Transcript Reliability Block Diagram Modeling – A Comparison of the

Reliability Block Diagram Modeling –
A Comparison of Three Software Packages
Aron Brall, SRS Technologies, Mission Support Division
William Hagen, Ford Motor Company, Powertrain Manufacturing Engineering
Hung Tran, SRS Technologies, Mission Support Division
1
THE SOFTWARE PACKAGES - 1

ARINC RAPTOR 7.0.07
 From RAPTOR web site:
 “Raptor is a software tool that simulates the operations
of any system.”
 “Sophisticated Monte Carlo simulation algorithms are
used to achieve these results.”
 Our Take:
 Pure Monte Carlo simulation tool to solve reliability
block diagrams.
2007 RAMS – Brall, Hagen, Tran
2
THE SOFTWARE PACKAGES - 2

Reliasoft BlockSim 6.5.2
 From BlockSim web site:
 “Flexible Reliability Block Diagram (RBD) creation.”
 “Exact reliability results/plots and optimum reliability
allocation.”
 “Repairable system analysis via simulation (reliability,
maintainability, availability) plus throughput, life cycle
cost and related analyses.”
 Our Take:
 Monte Carlo simulation with algorithms used to speed
the processing time.
 Also provides analytical calculation of reliability.
2007 RAMS – Brall, Hagen, Tran
3
THE SOFTWARE PACKAGES - 3

Relex Reliability Block Diagram
 From Relex web site:
 “At the core of Relex RBD is a highly intelligent
computational engine.”
 “First, each diagram is analyzed to determine the best
approach for problem solving using pure analytical solutions,
simulation, or a combination of both.”
 “Once a methodology is determined, the powerful Relex RBD
calculations are engaged to produce fast, accurate results.”
 Our Take:
 Relex RBD appears to be a hybrid tool that uses algorithms
and simulation in varying combinations to solve reliability
block diagrams.
2007 RAMS – Brall, Hagen, Tran
4
Why Compare Reliability Software
Analysts (especially new analysts) tend to report reliability
software results as exact values
 Engineering judgment, caution and experience are being
supplanted by software analysis
 Error checking is often absent
 Number of runs; confidence limits; garbage in, garbage out all
impact value of software analysis

2007 RAMS – Brall, Hagen, Tran
5
One Block Model
Block
Block
Parameter
Failure
a
Distribution
Repair
Distribution
a
Probability
Distribution
Weibull
Lognormal
Parameter 1 Parameter 2
Shape 1.5
Scale 1000
Mu 5
Sigma 0.5
2007 RAMS – Brall, Hagen, Tran
6
Simple Model
Block
Name
Sta rt
En d
1 ::1
m
F a ilu r e : W e ib u l
Ch a r . L if e : 1 0 0 0
Sh a p e F a c t . : 2
t 0: 0
Q t y: 1
R: 0 . 9 9 0 0 5
b
F a ilu r e : No r m a l
M e an : 25 0
St d De v : 5 0
Q t y: 1
R: 0 . 9 9 8 6 5
d
F a ilu r e : L o g No r m a l
M u: 6
Sig m a : 2
Q t y: 1
R: 0 . 7 5 7 2 2 8
c
M T BF : 1 0 0 0 0
Q t y: 1
R: 0 . 9 9 0 0 5
1 ::1
g
M T BF : 1 0 0 0 0
Q t y: 1
R: 0 . 9 9 0 0 5
1 ::1
e
F a ilu r e : W e ib u l
Ch a r . L if e : 2 3 0 0
Sh a p e F a c t . : 1 . 5
t 0: 0
Q t y: 1
R: 0 . 9 9 0 9 7 5
1 ::2
n
F a ilu r e : W e ib u l
Ch a r . L if e : 1 0 0 0
Sh a p e F a c t . : 3
t 0: 0
Q t y: 1
R: 0 . 9 9 9
1 ::1
i
F a ilu r e : W e ib u l
Ch a r . L if e : 1 0 0 0
Sh a p e F a c t . : 1 . 5
t 0: 0
Q t y: 1
R: 0 . 9 6 8 8 7 2
Parameter
1
Parameter
2
a
Weibull
Shape 1.5
Scale 1000
b
Normal
Mean 250
Std Dev 50
c
Exponential
10000
0
d
Lognormal
Mu 6
Sigma 2
e
Weibull
Shape 1.5
Scale 2300
f
Normal
Mean 250
Std Dev 50
g
Exponential
10000
0
h
Lognormal
Mu 8
Sigma 1
i
Weibull
Shape 1.5
Scale 1000
j
Normal
Mean 250
Std Dev 50
k
Exponential
10000
0
l
Lognormal
Mu 8
Sigma 3
m
Weibull
Shape 2.0
Scale 1000
n
Weibull
Shape 3.0
Scale 1000
o
Weibull
Shape 4.0
Scale 1000
p
Weibull
Shape 0.5
Scale 1000
q
Weibull
Shape 0.4
Scale 1000
3 ::6
f
F a ilu r e : No r m a l
M e an : 25 0
St d De v : 5 0
Q t y: 1
R: 0 . 9 9 8 6 5
h
F a ilu r e : L o g No r m a l
M u: 8
Sig m a : 1
Q t y: 1
R: 0 . 9 9 9 6 5 7
Failure
Distribution
l
F a ilu r e : L o g No r m a l
M u: 8
Sig m a : 3
Q t y: 1
R: 0 . 8 7 1 1 0 1
1 ::2
j
F a ilu r e : No r m a l
M e an : 25 0
St d De v : 5 0
Q t y: 1
R: 0 . 9 9 8 6 5
o
F a ilu r e : W e ib u l
Ch a r . L if e : 1 0 0 0
Sh a p e F a c t . : 4
t 0: 0
Q t y: 1
R: 0 . 9 9 9 9
1 ::1
a
F a ilu r e : W e ib u l
Ch a r . L if e : 1 0 0 0
Sh a p e F a c t . : 1 . 5
t 0: 0
Q t y: 1
R: 0 . 9 6 8 8 7 2
q
F a ilu r e : W e ib u l
Ch a r . L if e : 1 0 0 0
Sh a p e F a c t . : 0 . 4
t 0: 0
Q t y: 1
R: 0 . 6 7 1 5 9
p
F a ilu r e : W e ib u l
Ch a r . L if e : 1 0 0 0
Sh a p e F a c t . : 0 . 5
t 0: 0
Q t y: 1
R: 0 . 7 2 8 8 9 3
k
M T BF : 1 0 0 0 0
Q t y: 1
R: 0 . 9 9 0 0 5
2007 RAMS – Brall, Hagen, Tran
7
Large Model
2007 RAMS – Brall, Hagen, Tran
8
Complex Model
2007 RAMS – Brall, Hagen, Tran
9
Results of Simulations
Model
One Block
One Block
Simple
Simple
Simple
Large
Large
Large
Large
Large
Complex
Complex
Complex
Complex
Model Data
Parameter
Reliability
Availability
Reliability
Availability
System Failures
Reliability
Reliability
Availability
Availability
MTTFF: (Hours)
Reliability
Availability
MTBF (MTBDE)(Hrs)
MTTR (MDT)(Hrs.)
Trials or
Runs
1,000
1,000
1,000
1,000
1,000
10,000
1,000
1,000
10,000
10,000
10,000
10,000
10,000
10,000
Time
(hours)
1,000
1,000
100
100
100
61,362
61,362
61,362
61,362
61,362
100
100
100
100
Software Package
Raptor BlockSim
Relex
0.3797
0.8927
0.983
0.9955
0.017
0.7024
0.718
0.858
0.847
0.3663
0.8894
0.977
0.9892
0.023
0.737
0.729
0.861
0.865
0.365
0.843
0.978
0.978
Not Reported
0.6914
0.707
0.691
0.6866
144,775.99 201,679.13
146,321.53
0.1313
0.1315
0.0988
0.3877
0.3741
0.3333
2007 RAMS – Brall, Hagen, Tran
36.2732
68.3853
39.3565
62.7677
33.92
74.51
10
What Do the Results Tell Us

If precision is required, it isn’t there
One to two significant figure agreement at best between packages
 Confidence limits are necessary for data

Some parameters are either defined differently, or
calculated using such diverse algorithms or methodologies
that they aren’t comparable
 Errors in modeling or application of the software can go
undiscovered when only one software package and one
analyst are used

The complexity of large models and different issues with each
software interface opens up many opportunities for human failure
 Checking a model for errors can be more time intensive than
creating the original model

2007 RAMS – Brall, Hagen, Tran
11
Cautions - 1


Use of a single model, especially a highly complex model, to
demonstrate compliance with a requirement is error prone and risky
Many times the results of these simulations are used to demonstrate
compliance with a specified reliability or availability requirement.
 A result that would show a Reliability of 0.85 when the
requirement was 0.90 might cause redesign, request for waiver,
or other action to address the shortfall.
 The shortfall may be due to the parameters used for the
simulation, the algorithms used by the software, a lack of
understanding of how long to simulate, how many independent
random number streams to use, and/or how many runs to use.
 Analytical solutions for highly complex models are based on
approximations.
2007 RAMS – Brall, Hagen, Tran
12
Cautions - 2

The programs do not necessarily describe variables in the same
manner.
 i.e.When using the Lognormal distribution, there was a
difference in terminology between Raptor and BlockSim.
 Raptor allows the Lognormal to be entered as Mean and
Std Dev. or Mu and Sigma.
 BlockSim only uses Mean and Std. Dev., but this is the
same as Raptor’s Mu and Sigma.
 A novice could waste a great deal of time clarifying
what needs to be entered as data.
2007 RAMS – Brall, Hagen, Tran
13
Cautions - 3
Modeling special cases can be difficult because of the way the
programs handle standby (which was in our models) and
phasing (which was not in our models).
 Output parameters were not consistently labeled. The user
should understand the difference between MTTF, MTTFF,
MTBDE, and MTBF for reliability and MDT and MTTR for
maintainability.

2007 RAMS – Brall, Hagen, Tran
14
Cautions - 4

The products provide reliability and availability results with
various adjectives such as “mean”, “point”, “conditional”, etc.
 A review of the literature provided with the packages is
necessary to understand these terms and relate them to those
found in specifications, handbooks, references, and texts.
 It is a serious issue that there doesn’t appear to be standard
and/or consistent terminology and notation from one
program to another as well as to standard literature in the
field.
2007 RAMS – Brall, Hagen, Tran
15
Cautions - 5

Flexibility
 Each package has tabs, checkboxes, preferences, defaults, multiple
random number streams, selectable seeds for random numbers, etc to
facilitate the modeling, analysis, and simulation process.
 Flexibility can provide huge pitfalls to the analyst.
 Care in modeling, and use of support services provided by the
software supplier is a good practice.
 Numerous runs and reruns may be necessary due to idiosyncrasies of
the software,
 Beware of errors in modeling, confusion of parameter definition, etc.
 Problems compound as a variety of failure distributions are
intermixed with a similar grouping of repair distributions.
 As a model becomes more complex, simulation becomes mandatory
2007 RAMS – Brall, Hagen, Tran
16
Observations - 1


The models can run quickly even on old Pentium II PCs, or they can
take hours to run.
 Length of simulation time, number of runs, and failure rate of the
system can all contribute to lengthening of simulation time.
 One of the models took in excess of 1 hour on a 3 GHz Pentium
IV.
Convergence of the results is heavily dependent on how consistent
the block failure rates are.
 For example, one block with an MTBF of 1000 hours, can double
or triple simulation time.
 The display during simulation on some of the packages shows the
general trend, but there can be a lot of outliers.
 One model failed to converge on one of the packages – again this
may have been due to a subtle preference selection (or nonselection).
2007 RAMS – Brall, Hagen, Tran
17
Observations - 2


The display of Availability and or Reliability during simulation can
be useful for seeing how the simulation is behaving.
 For most models, this rapidly stabilizes to the first decimal place,
then the second decimal place tends to bounce around.
 Usually you get the first 2 significant figures in a hundred runs.
We have the impression that most of the user interfaces were
designed by software designers, working with R&M engineers.
 The problem is that we seem to have gotten what an R&M
engineer would tell someone never having used the product.
 For example, it’s really annoying that you have to double
click and work through tabs to put data into blocks in the
block diagrams; the alternative is to use the Item Properties
Table, which doesn't let you create blocks and in some
cases change probability distributions.
2007 RAMS – Brall, Hagen, Tran
18
Recommendations

When demonstrating compliance to a requirement is required
 Model system using one of the following approaches to reduce
human error
 Have one analyst model in two different software packages
 Software methodologies are sufficiently different to
avoid repeating errors
 Have second analyst perform detailed audit of model and
data entry
 Have two analysts independently model and enter data
 Compare results
 Results should agree within +/- 3 Standard Errors of the
Mean
 Make detailed notes of assumptions, methods, simulation values,
etc. to provide an audit trail
2007 RAMS – Brall, Hagen, Tran
19