Ageing probabilistic safety assessment network Recent developments


Nitoi, M.; Rodionov, A.

Progress in Nuclear Energy 56: 71-78

2012


Contents
lists
available
at
SciVerse
ScienceDirect
Progress
in
Nuclear
Energy
journal
homepage:
www.elsevier.com/locate/pnucene
ELSEV
I
ER
2
PROGRESS
IN
NUCLEAR
ENERGY
Ageing
probabilistic
safety
assessment
network
Recent
developments
M.
Nitoi
a
'
*
,
A.
Rodionov
b
a
Safety
of
Present
Nuclear
Reactors
Unit,
European
Commission
joint
Research
Centre,
Institute
for
Energy,
Westerduinweg
3,
1755
LE
Petten,
The
Netherlands
b
Institute
for
Radiological
Protection
and
Nuclear
Safety,
92262
Fontenay-aux-Roses
Cedex,
France
ARTICLE INFO
ABSTRACT
Progress
in
Nuclear
Energy
56
(2012)
71-78
Article
history:
Received
14
June
2011
Received
in
revised
form
7
December
2011
Accepted
13
December
2011
Keywords:
Ageing
Probabilistic
safety
assessment
Network
This
paper
presents
the
recent
results
obtained
in
the
frame
of
Joint
Research
Centre
-
Institute
for
Energy
(JRC-IE)
institutional
project
"Use
of
Probabilistic
Safety
Assessment
for
Evaluation
of
Ageing
Effects".
The
project
started
in
2004,
having
as
participants,
organizations
(utilities,
regulatory
authori-
ties,
research
and
technical
support
organizations)
from
EC
MS,
Swiss,
RF,
Armenia
and
Korea.
The
network
intends
to
contribute
to
the
understanding
and
assessment
of
ageing
effects
on
the
performances
of
the
plants,
and
to
promote
the
use
of
PSA
for
ageing
management
and
for
risk-informed
decisions.
©
2011
Elsevier
Ltd.
All
rights
reserved.
1.
Introduction
The
ageing
phenomena
have
certain
effects
on
any
equipment,
and
as
the
time
is
passing,
it
becomes
highly
likely
that
their
performances
will
be
lower
than
those
expected
at
the
design
phase.
No
installation
could
be
considered
immune
to
the
ageing
effects,
and
to
determine
the
impact
level,
is
necessary
to
assess
the
ageing
effects
on
component
behaviour
and
on
plant
safety
level.
Currently,
activities
related
to
ageing
evaluation
are
performed
in
the
frame
of
the
following
programs:
Periodic
Safety
Review,
Ageing
Management,
Maintenance
Optimization,
Long
Term
Operation
According
to
Power
Reactor
Information
System
(PRIS)
statistics
(PRIS,
2011),
more
than
81%
of
operational
nuclear
reactors
have
more
than
20
years
of
operation,
which
means
that
in
the
next
decade
ageing
management
and
long-term
operation
issues
will
became
one
of
the
key
points
of
nuclear
safety,
and
the
evaluation
of
ageing
effects
on
the
overall
plant
safety
will
become
a
necessity
(Fig.
1).
*
Corresponding
author.
Tel.:
+31
224
56
52
06;
fax:
+31
224
56
56
37.
E-mail
addresses:
Mirela.NIT01@ec.europa.eu
,
mirela.nitoi@nuclear.ro
(M.
Nitoi),
andreixodionov@irsn.fr
(A.
Rodionov).
0149-1970/$
see
front
matter
©
2011
Elsevier
Ltd.
All
rights
reserved.
doi:10.1016/j.pnucene.2011.12.006
The
basic
concern
to
use
the
Probabilistic
Safety
Assessment
(PSA)
for
ageing
evaluation
came
from
the
requirement
to
accom-
plish
the
safety
goals
during
the
whole
life
cycle
of
the
nuclear
installation
(including
the
extended
lifetime).
In
probabilistic
terms,
INSAG12
(IAEA,
1999)
specified
a
safety
goal
as
follows:
"The
target
for
existing
nuclear
power
plants
consistent
with
the
technical
safety
objective
is
a
frequency
of
occurrence
of
severe
core
damage
that
is
below
about
10
-4
events
per
plant
operating
year.
Severe
accident
management
and
mitigation
measures
could
reduce
by
a
factor
of
at
least
ten
the
probability
of
large
off-site
releases
requiring
short
term
off-site
response".
For
the
units
which
are
approaching
the
end
of
initial
design
lifetime
and
especially
for
those
which
are
planning
to
extend
the
lifetime,
it
has
to
be
demonstrated
that
the
plant
safety
level
will
remain
in
accordance
with
this
target
until
the
end
of
operation,
and
to
do
that,
is
necessary
to
evaluate
the
effects
of
ageing
phenomena
on
the
plant
performance.
In
the
last
years,
the
PSA
tools
have
been
reached
a
certain
level
of
maturity,
and
it
was
acknowledged
that
PSA
results
can
bring
important
issues
to
successfully
complement
deterministic
anal-
ysis.
As
together
with
maintaining
the
established
safety
goals,
it
is
necessary
to
prioritize
the
actions
dedicated
to
Ageing
Manage-
ment
or
Long
Term
Operation,
the
results
of
PSA
could
be
used
successfully
in
this
area.
The
effects
of
ageing
could
be
multiple
and
variable.
On
system
availability
level,
ageing
could
induce
the
modification
of
system
success
criteria,
could
increase
the
CCF
probability
for
highly
redundant
systems,
and
could
change
the
list
of
contributors
to
overall
system
unavailability.
On
overall
plant
level,
ageing
could
2
22
21
16
14
14
3
_
12
11
19
7
72
M.
Nitoi,
A.
Rodionov
/
Progress
in
Nuclear
Energy
56
(2012)
71-78
Number
of
Operating
Reactors
by
Age
35
30
25
20
15
10
6
6
5
5
5
0
ii
°
jijil
-'3
1
ii•
-
•1' rJ
Note
Age
of
a
reactor
is
determined
by
its
first
grid
connection.
Fig.
1.
Ageing
profile
of
nuclear
operating
reactors.
6 6
5
J
iilliiiiiiiiiiiil
i
o2
0
1
2
3
4
5
6
7
0
9
10
1
12
13
4
5
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
30
39
40
41
42
43
[year]
induce
the
modification
of
initiating
event
frequencies,
probability
of
mitigation
of
undesired
events
(probability
of
unavailability
for
safety
systems),
the
occurrence
of
inter-systems
CCF,
the
dominant
sequences,
and
could
change
the
list
of
major
contributors
to
the
accident
sequences.
Without
the
evaluation
of
ageing
effects
based
on
probabilistic
assessment,
study
of
sustainability
of
a
nuclear
generation
fleet
in
the
next
decade
will
not
be
completely,
and
safety
authorities
could
emit
doubt
on
the
lifetime
extension
for
any
facility.
2.
Ageing
PSA
network
The
initiation
of
the
JRC
Institute
of
Energy
(IE)
project
"Use
of
Probabilistic
Safety
Assessment
for
Evaluation
of
Ageing
Effects"
(APSA),
has
been
induced
by
the
fact
that
current
standard
PSA
tools
do
not
adequately
address
important
ageing
issues,
which
could
have
a
significant
impact
on
the
conclusions
made
from
PSA
studies
and
applications,
especially
in
case
of
plants
which
operate
in
advanced
aged
conditions.
2.1.
Objective
The
JRC-IE
project
has
established
the
bases
for
Ageing
Proba-
bilistic
Safety
Assessment
(APSA)
Network.
(Patrik
and
Kirchsteiger,
2004)
Main
objective
of
the
APSA
Network
is
to
use
common
resources
of
Network
participants
for
identification,
development
and
demonstration
of
methods
and
approaches
which
could
help
PSA
developers
and
users
for
the
following
activities:
to
investigate
and
to
evaluate
the
effects
that
ageing
phenomena
could
induce
on
the
plant
performance,
to
incorporate
the
effects
of
equipment
ageing
into
current
PSA
models
to
perform
engineering
analysis,
to
provide
the
necessary
support
for
identification
and
priori-
tization
of
reliability
monitoring
actions/approaches
to
assure
that
potential
decreasing
of
reliability
of
SSC
would
be
identi-
fied
and
corrected
in
time,
to
promote
the
use
of
PSA
for
ageing
management
and
risk-
informed
applications
of
Nuclear
Power
Plants
in
Long
Term
Operation
(LTO).
22.
Network
participants
Any
organization
could
become
member
of
the
APSA
Network
if
posses
a
certain
experience
in
the
field
and
allocated
resources
for
PSA
applications
on
ageing
assessment
and
system
reliability
optimization
approaches.
Another
condition
for
participation
is
to
have
the
good
will
to
share
this
experience
and
the
results
with
partners
and
to
undertake
new
initiatives
in
this
particular
field.
The
Network
partners
are
experts
from
European
and
non-
European
Institutions
(utilities,
regulatory
authorities,
research
and
scientific
organizations).
The
network
structure
is
presented
in
Fig.
2.
The
APSA
Network
uses
two
kinds
of
contributions:
in-kind
contributions,
when
the
costs
related
to
carrying
out
the
Network
activities
are
bear
by
the
partner
organization
the
activities
performed
under
contracted
work
(case
when
agreement
on
financial
terms
exists)
23.
Network
achievements
The
main
tasks
of
the
APSA
Network
are
strongly
related
to
the
deterministic
ageing
assessments
tasks
(ageing
management,
long
time
operation,
maintenance
optimization),
and
even
intermediate
APSA
tasks
results
could
be
used
in
mentioned
deterministic
assessments
(trend
analysis,
system
reliability
analysis).
The
Network
participants
perform
active
research
in
the
following
area:
-
selection
of
aged
components
-
modelling
the
aged
components
(active
and
passive)
in
PSA
studies
-
predictive
evaluations
(sensitivity
and
uncertainty
analysis)
APSA
model
can
be
used
for
prioritization
of
ageing
issues
and
for
evaluation
of
risk
profile
of
the
plant
(predictive
evaluation).
Investigating
the
possibility
to
develop
an
APSA
model
given
the
standard
PSA
model,
some
differences
between
the
models
were
found
and
they
highlighted
the
supplementary
features
that
an
APSA
model
should
have.
These
issues
are
listed
below:
M.
Nitoi,
A.
Rodionov
/
Progress
in
Nuclear
Energy
56
(2012)
71-78
73
Steering
Committee
Network
Coordinator
National/
international
institutions/
networks/
projects
JRC-IE
Operating
agent
TASK
1
Network
management
-
r
T
-11-
-
n
TASK
2
leader
TASK
3
leader
TASK
4
leader
TASK
5
leader
TASK
6
leader
TASK
7
leader
TASK
8
leader
Working
Group
2
Working
Group
Working
Group
Working
Group
Working
Group
Working
Group
3
4
5
6
7
Working
Group
Analysis
of
main
8
PSA
tasks
with
Selection
of
the
Reliability
and
Investigation
of
Reliability
and
Incorporation
of
regards
to
APSA
SSC
to
be
considered
in
data
analysis
for
active
ageing
impact
on
Common
data
analysis
for
passive
age-depended
reliability
Applications
of
APSA
results
APSA
components
Cause
Failures
components
parameters
and
data
into
PSA
model
Fig.
2.
APSA
network
structure.
while
a
standard
PSA
assumes
that
component
failure
rates
are
constant,
APSA
should
explicitly
models
ageing
effects
in
component
failure
rates,
which
generally
cause
the
failure
rates
to
increase
with
age
APSA
should
explicitly
model
the
effects
of
test
and
mainte-
nances
in
controlling
the
ageing
of
components
Standard
PSA
neglects
the
components
that
have
small
failure
probabilities,
not
taken
into
account
the
fact
that
these
prob-
abilities
could
suffer
dramatic
changes
in
time;
passive
components
are
typically
not
included
in
the
analysis,
while
in
APSA
model
they
should
be
taken
into
account
A
standard
PSA
calculates
constant
values
for
the
core
damage
frequency
and
systems
unavailability;
APSA
could
explicitly
calculate
the
ageing
effects
and
age
dependence
on
the
core
damage
frequency,
initiating
event
frequency
and
systems
unavailability
After
considering
the
possibility
to
incorporate
the
ageing
effects
in
PSA
models,
it
was
concluded
after
discussions
that
the
following
issues
should
be
considered:
system
fault
trees
need
to
be
developed
in
sufficient
level
of
detail,
to
permit
the
modelling
of
ageing
effects
component
unavailability
should
include
age-dependent
failure
contributors
the
impact
of
ageing
on
CCF
should
be
investigated
and
included
in
CCF
parameters
the
passive
component
failures
should
be
modelled
and
included
in
system
logic
model
component
test
and
maintenance
models
should
include
the
effects
of
test
and
maintenance
in
terms
of
"good
as
old"
and
"good
as
new"
cases,
and
should
include
replacements
and
renewals
of
components
any
additional
dependency
caused
by
age
related
degradation
of
components
and
structures
should
be
included
in
the
model
appropriate
computer
packages
should
be
developed
and
used
Specific
activities
of
the
Network
are
presented
below.
2.3.1.
Data
collection
aspects
Providing
more
information
than
the
standard
PSA,
it
is
normal
that
APSA
requires
more
data
and
more
extended
models
than
the
standard
PSA.
The
availability
of
data
necessary
to
develop
an
ageing
PSA
has
required
a
specific
attention
from
the
project
participants.
It
has
to
be
recognized
that
any
data
collection
process
is
resources
consuming,
mainly
because
it
involves
a
considerable
effort
to
process
data
available
in
operational
logs
to
the
format
needed
for
parameters
estimation.
Moreover,
in
order
to
perform
an
age-dependent
reliability
analysis,
additional
data
have
to
be
obtained.
At
the
beginning
of
research,
data
necessary
for
developing
a
time-dependent
reliability
model
could
be
specified,
but
some
issues
were
not
clear,
as
the
following:
if
those
data
are
available
in
any
of
the
listed
sources;
how
easy
it
will
be
to
extract
and
to
process
them;
what
will
be
the
quality/completeness
of
obtained
information.
To
answer
to
these
questions
it
was
performed
a
study
among
the
project
participants.
The
study
was
focused
on
data
availability
and
accessibility,
as
well
as
on
finding
possible
ways
to
improve
data
collection.
(Rodionov,
2008)
During
discussions,
it
was
specified
that
the
most
important
information
is
collected
during
parameter's
elaboration,
normally
documented
in
Initiating
Event
frequencies
and
Component
Reli-
ability
Parameters
Evaluation
task
reports
and/or
databases.
Pro-
cessed
data
about
failures
and
component
performance
are
usually
well
structured
and
have
a
high
quality,
and
these
data
could
be
certainly
used
for
age-dependent
reliability
analysis.
Raw
data
sources
Operating
(defect)
and
maintenance
logs
Abnormal
Operational
Events
(LER)
Reporting
System
Operating
and
maintenance
procedures
Design,
commissioning
and
manufactory
information
zP
3
74
M.
Nitoi,
A.
Rodionov
/
Progress
in
Nuclear
Energy
56
(2012)
71-78
Answering
to
the
questionnaire,
the
project
participants
proposed
additional
data
categories
considered
as
needed
for
ageing
evaluation.
Some
of
these
categories
have
normally
to
be
addressed
in
PSA
reliability
data
analysis,
as
for
example:
for
each
component
failure,
information
on
failure
mode,
mechanism
and
cause;
component
type
(active
or
passive
components);
type
of
failure
(degradation
or
critical);
information
about
common
cause
failure
mode
(CCF-
candidate).
Others
proposed
categories
were
more
specific
to
the
ageing
characterization
and
surveillance
activities,
as
the
following:
component
life
cycle
or
ageing
management
strategy,
information
on
component
degradation
mechanisms,
information
on
component
failure
precursors
(maybe
these
occurred
before
the
component
really
failed
as
a
result
of
preventive
inspection/maintenance),
results
of
non-destructive
tests
for
passive
components,
plant
history
of
component
failures
of
all
component
types,
including
date
of
failures,
relation
between
random
failures
and
ageing
dependent
failures,
number
of
failures
per
component
on
one-year
interval.
It
was
concluded
that
for
age-dependent
reliability
analysis
we
could
use
three
types
of
data
sources:
PSA
reliability
database,
other
reliability
databases
(good
if
is
necessary
to
enlarge
the
statistic
for
a
component,
using
data
from
similar
components),
raw
data
sour-
ces
(operational,
maintenance
and
design
documentation)
(Fig.
3).
Taking
into
account
the
variety
of
time-dependent
reliability
models
(Atwood
et
al.,
2007;
Antonov
et
al.,
2008)
proposed
in
recent
publications,
the
following
additional
data
categories
were
proposed
for
evaluation:
a
Component
commissioning
date
(age
considered
as
0)
b
Failure/censoring
times
(age
in
the
moment
of
failure/
censoring)
c
Component
replacement
date
and
cause
of
replacement
(corrective
maintenance,
preventive
maintenance,
modifica-
tions/design
changes)
PSA
Reliability
DB
IE
frequencies
Component
reliability
parameters
Processed
component's
failure,
performance
and
maintenance
data
Processed
Operational
Events
data
Generic
data
(Rel
ability
parameters
and
IE
frequencies)
Others
reliability
DBs
(vendors
DB,
NDE
of
piping
systems
DB,
l&C
elements
DB,
etc.)
Fig.
3.
Potential
data
sources
for
age-dependent
reliability
analysis.
d
Characteristics
of
applied
tests
and
maintenance
strategy
type
and
frequency
e
Degree
of
component
renewal
during
the
maintenance
(corrective
maintenance,
preventive
maintenance)
f
Component
lifetime
(design/manufacture
specification,
quali-
fication
tests
results)
g
real
cumulated
number
of
hours
in
operation,
number
of
demands
h
information
about
the
average
and
extreme
levels
of
the
operating
stressors
(pressure,
temperature,
mechanical
loads,
vibration,
water
chemistry,
neutron
flow,
current
intensity,
frequency,
voltage,
etc.)
i
information
about
the
average
and
extreme
levels
of
environ-
mental
stressors
(pressure,
temperature,
humidity,
neutron
flow,
etc.)
For
each
possible
time-dependent
model
that
could
be
devel-
oped,
the
necessary
information
was
identified,
as
mentioned
below:
Case
1
simple
age-dependent
reliability
model
or
trend
analysis.
For
this
model,
the
data
needed
falls
into
categories
a—c.
Case
2
age-dependent
reliability
models
including
test
and
maintenance
evaluations.
This
type
of
model
requires
data
supplementary
to
case
1,
as
test
and
maintenance
information
(data
characteristic
to
category
d),
and
information
about
renewal
of
component
during
maintenance
(category
e)
Case
3
comprehensive
age-dependent
reliability
models.
Types
of
data
needed
in
this
case
are
the
data
for
Case
1
and
Case
2,
and
supplementary
data
about
component
lifetime
(category
f),
cumulated
number
of
hours
in
operation/number
of
demands
(category
g),
information
about
stressors
(category
h
for
operational
stressors
and
category
i
for
environmental
stressors)
From
applications
point
of
view,
availability
and
accessibility
of
data
for
different
types
of
reliability
models
were
evaluated
during
the
survey
and
the
results
are
shown
on
the
Fig.
4.
The
results
demonstrated
that
even
for
simple
age-dependent
reliability
assessments
for
which
most
of
the
data
are
available,
the
cost
of
data
processing
could
be
quite
high,
because
not
all
the
data
are
easily
accessible.
If
is
intended
to
apply
reliability
models
for
maintenance
analysis
and
optimization,
or
for
lifetime
evalua-
tion
and
prediction,
a
large
additional
investment
for
data
collec-
tion
and
processing
have
to
be
envisaged.
This
has
to
be
taken
into
account
as
for
further
Ageing
PSA
activities
on
models
development,
as
for
the
specification
of
addi-
tional
efforts
needed
in
data
collection.
Availability
and
accessibility
of
data
for
different
types
of
reliability
models
Case
1
Case
2
Case
3
K1
-
availability
K2
-
accessibility
Fig.
4.
Availability
and
accessibility
of
data
for
different
types
of
reliability
models.
100%
-
80%
60%
40%
-
20%
0%
M.
Nitoi,
A.
Rodionov
/
Progress
in
Nuclear
Energy
56
(2012)
71-78
75
2.32.
SSC
selection
aspects
Selection
of
the
SSC
represents
another
area
of
interest
for
Network
activities.
To
evaluate
the
ageing
impact
on
all
components
of
the
plant
and
to
incorporate
these
effects
into
PSA
studies
requires
a
huge
effort,
but
fortunately
it
was
demonstrated
that
the
process
is
enough
to
be
performed
only
for
the
key
components,
which
are
both
sensitive
to
ageing
and
important
for
plant
safety.
For
identification
of
these
key
ageing
SSC,
their
risk
significance
and
their
sensitivity
to
ageing
should
be
evaluated
(taking
into
account
the
risk
implications,
the
various
stressors
contributing
to
performances
degradation,
the
different
ageing
mechanisms
and
the
available
measures
for
detection
of
ageing
degradation
and
for
mitigation).
The
guideline
(Nitoi
and
Rodionov,
2010)
developed
in
the
frame
of
the
APSA
Network
provides
a
practical
approach
and
makes
recommendations
about
the
methods
that
could
be
used
in
selection/prioritization
of
SSC
sensitive
to
ageing
and
important
from
risk
point
of
view
in
operating
nuclear
power
plants.
The
methods
suitable
for
selection
were
briefly
presented,
together
with
their
advantages
and
disadvantages.
It
was
developed
a
list
of
ageing
mechanisms
susceptible
to
appear,
the
factors
favourable
for
their
occurrence
and
related
sensitive
materials
to
ageing
mechanisms.
The
guideline
proposes
to
use
a
combined
approach
for
the
selection
of
SSC,
using
the
trend
analysis,
qualitative
analysis,
PSA
importance
measures
and
expert
judgements.
Proposed
steps
of
selection
process
are
presented
in
the
Fig.
5.
After
carefully
consideration
of
the
selection
issues,
it
was
decided
that
SSC
need
be
evaluated
from
two
points
of
view:
susceptibility
to
ageing
and
importance
from
risk
point
of
view.
The
analysis
of
SSC
susceptibility
to
ageing
is
necessary,
considering
that
there
is
no
reason
to
model
the
time-behaviour
of
SSC
for
Ageing
PSA
purposes
if
they
do
not
have
such
vulnerability.
To
identify
the
SSC
potentially
sensitive
to
ageing,
trend
analysis
of
available
reliability
data,
or
qualitative
assessment
could
be
used.
Qualitative
assessment
could
cover
SSC
which
were
initially
neglected
in
PSA
models
and
SSC
without
enough
failure
statistics
to
perform
trend
analysis.
A
qualitative
method
for
identification
of
SSC
sensitive
to
ageing
Ageing
Failure
Modes
and
Effect
Analysis
(AFMEA)
was
developed,
and
data
and
information
sources
that
could
be
useful
for
the
analysis
were
specified.
The
applicability
of
the
methods
proposed
for
selection
of
SSC
susceptible
to
ageing
was
demon-
strated
by
results
of
a
case
study
which
had
used
as
an
example
I
Colecting
information
and
creating
SSC
list
Evaluation
of
SSC
risk
importance
Evaluation
of
SSC
susceptibility
to
ageing
Trend
analysis
Qualitative
analysis
List
of
SSC
important
from
risk
point
of
view
List
of
SSC
which
show
age
tendency
in
data
List
of
SSC
which
are
considered
susceptible
to
ageing
SC
selection
&
prioritizaton
List
of
prioritized
SSC
important
for
APSA
C
Final
report
Fig.
5.
Selection
procedure
for
SSC
which
are
risk
important
and
ageing
vulnerable.
PSA
Incorporation
into
PSA
Statistical
Tests
Detailed
Ana
ysis
Parameter
Estimation
Data
Collection
and
Analysis
Visual
Inspection
Nonparametric
Tests
No
line
trends
bservable?
vig
Selection
of
Best
Pert°
ming
Model
Sensitivity
Study
with
existing
PSA
model
No-
Documentation
of
Sensitivity
Study
SA
results
affected?
Yes
Documentation
of
Analysis
results
Final
Report
Yes
Analysis
of
Homogenity
of
Data
Sets
76
M.
Nitoi,
A.
Rodionov
/
Progress
in
Nuclear
Energy
56
(2012)
71-78
Table
'1
Decision
table.
Risk
importance
results
2
3
AFMEA
results
LOW
1LOW
2LOW
3LOW
MEDIUM
1MEDIUM
2MEDIUM
3MEDIUM
HIGH
1HIGH
2HIGH
3HIGH
systems
of
Institute
for
Nuclear
Research
TRIGA
research
reactor.
(Nitoi
and
Rodionov,
2008)
To
identify
the
SSC
which
are
modelled
in
PSA
and
are
important
from
risk
point
of
view,
the
Risk
Importance
Measures
were
used.
After
each
of
these
analyses,
attributes
concerning
risk
signifi-
cance
and
ageing
sensitivity
were
allocated
to
each
SSC.
The
next
step
of
the
analysis
was
to
prioritize
the
previously
obtained
results,
using
defined
prioritization
criteria.
It
was
developed
a
decision
table
(Table
1),
with
the
purpose
of
ranking
the
components
which
are
both
susceptible
to
ageing
and
risk-
significant.
If
is
necessary,
some
weighed
criteria
could
be
used
for
prioritization.
The
selection
procedure
allows
the
identification
of
specific
types
of
components:
components
which
are
vulnerable
to
ageing
(this
is
revealed
by
the
presence
of
ageing
trends
in
data
or
by
the
results
of
qualitative
analysis)
Fig.
6.
Investigation
of
time
-
dependent
trends.
WinBUGS
constant
Uncertainty
range
of
real
data
WinBUGS
Loglinear
4
Frequentist
estimate
of
real
data
0.000014
0.000012
0.00001
0.000008
0.000006
0.000004
0.000002
18
20
22
24
26
28
0
16
M.
Nitoi,
A.
Rodionov
/
Progress
in
Nuclear
Energy
56
(2012)
71-78
77
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
16
.....
i•m••••
.....
18
20
22
24
26
28
—WinBUGS
constant
—WinBUGS
Probit
—WinBUGS
Cloglog
—WinBUGS
Logit
Uncertainty
range
of
real
data
Frequentist
estimate
of
real
data
Fig
7.
Results
for
DG-FS.
.
components
which
are
important
from
safety
point
of
view
(identified
by
the
values
of
risk
importance
factors)
.
components
for
which
is
necessary
to
develop
time-dependent
reliability
models
2.33.
Data
analysis
aspects
A
guideline
for
reliability
and
data
analysis
for
active
compo-
nents
was
developed
in
the
frame
of
the
Network
activities.
The
guideline
provides
practical
methods
for
analyzing
component
and
system
reliability
data,
focussing
on
the
identification
and
model-
ling
of
ageing.
(Rodionov
et
al.,
2009)
The
emphasis
was
on
fre-
quentist
and
Bayesian
approaches,
implemented
with
MS
EXEL
and
the
open-source
software
package
WinBUGS,
but
the
methods
described
can
be
implemented
with
other
software
packages.
Fig.
6
shows
the
steps
for
the
investigation
of
time-dependent
trends
in
failure
data
of
active
components
(starting
from
data
collection
and
analysis
up
to
incorporation
of
the
results
into
a
plant
specific
PSA).
In
case
of
data
collection
for
parameters
having
time
depen-
dence
(relevant
raw
information
obtainable
from
operational
and
maintenance
records),
it
is
important
to
obtain
more
accurate
information
on
the
life
history
of
the
investigated
components
(date
of
commission,
dates
of
replacement,
dates
of
complete
overhaul
work)
and
on
the
time
of
occurrence
of
failure
events.
A
qualitative
analysis
is
necessary
to
be
performed,
to
check
for
the
presence
of
time
trends.
If
nonparametric
tests
do
not
show
any
notable
time
trend
the
investigation
can
be
stopped.
If
a
significant
time
trend
is
suggested
by
qualitative
analysis
or
nonparametric
test,
then
a
more
detailed
analysis
is
needed.
During
this
detailed
analysis
the
presence
of
the
trend
will
be
confirmed
and
a
model
from
which
future
failure
events
can
be
predicted
will
be
developed.
The
model
will
permit
an
accurate
quantitative
description
of
the
time
trend.
This
is
especially
important
with
respect
to
ageing
effects,
as
the
capability
to
predict
future
failure
events
can
be
important
for
other
applications,
for
example,
for
planning
long
term
maintenance
work
and
investments.
The
quantitative
estimation
and
prediction
model
should
be
selected
on
the
basis
of
statistical
tests
measuring
the
performance
of
different
model
alternatives.
The
sensitivity
analysis
is
aimed
to
study
the
effect
of
time
trends
of
component
failure
rates
on
the
overall
PSA
results.
If
the
impact
of
time
dependency
on
CDF
or
LERF
is
small,
the
investi-
gation
can
be
stopped
by
documenting
the
results
in
a
final
report.
If
the
impact
of
time
dependency
is
not
negligible
(a
typical
criterion
is
a
change
of
the
mean
core
damage
frequency
by
more
Fig.
8.
Results
for
6
kV
CB
FC.
78
M.
Nitoi,
A.
Rodionov
/
Progress
in
Nuclear
Energy
56
(2012)
71-78
than
20%)
then
it
is
recommended
to
implement
the
results
of
the
investigation
directly
in
the
plant
specific
PSA.
The
more
detailed
predictive
models
are
then
used
to
calculate
the
time-dependent
mean
values
for
failure
rates
to
be
used
in
the
PSA
model.
A
case
study
that
used
operational
data
from
VVER-440
reactors
(Poghosyan
et
al.,
2010)
was
performed
to
demonstrate
the
appli-
cability
of
methods
for
reliability
parameters
estimation
presented
in
the
guideline.
The
case
study
was
performed
in
the
following
steps:
Selection
of
representative
components
for
further
assessment
Data
collection
from
similar
NPP
Preliminary
data
analysis
aimed
to
reveal
trend
model
Detailed
analysis
aimed
to
go
in
details
when
increase
trend
is
present
Data
were
collected
from
similar
VVER-440
plants
(from
oper-
ating
reactors
in
Czech
Republic,
Hungary,
Slovakia,
Republic
of
Armenia).
For
illustrating
the
work,
the
results
obtained
in
case
of
Diesel
Generators
and
6
kV
Circuit
Breakers
are
presented
in
the
Figs.
7
and
8.
From
the
components
analyzed
(using
two
methods,
Excel
functions
and
WinBUGS
scripts),
only
data
for
6
kV
circuit
breakers
have
shown
an
ageing
tendency,
so
it
was
decided
that
these
components
will
require
a
time-dependent
model
to
be
incorpo-
rated
in
PSA
studies.
Other
case
studies
performed
in
the
frame
of
the
Network
applied
the
statistical
evaluation
methods
of
ageing
trend
(Atwood
et
al.,
2007)
and
inversion
criteria
test
for
I&C
and
electrical
components.
Several
cases
of
Generalized
Linear
Model
were
also
proposed
and
investigated
using
continues
and
discrete
data
(Antonov
et
al.,
2008).
The
Fisher
Chit
minimization
approach
was
applied
for
goodness
of
fit
test
and
parameters
elaboration,
and
uncertainty
analysis
was
done
for
estimated
parameters
and
model
extrapola-
tions.
The
results
were
analyzed
and
compared
with
other
approaches.
The
results
obtained
in
frame
of
this
task
were
disseminated
in
a
training
program,
aiming
to
introduce
the
methods
and
approaches
for
Advanced
Time-dependent
Reliability
Data
Analysis
which
could
be
used
in
NPP
ageing
effects
evaluation
and
ageing
management.
The
training
included
lessons
and
many
practical
exercises
on
computation
techniques
using
available
software
tools
and
data
examples.
The
research
work
of
this
task
is
dedicated
to
develop
methods
to
elaborate
the
reliability
parameters
for
Ageing
PSA
model
and
to
demonstrate
their
validity
by
performing
case
studies.
It
is
ex-
pected
that
the
results
will
be
useful
for
the
following
activities:
choosing
the
appropriate
reliability
model
for
the
parameters
estimation,
addressing
ageing
and
maintenance
effects
in
component
failure
models,
improvement
of
reliability
and
maintenance
data
collection
system
to
fit
with
specific
requirements
of
age
and
mainte-
nance
dependent
reliability
models
for
the
purpose
of
APSA.
3.
Conclusions
The
resulting
knowledge
from
the
project
running
should
help
PSA
developers
and
users
to
efficiently
incorporate
the
effects
of
equipment
ageing
into
current
PSA
tools
and
models.
Summarizing,
the
APSA
expected
results
are
the
following:
O
Contributions
to
a
better
understanding
of
important
issues
in
modelling
of
ageing
phenomena
using
PSA
models;
O
Developed
set
of
feasible
approaches/models
and
meth-
odological
guidelines;
O
Proved
feasibility
of
the
proposed
approaches/models
by
case
studies
results;
O
Provided
support
&
training
for
correct
application
of
the
project
methodological
guidelines.
Practical
applications
that
use
the
addressing
of
ageing
effects
in
PSA
could
be
related
to:
O
prioritization
of
ageing
management
issues,
LTO
activities
using
APSA
findings;
O
predictive
evaluation
of
plant
safety
level;
O
prioritization
of
maintenance
activities
using
APSA
results.
Acknowledgement
The
authors
wish
to
thank
to
all
APSA
partners
for
their
work
and
commitment.
References
Antonov,
A.,
Chepurko,
V.,
Polyakov,
A.,
Rodionov,
A.,
2008.
A
Case
Study
on
Investigation
of
Component
Age
Dependent
Reliability
Models
JRC
IE
Report,
EUR
23079
EN,
Petten,
Netherlands.
Atwood,
C.,
Cronval,
0.,
Patrik,
M.,
Rodionov,
A.,
2007.
Models
and
Data
Used
for
Assessing
the
Ageing
of
Systems,
Structures
and
Components
JRC
IE
Report,
EUR
22483
EN,
May,
Petten,
Netherlands.
IAEA,
1999.
INSAG-12,
Basic
Safety
Principles
for
Nuclear
Power
Plants
75-INSAG-3
Rev.1
Vienna,
Austria.
Nitoi,
M.,
Rodionov,
A.,
2010.
Guideline
for
Selection
of
Systems,
Structures
and
Components
to
Be
Considered
in
Ageing
PSA,
ISBN
978-92-79-16496-5
JRC
IE
Report,
EUR
24503
EN,
doi:10.2790/21975,
Petten,
Netherlands.
Nitoi,
M.,
Rodionov,
A.,
2008.
Qualitative
Approach
for
Selection
of
Systems
Struc-
tures
and
Components
to
Be
Considered
in
Ageing
PSA
JRC
IE
Report,
EUR
23446
EN,
ISSN
1018-5593,
Petten,
Netherlands.
Patrik,
M.,
Kirchsteiger,
Ch,
2004.
Incorporating
Ageing
Effects
into
Probabilistic
Safety
Assessment
Applications
JRC
IE
Report,
EUR
21204
EN,
July,
Petten,
Netherlands.
Poghosyan,
S.,
Malkhasyan,
A.,
Rodionov,
A.,
Nitoi,
M.,
2010.
Analysis
of
Data
Related
to
Ageing
of
VVER-440
NPP
Components
JRC
Institute
for
Energy
Report,
ISSN
1018-5593,
Petten,
Netherlands.
PRIS.
2011.
PRIS
website,
http://www.iaea.orgicgi-binidb.page.plipris.agereac.htm.
Rodionov
,
A.,
Kelly,
D.,
Klugel,
J.,
2009.
Guidelines
for
Analysis
of
Data
Related
to
Ageing
of
Nuclear
Power
Plant
Components
and
Systems
JRC
IE
Report,
EUR
23954
EN,
Petten,
Netherlands.
Rodionov,
A.,
2008.
Overview
of
NPPs
Component
Reliability
Data
Collection
with
Regards
to
Time-dependent
Reliability
Analysis
Applications
JRC
IE
Report,
EUR
23084
EN,
Petten,
Netherlands.