
31
designated operator1 stationed locally for that purpose. Restoration actions must be
32
contained in a written procedure2, must be uncomplicated (a single action or a few simple
33
actions), must be capable of being restored in time to satisfy PRA success criteria and must
34
not require diagnosis or repair. Credit for a designated local operator can be taken only if
35
(s)he is positioned at the proper location throughout the duration of the test for the purpose of
36
restoration of the train should a valid demand occur. The intent of this paragraph is to allow
37
licensees to take credit for restoration actions that are virtually certain to be successful (i.e.,
38
probability nearly equal to 1) during accident conditions.
39
40
1 Operator in this circumstance refers to any plant personnel qualified and designated to perform
the restoration function.
2 Including restoration steps in an approved test procedure.
8
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
The individual performing the restoration function can be the person conducting the test and
1
must be in communication with the control room. Credit can also be taken for an operator in
2
the main control room provided (s)he is in close proximity to restore the equipment when
3
needed. Normal staffing for the test may satisfy the requirement for a dedicated operator,
4
depending on work assignments. In all cases, the staffing must be considered in advance and
5
an operator identified to perform the restoration actions independent of other control room
6
actions that may be required.
7
8
Under stressful, chaotic conditions, otherwise simple multiple actions may not be
9
accomplished with the virtual certainty called for by the guidance (e.g., lifting test leads and
10
landing wires; or clearing tags). In addition, some manual operations of systems designed to
11
operate automatically, such as manually controlling HPCI turbine to establish and control
12
injection flow, are not virtually certain to be successful. These situations should be resolved
13
on a case-by-case basis through the FAQ process.
14
15
2. During Maintenance
16
Unavailability of a risk-significant function during maintenance need not be included if the
17
risk-significant function can be promptly restored either by an operator in the control room or
18
by a designated operator3 stationed locally for that purpose. Restoration actions must be
19
contained in a written procedure4, must be uncomplicated (a single action or a few simple
20
actions), must be capable of being restored in time to satisfy PRA success criteria and must
21
not require diagnosis or repair. Credit for a designated local operator can be taken only if
22
(s)he is positioned at a proper location throughout the duration of the maintenance activity
23
for the purpose of restoration of the train should a valid demand occur. The intent of this
24
paragraph is to allow licensees to take credit for restoration of risk-significant functions that
25
are virtually certain to be successful (i.e., probability nearly equal to 1). The individual
26
performing the restoration function can be the person performing the maintenance and must
27
be in communication with the control room. Credit can also be taken for an operator in the
28
main control room provided (s)he is in close proximity to restore the equipment when
29
needed. Under stressful chaotic conditions otherwise simple multiple actions may not be
30
accomplished with the virtual certainty called for by the guidance (e.g., lifting test leads and
31
landing wires, or clearing tags). These situations should be resolved on a case-by-case basis
32
through the FAQ process.
33
34
3. Satisfying PRA success criteriaRisk Significant Mission Times
35
Risk significant operator actions to satisfy pre-determined train/system risk-significant
36
mission times can only be credited if they are modeled in the PRA.
37
38
Swing trains and components shared between units
39
40
3 Operator in this circumstance refers to any plant personnel qualified and designated to perform the
restoration function.
4 Including restoration steps in an approved test procedure.
9
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
Swing trains/components are trains/components that can be aligned to any unit. To be credited
1
as such, their swing capability should be modeled in the PRA to provide an appropriate Fussell-
2
Vesely value.
3
4
Unit Cross Tie Capability
5
6
Components that cross tie monitored systems between units should be considered active
7
components if they are modeled in the PRA and meet the active component criteria in Appendix
8
F. Such active components are counted in each units performance indicators.
9
10
Maintenance Trains and Installed Spares
11
12
Some power plants have systems with extra trains to allow preventive maintenance to be carried
13
out with the unit at power without impacting the risk-significant function of the system. That is,
14
one of the remaining trains may fail, but the system can still perform its risk significant function.
15
To be a maintenance train, a train must not be needed to perform the systems risk significant
16
function.
17
18
An "installed spare" is a component (or set of components) that is used as a replacement for other
19
equipment to allow for the removal of equipment from service for preventive or corrective
20
maintenance without impacting the risk-significant function of the system. To be an "installed
21
spare," a component must not be needed for the system to perform the risk significant function.
22
23
24
For unreliability, spare active components are included if they are modeled in the PRA.
25
Unavailability of the spare component/train is only counted in the index if the spare is substituted
26
for a primary train/component. Unavailability is not monitored for a component/train when that
27
component/train has been replaced by an installed spare or maintenance train.
28
29
Use of Plant-Specific PRA and SPAR Models
30
31
The MSPI is an approximation using some information from a plants actual PRA and is
32
intended as an indicator of system performance. Plant-specific PRAs and SPAR models cannot
33
be used to question the outcome of the PIs computed in accordance with this guideline.
34
35
Maintenance Rule Performance Monitoring
36
37
It is the intent that NUMARC 93-01 be revised to require consistent unavailability and
38
unreliability data gathering as required by this guideline.
39
40
10
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
ADDITIONAL GUIDANCE FOR SPECIFIC SYSTEMS
1
This guidance provides typical system scopes. Individual plants should include those systems
2
employed at their plant that are necessary to satisfy the specific risk-significant functions
3
described below and reflected in their PRAs.
4
Emergency AC Power Systems
5
Scope
6
The function monitored for the emergency AC power system is the ability of the emergency
7
generators to provide AC power to the class 1E buses upon a loss of off-site power while the
8
reactor is critical, including post-accident conditions. The emergency AC power system is
9
typically comprised of two or more independent emergency generators that provide AC power to
10
class 1E buses following a loss of off-site power. The emergency generator dedicated to
11
providing AC power to the high pressure core spray system in BWRs is not within the scope of
12
emergency AC power.
13
14
The electrical circuit breaker(s) that connect(s) an emergency generator to the class lE buses that
15
are normally served by that emergency generator are considered to be part of the emergency
16
generator train.
17
18
Emergency generators that are not safety grade, or that serve a backup role only (e.g., an
19
alternate AC power source), are not included in the performance reporting.
20
21
Train Determination
22
The number of emergency AC power system trains for a unit is equal to the number of class 1E
23
emergency generators that are available to power safe-shutdown loads in the event of a loss of
24
off-site power for that unit. There are three typical configurations for EDGs at a multi-unit
25
station:
26
27
1. EDGs dedicated to only one unit.
28
2. One or more EDGs are available to swing to either unit
29
3. All EDGs can supply all units
30
31
For configuration 1, the number of trains for a unit is equal to the number of EDGs dedicated to
32
the unit. For configuration 2, the number of trains for a unit is equal to the number of dedicated
33
EDGs for that unit plus the number of swing EDGs available to that unit (i.e., The swing
34
EDGs are included in the train count for each unit). For configuration 3, the number of trains is
35
equal to the number of EDGs.
36
37
Clarifying Notes
38
The emergency diesel generators are not considered to be available during the following portions
39
of periodic surveillance tests unless recovery from the test configuration during accident
40
conditions is virtually certain, as described in Credit for operator recovery actions during
41
11
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
testing, can be satisfied; or the duration of the condition is less than fifteen minutes per train at
1
one time:
2
3
Load-run testing
4
Barring
5
6
An EDG is not considered to have failed due to any of the following events:
7
8
spurious operation of a trip that would be bypassed in a loss of offsite power event
9
malfunction of equipment that is not required to operate during a loss of offsite power event
10
(e.g., circuitry used to synchronize the EDG with off-site power sources)
11
failure to start because a redundant portion of the starting system was intentionally disabled
12
for test purposes, if followed by a successful start with the starting system in its normal
13
alignment
14
Air compressors are not part of the EDG boundary. However, air receivers that provide starting
15
air for the diesel are included in the EDG boundary.
16
17
If an EDG has a dedicated battery independent of the stations normal DC distribution system,
18
the dedicated battery is included in the EDG system boundary.
19
20
If the EDG day tank is not sufficient to meet the EDG mission time, the fuel transfer function
21
should be modeled in the PRA. However, the fuel transfer pumps are not considered to be an
22
active component in the EDG system because they are considered to be a support system.
23
24
25
26
BWR High Pressure Injection Systems
27
(High Pressure Coolant Injection, High Pressure Core Spray, and Feedwater Coolant
28
Injection)
29
30
Scope
31
These systems function at high pressure to maintain reactor coolant inventory and to remove
32
decay heat following a small-break Loss of Coolant Accident (LOCA) event or a loss of main
33
feedwater event.
34
35
The function monitored for the indicator is the ability of the monitored system to take suction
36
from the suppression pool (and from the condensate storage tank, if credited in the plants
37
accident analysis) and inject into the reactor vessel.
38
39
Plants should monitor either the high-pressure coolant injection (HPCI), the high-pressure core
40
spray (HPCS), or the feedwater coolant injection (FWCI) system, whichever is installed. The
41
turbine and governor (or motor-driven FWCI pumps), and associated piping and valves for
42
turbine steam supply and exhaust are within the scope of these systems. Valves in the feedwater
43
line are not considered within the scope of these systems. The emergency generator dedicated to
44
12
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
providing AC power to the high-pressure core spray system is included in the scope of the
1
HPCS. The HPCS system typically includes a "water leg" pump to prevent water hammer in the
2
HPCS piping to the reactor vessel. The "water leg" pump and valves in the "water leg" pump
3
flow path are ancillary components and are not included in the scope of the HPCS system.
4
Unavailability is not included while critical if the system is below steam pressure specified in
5
technical specifications at which the system can be operated.
6
7
Train Determination
8
The HPCI and HPCS systems are considered single-train systems. The booster pump and other
9
small pumps are ancillary components not used in determining the number of trains. The effect
10
of these pumps on system performance is included in the system indicator to the extent their
11
failure detracts from the ability of the system to perform its risk-significant function. For the
12
FWCI system, the number of trains is determined by the number of feedwater pumps. The
13
number of condensate and feedwater booster pumps are not used to determine the number of
14
trains.
15
16
BWR Heat Removal Systems
17
(Reactor Core Isolation Cooling or Isolation Condenser)
18
19
Scope
20
This system functions at high pressure to remove decay heat following a loss of main feedwater
21
event. The RCIC system also functions to maintain reactor coolant inventory following a very
22
small LOCA event.
23
24
The function monitored for the indicator is the ability of the RCIC system to cool the reactor
25
vessel core and provide makeup water by taking a suction from either the condensate storage
26
tank or the suppression pool and injecting at rated pressure and flow into the reactor vessel.
27
28
The Reactor Core Isolation Cooling (RCIC) system turbine, governor, and associated piping and
29
valves for steam supply and exhaust are within the scope of the RCIC system. Valves in the
30
feedwater line are not considered within the scope of the RCIC system. The Isolation Condenser
31
and inlet valves are within the scope of Isolation Condenser system. Unavailability is not
32
included while critical if the system is below steam pressure specified in technical specifications
33
at which the system can be operated.
34
35
36
Train Determination
37
The RCIC system is considered a single-train system. The condensate and vacuum pumps are
38
ancillary components not used in determining the number of trains. The effect of these pumps on
39
RCIC performance is included in the system indicator to the extent that a component failure
40
results in an inability of the system to perform its risk-significant function.
41
13
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
1
BWR Residual Heat Removal Systems
2
Scope
3
The functions monitored for the BWR residual heat removal (RHR) system are the ability of the
4
RHR system to remove heat from the suppression pool, provide low pressure coolant injection,
5
and provide post-accident decay heat removal. The pumps, heat exchangers, and associated
6
piping and valves for those functions are included in the scope of the RHR system.
7
8
Train Determination
9
The number of trains in the RHR system is determined by the number of parallel RHR heat
10
exchangers.
11
12
PWR High Pressure Safety Injection Systems
13
Scope
14
These systems are used primarily to maintain reactor coolant inventory at high pressures
15
following a loss of reactor coolant. HPSI system operation following a small-break LOCA
16
involves transferring an initial supply of water from the refueling water storage tank (RWST) to
17
cold leg piping of the reactor coolant system. Once the RWST inventory is depleted,
18
recirculation of water from the reactor building emergency sump is required. The function
19
monitored for HPSI is the ability of a HPSI train to take a suction from the primary water source
20
(typically, a borated water tank), or from the containment emergency sump, and inject into the
21
reactor coolant system at rated flow and pressure.
22
23
The scope includes the pumps and associated piping and valves from both the refueling water
24
storage tank and from the containment sump to the pumps, and from the pumps into the reactor
25
coolant system piping. For plants where the high-pressure injection pump takes suction from the
26
residual heat removal pumps, the residual heat removal pump discharge header isolation valve to
27
the HPSI pump suction is included in the scope of HPSI system. Some components may be
28
included in the scope of more than one train. For example, cold-leg injection lines may be fed
29
from a common header that is supplied by both HPSI trains. In these cases, the effects of testing
30
or component failures in an injection line should be reported in both trains.
31
32
Train Determination
33
34
In general, the number of HPSI system trains is defined by the number of high head injection
35
paths that provide cold-leg and/or hot-leg injection capability, as applicable.
36
37
For Babcock and Wilcox (B&W) reactors, the design features centrifugal pumps used for high
38
pressure injection (about 2,500 psig) and no hot-leg injection path. Recirculation from the
39
containment sump requires operation of pumps in the residual heat removal system. They are
40
14
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
typically a two-train system, with an installed spare pump (depending on plant-specific design)
1
that can be aligned to either train.
2
3
For two-loop Westinghouse plants, the pumps operate at a lower pressure (about 1600 psig) and
4
there may be a hot-leg injection path in addition to a cold-leg injection path (both are included as
5
a part of the train).
6
7
For Combustion Engineering (CE) plants, the design features three centrifugal pumps that
8
operate at intermediate pressure (about 1300 psig) and provide flow to two cold-leg injection
9
paths or two hot-leg injection paths. In most designs, the HPSI pumps take suction directly from
10
the containment sump for recirculation. In these cases, the sump suction valves are included
11
within the scope of the HPSI system. This is a two-train system (two trains of combined cold-leg
12
and hot-leg injection capability). One of the three pumps is typically an installed spare that can
13
be aligned to either train or only to one of the trains (depending on plant-specific design).
14
15
For Westinghouse three-loop plants, the design features three centrifugal pumps that operate at
16
high pressure (about 2500 psig), a cold-leg injection path through the BIT (with two trains of
17
redundant valves), an alternate cold-leg injection path, and two hot-leg injection paths. One of
18
the pumps is considered an installed spare. Recirculation is provided by taking suction from the
19
RHR pump discharges. A train consists of a pump, the pump suction valves and boron injection
20
tank (BIT) injection line valves electrically associated with the pump, and the associated hot-leg
21
injection path. The alternate cold-leg injection path is required for recirculation, and should be
22
included in the train with which its isolation valve is electrically associated. This represents a
23
two-train HPSI system.
24
25
For Four-loop Westinghouse plants, the design features two centrifugal pumps that operate at
26
high pressure (about 2500 psig), two centrifugal pumps that operate at an intermediate pressure
27
(about 1600 psig), a BIT injection path (with two trains of injection valves), a cold-leg safety
28
injection path, and two hot-leg injection paths. Recirculation is provided by taking suction from
29
the RHR pump discharges. Each of two high pressure trains is comprised of a high pressure
30
centrifugal pump, the pump suction valves and BIT valves that are electrically associated with
31
the pump. Each of two intermediate pressure trains is comprised of the safety injection pump, the
32
suction valves and the hot-leg injection valves electrically associated with the pump. The cold-
33
leg safety injection path can be fed with either safety injection pump, thus it should be associated
34
with both intermediate pressure trains. This HPSI system is considered a four-train system for
35
monitoring purposes.
36
37
38
39
PWR Auxiliary Feedwater Systems
40
Scope
41
The AFW system provides decay heat removal via the steam generators to cool down and
42
depressurize the reactor coolant system following a reactor trip. The AFW system is assumed to
43
be required for an extended period of operation during which the initial supply of water from the
44
condensate storage tank is depleted and water from an alternative water source (e.g., the service
45
water system) is required. Therefore components in the flow paths from both of these water
46
15
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
sources are included; however, the alternative water source (e.g., service water system) is not
1
included.
2
3
The function monitored for the indicator is the ability of the AFW system to take a suction from
4
the primary water source (typically, the condensate storage tank) or, if required, from an
5
emergency source (typically, a lake or river via the service water system) and inject into at least
6
one steam generator at rated flow and pressure.
7
8
The scope of the auxiliary feedwater (AFW) or emergency feedwater (EFW) systems includes
9
the pumps and the components in the flow paths from the condensate storage tank and, if
10
required, the valve(s) that connect the alternative water source to the auxiliary feedwater system.
11
Startup feedwater pumps are not included in the scope of this indicator.
12
13
Train Determination
14
The number of trains is determined primarily by the number of parallel pumps. For example, a
15
system with three pumps is defined as a three-train system, whether it feeds two, three, or four
16
injection lines, and regardless of the flow capacity of the pumps. Some components may be
17
included in the scope of more than one train. For example, one set of flow regulating valves and
18
isolation valves in a three-pump, two-steam generator system are included in the motor-driven
19
pump train with which they are electrically associated, but they are also included (along with the
20
redundant set of valves) in the turbine-driven pump train. In these instances, the effects of testing
21
or failure of the valves should be reported in both affected trains. Similarly, when two trains
22
provide flow to a common header, the effect of isolation or flow regulating valve failures in
23
paths connected to the header should be considered in both trains.
24
25
PWR Residual Heat Removal System
26
Scope
27
The functions monitored for the PWR residual heat removal (RHR) system are those that are
28
required to be available when the reactor is critical. These typically include the low-pressure
29
injection function (if risk-significant) and the post-accident recirculation mode used to cool and
30
recirculate water from the containment sump following depletion of RWST inventory to provide
31
post-accident decay heat removal. The pumps, heat exchangers, and associated piping and valves
32
for those functions are included in the scope of the RHR system. Containment spray function
33
should be included if it is identified in the PRA as a risk-significant post accident decay heat
34
removal function. Containment spray systems that only provide containment pressure control are
35
not included.
36
37
38
39
Train Determination
40
The number of trains in the RHR system is determined by the number of parallel RHR heat
41
exchangers. Some components are used to provide more than one function of RHR. If a
42
component cannot perform as designed, rendering its associated train incapable of meeting one
43
16
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
of the risk-significant functions, then the train is considered to be failed. Unavailable hours
1
would be reported as a result of the component failure.
2
Cooling Water Support System
3
Scope
4
The function of the cooling water support system is to provide for direct cooling of the
5
components in the other monitored systems. It does not include indirect cooling provided by
6
room coolers or other HVAC features.
7
8
Systems that provide this function typically include service water and component cooling water
9
or their cooling water equivalents. Pumps, valves, heat exchangers and line segments that are
10
necessary to provide cooling to the other monitored systems are included in the system scope up
11
to, but not including, the last valve that connects the cooling water support system to the other
12
monitored systems. This last valve is included in the other monitored system boundary.
13
14
Valves in the cooling water support system that must close to ensure sufficient cooling to the
15
other monitored system components to meet risk significant functions are included in the system
16
boundary.
17
18
19
20
Train Determination
21
The number of trains in the Cooling Water Support System will vary considerably from plant to
22
plant. The way these functions are modeled in the plant-specific PRA will determine a logical
23
approach for train determination. For example, if the PRA modeled separate pump and line
24
segments, then the number of pumps and line segments would be the number of trains.
25
26
Clarifying Notes
27
Service water pump strainers and traveling screens are not considered to be active components
28
and are therefore not part of URI. However, clogging of strainers and screens due to expected or
29
routinely predictable environmental conditions that render the train unavailable to perform its
30
risk significant cooling function (which includes the risk-significant mission times)are included
31
in UAI.
32
33
Unpredictable extreme environmental conditions that render the train unavailable to perform its
34
risk significant cooling function should be addressed through the FAQ process to determine if
35
resulting unavailability should be included in UAI.
36
37
Attachment 2
NEI 99-02, Appendix F, Methodologies For Computing the Unavailability Index, the
Unreliability Index and Determining Performance Index Validity (Draft).
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-1
APPENDIX F
1
2
METHODOLOGIES FOR COMPUTING THE UNAVAILABILITY
3
INDEX, THE UNRELIABILITY INDEX AND DETERMINING
4
PERFORMANCE INDEX VALIDITY
5
This appendix provides the details of three calculations, calculation of the System
6
Unavailability Index, the System Unreliability Index, and the criteria for determining
7
when the Mitigating System Performance Index is unsuitable for use as a performance
8
index.
9
System Unavailability Index (UAI) Due to Changes in Train Unavailability
10
Calculation of System UAI due to changes in train unavailability is as follows:
11
UAI
UAItj
j 1
n
Eq. 1
12
where the summation is over the number of trains (n) and UAIt is the unavailability index
13
for a train.
14
Calculation of UAIt for each train due to changes in train unavailability is as follows:
15
)
(
max
BLt
t
p
UAp
p
t
UA
UAI
,
Eq. 2
16
where:
17
CDFp is the plant-specific, internal events, at power Core Damage Frequency,
18
FVUAp is the train-specific Fussell-Vesely value for unavailability,
19
UAP is the plant-specific PRA value of unavailability for the train,
20
UAt is the actual unavailability of train t, defined as:
21
quarters
12
previous
the
during
hours
Critical
critical
while
quarters
12
previous
the
during
hours
e
Unavailabl
t
22
and,
23
UABLt is the historical baseline unavailability value for the train determined
24
as described below.
25
UABLt is the sum of two elements: planned and unplanned unavailability. Planned
26
unavailability is the actual, plant-specific three-year total planned unavailability
27
for the train for the years 1999 through 2001 (see clarifying notes for details).
28
This period is chosen as the most representative of how the plant intends to
29
perform routine maintenance and surveillances at power. Unplanned
30
unavailability is the historical industry average for unplanned unavailability for
31
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-2
the years 1999 through 2001. See Table 1 for historical train values for
1
unplanned unavailability.
2
Calculation of the quantity inside the square bracket in equation 2 will be discussed at the
3
end of the next section. See clarifying notes for calculation of UAI for cooling water
4
support system.
5
6
System Unreliability Index (URI) Due to Changes in Component Unreliability
7
Unreliability is monitored at the component level and calculated at the system level.
8
Calculation of system URI due to changes in component unreliability is as follows:
9
)
(
1
max
BLcj
Bcj
m
j
pcj
URcj
p
UR
UR
UR
Eq. 3
10
Where the summation is over the number of active components (m) in the system, and:
11
CDFp is the plant-specific internal events, at power, core damage frequency,
12
FVURc is the component-specific Fussell-Vesely value for unreliability,
13
URPc is the plant-specific PRA value of component unreliability,
14
URBc is the Bayesian corrected component unreliability for the previous 12
15
quarters,
16
and
17
URBLc is the historical industry baseline calculated from unreliability mean values
18
for each monitored component in the system. The calculation is performed in a
19
manner similar to equation 4 below using the industry average values in Table 2.
20
Calculation of the quantity inside the square bracket in equation 3 will be discussed at the
21
end of this section.
22
Component unreliability is calculated as follows.
23
URBc PD Tm
Eq 4
24
where:
25
PD is the component failure on demand probability calculated based on data
26
collected during the previous 12 quarters,
27
is the component failure rate (per hour) for failure to run calculated based on
28
data collected during the previous 12 quarters,
29
and
30
Tm is the risk-significant mission time for the component based on plant specific
31
PRA model assumptions. Add acceptable methodologies for determining mission
32
time.
33
34
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-3
NOTE:
1
For valves only the PD term applies
2
For pumps PD + Tm applies
3
For diesels PD start + PD load run + Tm applies
4
5
The first term on the right side of equation 4 is calculated as follows.1
6
(Nd a)
(a b D)
Eq. 5
7
where:
8
Nd is the total number of failures on demand during the previous 12 quarters,
9
D is the total number of demands during the previous 12 quarters (actual ESF
10
demands plus estimated test and estimated operational/alignment demands. An
11
update to the estimated demands is required if a change to the basis for the
12
estimated demands results in a >25% change in the estimate),
13
and
14
a and b are parameters of the industry prior, derived from industry experience (see
15
Table 2).
16
In the calculation of equation 5 the numbers of demands and failures is the sum of all
17
demands and failures for similar components within each system. Do not sum across
18
units for a multi-unit plant. For example, for a plant with two trains of Emergency Diesel
19
Generators, the demands and failures for both trains would be added together for one
20
evaluation of PD which would be used for both trains of EDGs.
21
In the second term on the right side of equation 4, is calculated as follows.
22
(Nr a)
(Tr b)
Eq. 6
23
where:
24
Nr is the total number of failures to run during the previous 12 quarters,
25
Tr is the total number of run hours during the previous 12 quarters (actual ESF run
26
hours plus estimated test and estimated operational/alignment run hours. An
27
update to the estimated run hours is required if a change to the basis for the
28
estimated hours results in a >25% change in the estimate).
29
and
30
1 Atwood, Corwin L., Constrained noninformative priors in risk assessment, Reliability
Engineering and System Safety, 53 (1996; 37-46)
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-4
a and b are parameters of the industry prior, derived from industry experience (see
1
Table 2).
2
In the calculation of equation 6 the numbers of demands and run hours is the sum of all
3
run hours and failures for similar components within each system. Do not sum across
4
units for a multi-unit plant. For example, a plant with two trains of Emergency Diesel
5
Generators, the run hours and failures for both trains would be added together for one
6
evaluation of which would be used for both trains of EDGs.
7
Fussell-Vesely, Unavailability and Unreliability
8
Equations 2 and 3 include a term that is the ratio of a Fussell-Vesely importance value
9
divided by the related unreliability or unavailability. Calculation of these quantities is
10
generally complex, but in the specific application used here, can be greatly simplified.
11
The simplifying feature of this application is that only those components (or the
12
associated basic events) that can fail a train are included in the performance index.
13
Components within a train that can each fail the train are logically equivalent and the
14
ratio FV/UR is a constant value for any basic event in that train. It can also be shown that
15
for a given component or train represented by multiple basic events, the ratio of the two
16
values for the component or train is equal to the ratio of values for any basic event within
17
the train. Or:
18
FVbe
URbe FVURc
URPc FVt
URt Constant
19
and
20
FVbe
UAbe FVUAp
UAp Constant
21
Note that the constant value may be different for the unreliability ratio and the
22
unavailability ratio because the two types of events are frequently not logically
23
equivalent. For example recovery actions may be modeled in the PRA for one but not the
24
other.
25
Thus, the process for determining the value of this ratio for any component or train is to
26
identify a basic event that fails the component or train, determine the failure probability
27
or unavailability for the event, determine the associated FV value for the event and then
28
calculate the ratio. Use the basic event in the component or train with the largest failure
29
probability (hence the maximum notation on the bracket) to minimize the effects of
30
truncation on the calculation. Exclude common cause events, which are not within the
31
scope of this performance index
32
Some systems have multiple modes of operation, such as PWR HPSI systems that operate
33
in injection as well as recirculation modes. In these systems all active components are not
34
logically equivalent, unavailability of the pump fails all operating modes while
35
unavailability of the sump suction valves only fails the recirculation mode. In cases such
36
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-5
as these, if unavailability events exist separately for the components within a train, the
1
appropriate ratio to use is the maximum.
2
Determination of systems for which the performance index is not valid
3
The performance index relies on the existing testing programs as the source of the data
4
that is input to the calculations. Thus, the number of demands in the monitoring period is
5
based on the frequency of testing required by the current test programs. In most cases this
6
will provide a sufficient number of demands to result in a valid statistical result.
7
However, in some cases, the number of demands will be insufficient to resolve the
8
change in the performance index (1.0x10-6) that corresponds to movement from a green
9
performance to a white performance level. In these cases, one failure is the difference
10
between baseline performance and performance in the white performance band. The
11
performance index is not suitable for monitoring such systems and monitoring is
12
performed through the inspection process.
13
This section will define the method to be used to identify systems for which the
14
performance index is not valid, and will not be used.
15
The criteria to be used to identify an invalid performance index is:
16
If, for any failure mode for any component in a system, the risk increase
17
(CDF) associated with the change in unreliability resulting from single
18
failure is larger than 1.0x10-6, then the performance index will be
19
considered invalid for that system.
20
The increase in risk associated with a component failure is the sum of the contribution
21
from the decrease in calculated reliability as a result of the failure and the decrease in
22
availability resulting from the time required to affect the repair of the failed component.
23
The change in CDF that results from a demand type failure is given by:
24
25
CR
Mean
p
UAp
p
comp
similar
N
pc
URc
p
T
T
D
b
a
UR
Repair
1
Eq. 7
26
27
Likewise, the change in CDF per run type failure is given by:
28
29
CR
p
UAp
p
comp
similar
N
r
m
pc
URc
p
T
T
T
b
T
UR
Repair
Mean
Eq. 8
30
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-6
In these expressions, the variables are as defined earlier and additionally
1
TMR is the mean time to repair for the component
2
and
3
TCR is the number of critical hours in the monitoring period.
4
The summation in the equations is taken over all similar components within a system.
5
With multiple components of a given type in one system, the impact of the failure on
6
CDF is included in the increased unavailability of all components of that type due to
7
pooling the demand and failure data.
8
The mean time to repair can be estimate as one-half the Technical Specification Allowed
9
Outage Time for the component and the number of critical hours should correspond to the
10
1999 - 2001 actual number of critical hours.
11
These equations are be used for all failure modes for each component in a system. If the
12
resulting value of CDF is greater than 1.0x10-6 for any failure mode of any component,
13
then the performance index for that system is not considered valid.
14
15
Definitions
16
17
Train Unavailability: Train unavailability is the ratio of the hours the train was
18
unavailable to perform its risk-significant functions due to planned or unplanned
19
maintenance or test during the previous 12 quarters while critical to the number of critical
20
hours during the previous 12 quarters. (Fault exposure hours are not included;
21
unavailable hours are counted only for the time required to recover the trains risk-
22
significant functions.)
23
Train unavailable hours: The hours the train was not able to perform its risk significant
24
function due to maintenance, testing, equipment modification, electively removed from
25
service, corrective maintenance, or the elapsed time between the discovery and the
26
restoration to service of an equipment failure or human error that makes the train
27
unavailable (such as a misalignment) while the reactor is critical.
28
Fussell-Vesely (FV) Importance:
29
The Fussell-Vesely importance for a feature (component, sub-system, train, etc.) of a
30
system is representative of the fractional contribution that feature makes to the to the total
31
risk of the system.
32
The Fussell-Vesely importance of a basic event or group of basic events that represent a
33
feature of a system is represented by:
34
0
1
R
R
i
35
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-7
Where:
1
R0 is the base (reference) case overall model risk,
2
Ri is the decreased risk level with feature i completely reliable.
3
In this expression, the second term on the right represents the fraction of the reference
4
risk remaining assuming the feature of interest is perfect. Thus 1 minus the second term is
5
the fraction of the reference risk attributed to the feature of interest.
6
The Fussell-Vesely importance is calculated according to the following equation:
7
m
j
j
n
j
j
i
C
C
,1
0
,1
1
,
8
where the denominator represents the union of m minimal cutsets C0 generated with the
9
reference (baseline) model, and the numerator represents the union of n minimal cutsets
10
Ci generated assuming events related to the feature are perfectly reliable, or their failure
11
probability is False.
12
Critical hours: The number of hours the reactor was critical during a specified period of
13
time.
14
Component Unreliability: Component unreliability is the probability that the component
15
would not perform its risk-significant functions when called upon during the previous 12
16
quarters.
17
Active Component: A component whose failure to change state renders the train incapable
18
of performing its risk-significant functions. In addition, all pumps and diesels in the
19
monitored systems are included as active components. (See clarifying notes.)
20
Manual Valve: A valve that can only be operated by a person. An MOV or AOV that is
21
remotely operated by a person may be an active component.
22
Start demand: Any demand for the component to successfully start to perform its risk-
23
significant functions, actual or test. (Exclude post maintenance tests, unless in case of a
24
failure the cause of failure was independent of the maintenance performed.)
25
Post maintenance tests: Tests performed following maintenance but prior to declaring the
26
train/component operable, consistent with Maintenance Rule implementation.
27
Run demand: Any demand for the component, given that it has successfully started, to
28
run/operate for its mission time to perform its risk-significant functions. (Exclude post
29
maintenance tests, unless in case of a failure the cause of failure was independent of the
30
maintenance performed.)
31
EDG failure to start: A failure to start includes those failures up to the point the EDG has
32
achieved rated speed and voltage. (Exclude post maintenance tests, unless the cause of
33
failure was independent of the maintenance performed.)
34
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-8
EDG failure to load/run: Given that it has successfully started, a failure of the EDG
1
output breaker to close, loads successfully sequence and to run/operate for one hour to
2
perform its risk-significant functions. This failure mode is treated as a demand failure for
3
calculation purposes. (Exclude post maintenance tests, unless the cause of failure was
4
independent of the maintenance performed.)
5
EDG failure to run: Given that it has successfully started and loaded and run for an hour,
6
a failure of an EDG to run/operate. for its mission time to perform its risk-significant
7
functions. (Exclude post maintenance tests, unless the cause of failure was independent of
8
the maintenance performed.)
9
Pump failure on demand: A failure to start and run for at least one hour is counted as
10
failure on demand. (Exclude post maintenance tests, unless the cause of failure was
11
independent of the maintenance performed.)
12
Pump failure to run: Given that it has successfully started and run for an hour, a failure of
13
a pump to run/operate. for its mission time to perform its risk-significant functions.
14
(Exclude post maintenance tests, unless the cause of failure was independent of the
15
maintenance performed.)
16
Valve failure on demand: A failure to open or close is counted as failure on demand.
17
(Exclude post maintenance tests, unless the cause of failure was independent of the
18
maintenance performed.)
19
Clarifying Notes
20
Train Boundaries and Unavailable Hours
21
Include all components that are required to satisfy the risk-significant function of the
22
train. For example, high-pressure injection may have both an injection mode with
23
suction from the refueling water storage tank and a recirculation mode with suction from
24
the containment sump. Some components may be included in the scope of more than one
25
train. For example, one set of flow regulating valves and isolation valves in a three-pump,
26
two-steam generator system are included in the motor-driven pump train with which they
27
are electrically associated, but they are also included (along with the redundant set of
28
valves) in the turbine-driven pump train. In these instances, the effects of unavailability
29
of the valves should be reported in both affected trains. Similarly, when two trains
30
provide flow to a common header, the effect of isolation or flow regulating valve failures
31
in paths connected to the header should be considered in both trains
32
Cooling Water Support System Trains
33
The number of trains in the Cooling Water Support System will vary considerably from
34
plant to plant. The way these functions are modeled in the plant-specific PRA will
35
determine a logical approach for train determination. For example, if the PRA modeled
36
separate pump and line segments, then the number of pumps and line segments would be
37
the number of trains. A separate value for UAI and URI will be calculated for each of the
38
systems in this indicator and then they will be added together to calculate the MSPI.
39
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-9
1
Active Components
2
For unreliability, use the following criteria for determining those components that should
3
be monitored:
4
Components that are normally running or have to change state to achieve the risk
5
significant function will be included in the performance index. Active failures of
6
check valves and manual valves are excluded from the performance index and will be
7
evaluated in the NRC inspection program.
8
Redundant valves within a train are not included in the performance index. Only
9
those valves whose failure alone can fail a train will be included. The PRA success
10
criteria are to be used to identify these valves.
11
Redundant valves within a multi-train system, whether in series or parallel, where the
12
failure of both valves would prevent all trains in the system from performing a risk-
13
significant function are included. (See Figure F-5)
14
All pumps and diesels are included in the performance index
15
Table 3 defines the boundaries of components, and Figures F-1, F-2, F-3 and F-4 provide
16
examples of typical component boundaries as described in Table 3. Each plant will
17
determine their system boundaries, active components, and support components, and
18
have them available for NRC inspection.
19
Failures of Non-Active Components
20
Failures of SSCs that are not included in the performance index will not be counted as a
21
failure or a demand. Failures of SSCs that cause an SSC within the scope of the
22
performance index to fail will not be counted as a failure or demand. An example could
23
be a manual suction isolation valve left closed which causes a pump to fail. This would
24
not be counted as a failure of the pump. Any mispositioning of the valve that caused the
25
train to be unavailable would be counted as unavailability from the time of discovery.
26
The significance of the mispositioned valve prior to discovery would be addressed
27
through the inspection process.
28
29
Baseline Values
30
The baseline values for unreliability are contained in Table 2 and remain fixed.
31
The baseline values for unavailability include both plant-specific planned unavailability
32
values and unplanned unavailability values. The unplanned unavailability values are
33
contained in Table 1 and remain fixed. They are based on ROP PI industry data from
34
1999 through 2001. (Most baseline data used in PIs come from the 1995-1997 time
35
period. However, in this case, the 1999-2001 ROP data are preferable, because the ROP
36
data breaks out systems separately (some of the industry 1995-1997 INPO data combine
37
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-10
systems, such as HPCI and RCIC, and do not include PWR RHR). It is important to note
1
that the data for the two periods is very similar.)
2
Support cooling baseline data is based on plant specific maintenance rule unplanned and
3
planned unavailability for years 1999 to 2001. (Maintenance rule data does not
4
distinguish between planned and unplanned unavailability. There is no ROP support
5
cooling data.)
6
The baseline planned unavailability is based on actual plant-specific values for the period
7
1999 through 2001. These values are expected to remain fixed unless the plant
8
maintenance philosophy is substantially changed with respect to on-line maintenance or
9
preventive maintenance. In these cases, the planned unavailability baseline value can be
10
adjusted. A comment should be placed in the comment field of the quarterly report to
11
identify a substantial change in planned unavailability. To determine the planned
12
unavailability:
13
1. Record the total train unavailable hours reported under the Reactor Oversight Process
14
for 1999 through 2001.
15
2. Subtract any fault exposure hours still included in the 1999-2001 period.
16
3. Subtract unplanned unavailable hours
17
4. Add any on-line overhaul hours and any other planned unavailability excluded in
18
accordance with NEI 99-02. 2
19
5. Add any planned unavailable hours for functions monitored under MSPI which were
20
not monitored under SSU in NEI 99-02.
21
6. Subtract any unavailable hours reported when the reactor was not critical.
22
7. Subtract hours cascaded onto monitored systems by support systems.
23
8. Divide the hours derived from steps 1-6 above by the total critical hours during 1999-
24
2001. This is the baseline planned unavailability
25
Baseline unavailability is the sum of planned unavailability from step 7 and unplanned
26
unavailability from Table 1.
27
28
29
2 Note: The plant-specific PRA should model significant on-line overhaul hours.
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-11
Table 1. Historical Unplanned Maintenance Unavailability Train Values
1
(Based on ROP Industrywide Data for 1999 through 2001)
2
3
4
SYSTEM
UNPLANNED UNAVAILABILITY/TRAIN
1.7 E-03
6.1 E-04
9.1 E-04
6.9 E-04
7.6 E-04
4.2 E-04
1.1 E-03
3.3 E-03
5.4 E-04
2.9 E-03
1.2 E-03
Support Cooling
No Data Available Use plant specific Maintenance
Rule data for 1999-2001
5
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-12
Table 2. Industry Priors and Parameters for Unreliability
1
2
3
Component
Failure
Mode
a a
b a
Industry
Mean
Value b
Source(s)
Motor-operated
valve
Fail to open
(or close)
5.0E-1
2.4E+2
2.1E-3
NUREG/CR-5500, Vol.
4,7,8,9
Air-operated
valve
Fail to open
(or close)
5.0E-1
2.5E+2
2.0E-3
NUREG/CR-4550, Vol. 1
Fail to start
5.0E-1
2.4E+2
2.1E-3
NUREG/CR-5500, Vol.
1,8,9
Motor-driven
pump, standby
Fail to run
5.0E-1
5.0E+3h
1.0E-4/h
NUREG/CR-5500, Vol.
1,8,9
Fail to start
4.9E-1
1.6E+2
3.0E-3
NUREG/CR-4550, Vol. 1
Motor-driven
pump, running
or alternating
Fail to run
5.0E-1
1.7E+4h
3.0E-5/h
NUREG/CR-4550, Vol. 1
Fail to start
4.7E-1
2.4E+1
1.9E-2
NUREG/CR-5500, Vol. 1
Turbine-driven
pump, AFWS
Fail to run
5.0E-1
3.1E+2
1.6E-3/h
NUREG/CR-5500, Vol. 1
Fail to start
4.6E-1
1.7E+1
2.7E-2
NUREG/CR-5500, Vol.
4,7
Turbine-driven
pump, HPCI or
Fail to run
5.0E-1
3.1E+2h
1.6E-3/h
NUREG/CR-5500, Vol.
1,4,7
Fail to start
4.7E-1
2.4E+1
1.9E-2
NUREG/CR-5500, Vol. 1
Diesel-driven
pump, AFWS
Fail to run
5.0E-1
6.3E+2h
8.0E-4/h
NUREG/CR-4550, Vol. 1
Fail to start
4.8E-1
4.3E+1
1.1E-2
NUREG/CR-5500, Vol. 5
Fail to
load/run
5.0E-1
2.9E+2
1.7E-3 c
NUREG/CR-5500, Vol. 5
Emergency
diesel generator
Fail to run
5.0E-1
2.2E+3h
2.3E-4/h
NUREG/CR-5500, Vol. 5
4
5
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-13
a. A constrained, non-informative prior is assumed. For failure to run events, a = 0.5 and
1
b = (a)/(mean rate). For failure upon demand events, a is a function of the mean
2
probability:
3
4
Mean Probability
a
5
0.0 to 0.0025
0.50
6
>0.0025 to 0.010
0.49
7
>0.010 to 0.016
0.48
8
>0.016 to 0.023
0.47
9
>0.023 to 0.027
0.46
10
11
Then b = (a)(1.0 - mean probability)/(mean probability).
12
13
b. Failure to run events occurring within the first hour of operation are included within
14
the fail to start failure mode. Failure to run events occurring after the first hour of
15
operation are included within the fail to run failure mode. Unless otherwise noted, the
16
mean failure probabilities and rates include the probability of non-recovery. Types of
17
allowable recovery are outlined in the clarifying notes, under Credit for Recovery
18
Actions.
19
20
c. Fail to load and run for one hour was calculated from the failure to run data in the
21
report indicated. The failure rate for 0.0 to 0.5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> (3.3E-3/h) multiplied by 0.5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />,
22
was added to the failure rate for 0.5 to 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> (2.3E-4/h) multiplied by 0.5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />.
23
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-14
Table 3. Component Boundary Definition
Component
Component boundary
Diesel
Generators
The diesel generator boundary includes the generator body, generator
actuator, lubrication system (local), fuel system (local), cooling components
(local), startup air system receiver, exhaust and combustion air system,
dedicated diesel battery (which is not part of the normal DC distribution
system), individual diesel generator control system, circuit breaker for supply
to safeguard buses and their associated local control circuit (coil, auxiliary
contacts, wiring and control circuit contacts, and breaker closure interlocks) .
Motor-Driven
Pumps
The pump boundary includes the pump body, motor/actuator, lubrication
system cooling components of the pump seals, the voltage supply breaker,
and its associated local control circuit (coil, auxiliary contacts, wiring and
control circuit contacts).
Turbine-
Driven Pumps
The turbine-driven pump boundary includes the pump body, turbine/actuator,
lubrication system (including pump), extractions, turbo-pump seal, cooling
components, and local turbine control system (speed).
Motor-
Operated
Valves
The valve boundary inc1udes the valve body, motor/actuator, the voltage
supply breaker (both motive and control power) and its associated local
open/close circuit (open/close switches, auxiliary and switch contacts, and
wiring and switch energization contacts).
Air-Operated
Valves
The valve boundary includes the valve body, the air operator, associated
solenoid-operated valve, the power supply breaker or fuse for the solenoid
valve, and its associated control circuit (open/close switches and local
auxiliary and switch contacts).
1
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Figure F-1
23
Diesel Engine
Control and
Protection System
Starting Air
System Receiver
Combustion Air
System and
Supply
Jacket
Water
Fuel Oil
System
Fuel Oil Day
Tank
Generator
Exciter and
Voltage
Regulator
Exhaust
System
Governor and
Control System
Lubrication
System
Breaker
ESFAS/Sequencer
DC Power
Cooling Water
Class 1E Bus
EDG Boundary
Isol.
Valve
Fuel Storage and
Transfer System
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-16
1
2
Figure F-2
3
4
5
Controls
Breaker
Motor Operator
Motor Driven Pump Boundary
Pump
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-17
Figure F-3
1
2
Controls
Breaker
Motor Operator
MOV Boundary
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-18
1
2
Figure F-4
3
4
Controls
Turbine
Turbine Driven Pump Boundary
Pump
DRAFT NEI 99-02 MSPI 8/28/20028/23/20028/9/2002
F-19
1
T
A
N
K
Figure F-5
Active
Components
Active
Components
Non-active
Components