Selecting HLR and SR Standards for PRA Applications

1
Selecting HLR and SR Standards for PRA
Applications

• ICONE16-48598

2
Introduction

• Identify which ASME PRA Standard High Level
  Requirements (HLRs) fall in the scope of a
  particular risk informed application to the NRC
  that will, in part, justify the requested change
  with risk metrics.
• What should be changed in the model to represent
  the "risk informed" topic at hand?

3
(No Transcript)
4
HLRs

• 4.5.1 Initiating Events Analysis (IE)
• 4.5.2 Accident Sequence Analysis (AS)
• 4.5.3 Success Criteria (SC)
• 4.5.4 Systems Analysis (SY)
• 4.5.5 Human Reliability Analysis (HR)
• 4.5.6 Data Analysis (DA)
• 4.5.7 Internal Flooding (IF)
• 4.5.8 Quantification (QU)
• 4.5.9 LERF Analysis (LE)

5
(No Transcript)
6

• Evaluate the plant or operational change being
  assessed.

• Identify SSCs and activities affected by the
  plant change, including cause and effect
  relationship.

• Identify PRA scope and risk metrics needed to
  evaluate plant changes and to support application.

Parts irrelevant to application
Parts needing Capability Category I

• Determine capability category needed for each
  part of PRA to support application

Parts needing Capability Category II
Parts needing Capability Category III
7
Assign SSCs to Partition I Partition II Partition
III
Assign SSCs in Partition I to Partition
I-APartition I-B Partition I-C Partition I-D

• Identify SSCs and activities affected by the
  plant change, including cause and effect
  relationship.

Assignment to Partitions I-A to I-D results in a
mapping of a SSC to a model element.

• Identify PRA scope and risk metrics needed to
  evaluate plant changes and to support application.

A model elementis either an Initiator Mitigati
ng event probability Human Recovery Action /
Operator Action
The type of model element isdirectly
correlated toone or more ASME HLRs
8
Applications

• When the application involves changes to success
  criteria and process parameters, e.g., PWR
  hot-leg temperature (Thot) reduction, the effect
  is on the logic.
• When plant changes involve addition or deletion
  of equipment or modification in the use of
  existing equipment (aka. procedure changes),
  those changes should consider the risk impact
  from the perspectives of event mitigation and
  initiation.

logic
event mitigation
initiation
9
Assign SSCs to Partition I Partition II Partition
III
Assign SSCs in Partition I to Partition
I-APartition I-B Partition I-C Partition I-D

• Identify SSCs and activities affected by the
  plant change, including cause and effect
  relationship.

Assignment to Partitions I-A to I-D results in a
mapping of a SSC to a model element.

• Identify PRA scope and risk metrics needed to
  evaluate plant changes and to support application.

A model elementis either an Initiator Mitigati
ng event probability Human Recovery Action /
Operator Action
The type of model element isdirectly
correlated toone or more ASME HLRs
10
Assign SSCs to Partition I Partition II Partition
III
The PRA model addresses (or should address)
failure of the SSC. The SSC may be modeled
explicitly, modeled via a subsuming component, or
modeled via a surrogate event.A SSC in this
group will be further classified.
11
Assign SSCs to Partition I Partition II Partition
III
A SSC excluded from the PRA model cannot affect
calculated risk. It is likely that any SSC that
can mitigate risk in any way is already part of
the model. While the licensing basis of the
SSCs are typically well defined, many SSCs may be
employed as mitigating equipment for core damage
sequences not envisioned during the licensing of
the NPP. Typical equipment in this category
includes startup feedwater pumps, fire protection
pumps, and condensate pumps.Thus, an SSC
excluded from the PRA is unlikely to affect a
risk-informed application.
12
Assign SSCs to Partition I Partition II Partition
III
The feature of the SSC being relied on in this
risk-informed application has been previously
reviewed in accordance with NRC guidance.
Presumably, this means that the SSC feature is
already modeled in conformance with the
appropriate Capability Category 2 SRs. The most
common example is RCP Seal Modeling.
13
Assign SSCs to Partition I Partition II Partition
III
A SSC in this partition will be represented by
one of three types.
An initiating event most likely a system level
initiator. It is also possible to affect an
inclusive initiator such as general transient or
loss of offsite power, e.g., changing test
procedures for main steam isolation valves.
A typical basic event representing a failure to
mitigate, e.g., failure to run, failure to open.
A recovery term derived from the HRA.
14
Assign SSCs to Partition I Partition II Partition
III
The PRA model is able to calculate a risk metric
for Partition I SSCs. The question becomes,
which part of the model have to be manipulated to
find the change in risk?
15
Assign SSCs to Partition I Partition II Partition
III
Assign Partition I SSCs to Partition
I-A Partition I-B Partition I-C Partition I-D
16
Assign SSCs to Partition I Partition II Partition
III
Assign Partition I SSCs to Partition
I-A Partition I-B Partition I-C Partition I-D
17
Partition I-A Partition I-B Partition
I-C Partition I-D
The SSC and the failure mode affected by the
application are explicitly modeled. For
example, a modern PRA is likely to well estimate
risk associated with lengthening the surveillance
interval for the emergency diesel generators
(EDGs).
18
Partition I-A Partition I-B Partition
I-C Partition I-D
The SSC is not modeled explicitly, but the model
does include another component that subsumes the
SSC affected by the application. For example,
replacing an electro-mechanical time-delay relay
with a solid state time-delay relay in the EDG
load sequencer.In this case, the PRA analyst
will determine the relative reliability and
availability contribution of the relay to the
overall sequencer and adjust the sequencer
probabilities
19
Partition I-A Partition I-B Partition
I-C Partition I-D
These SSCs have a potential adverse indirect
impact on a modeled component, and the indirect
effect is covered appropriately in the PRA.An
example of a Partition I-C component would be a
DC breaker providing control power to a large AC
switchgear compartment.
20
Partition I-A Partition I-B Partition
I-C Partition I-D
Partition I-D SSCs would be Partition I-C, except
the indirect impact is not modeled. This is
most typically the result of not having a
particular failure mode for an SSC that is
otherwise already in the model.
21
HLRs for the Category I SSCs

• Once the model element is identified, it is a
  straight forward matter to pick the associated
  HLRs, of which there are nine.
• 1. Initiating Event Analysis
• 2. Accident Sequence Analysis
• 3. Success Criteria
• 4. System Analysis
• 5. Human Reliability Analysis
• 6. Data Analysis
• 7. Internal Flooding
• 8. Quantification
• 9. LERF Analysis

22
SRs of the HLRs

• Each of the nine HLRs has a set of SRs. It is
  those SRs that allow a distinction between
  Capability Categories.
• Fortunately, not all of SRs distinguish between
  the three Capability Categories.

23
Capability Category 2 PRA Model for the
Application

• Depending on which Partition I SSCs are in-scope
  for the application, it is possible that the
  application need only conform to Capability
  Category 2 for just some of the HLRs.
• This means, for a given risk-informed
  application, the applicant need only resolve the
  FO list associated with the plant change and the
  risk informed application.

24
PRA Model for an Application

• The peer review will have identified the
  capability category for each ASME SR for the
  model.
• The model changes supporting the application sent
  to the NRC need then only concern the model
  elements that had been only Capability
  Category 1.
• For the NRC to accept the risk-informed
  application, the model elements and the
  associated HLRs need to be Capability Category 2.
  
• The PRA aspect of the submittal then becomes, in
  part, making changes to the model or its
  underlying data (e.g., event probabilities) as
  needed to raise particular SRs to Capability
  Category 2.

A model elementis either an Initiator Mitigati
ng event probability Human Recovery Action /
Operator Action
25
Administrative vs. Technical SRs

• An administrative SR would be one that requires
  documentation of inputs, tying values to
  references, associating HRA values with operator
  interviews, etc.
• A technical SR would be one that would change
  the risk metric calculated for the application.
  It may also require determination of an alternate
  risk metric as a sensitivity study.

26
Administrative SRs

• When the application needs to conform to a
  Capability Category 2 SR and it is an
  administrative SR, Regulatory Guide 1.200
  implies that an alternate (but equivalent)
  activity can be implemented, but the applicant
  would have to justify why the activity is
  appropriate for implementing the SR.

27
Technical SRs

• Raising a model from Capability Category 1 to
  Capability Category 2 in a technical aspect, by
  definition, changes the risk metric of the
  application.

30
25
20
T
15
A
10
5
0
IE
AS
SC
SY
HR
DA
IF
QU
LE
28
Technical SRs

• Implementing specific SRs may raise or lower risk
• For example, switching from a generic data source
  to a plant specific data source (or a Bayesian
  update of generic data with plant events) usually
  results in a net decrease in the risk metric.
• The beneficial effect on risk metrics as a result
  of moving to Capability Category 2 comes at a
  price, for example, more calculations, and more
  gathering of raw data followed by data analysis.

29
Summary

• ASME RA-Sb-2005 has dozens of SRs under nine
  HLRs.
• PRAs are at least Capability Category 1.
• Most Applications involve only a few SSCs.
• Important SSCs are already in the PRA.
• SSCs affect or are affected by just some of the
  HLRs.
• Only some of the SRs distinguish between
  Capability Category 1 and Capability Category 2.
• Only some of the SRs are technical, i.e.,
  implementation changes risk.

30
Assign SSCs to Partition I Partition II Partition
III
Assign SSCs in Partition I to Partition
I-APartition I-B Partition I-C Partition I-D

• Identify SSCs and activities affected by the
  plant change, including cause and effect
  relationship.

Assignment to Partitions I-A to I-D results in a
mapping of a SSC to a model element.

• Identify PRA scope and risk metrics needed to
  evaluate plant changes and to support application.

A model elementis either an Initiator Mitigati
ng event probability Human Recovery Action /
Operator Action
The type of model element isdirectly
correlated toone or more ASME HLRs
31
Conclusion

• Supporting a risk informed application with a
  Capability Category 2 PRA Model is manageable.
• Applications involve few SSCs
• A particular SSC affects some of the HLRs
• A limited number of SRs actually affect the
  risk metric.

32
www.bcpEngineers.com