Transcript For the top 5-10% quantities risk and performance analysis may be

ANALYTICAL METHODS IN
MAINTENANCE
• What Maintenance Analysis method
• Examples of Using Maintenance Data
Maintenance Analysis
 Dependant upon criticality, differing levels of analysis effort
should be applied.
 For the top 5-10% quantities risk and performance
analysis may be warranted
 For the next 40-50% use template a rule-based
methods (such as RCM).
 At the lower levels of criticality, not even simple
analysis is worthwhile, and instead may rely on
manufacturer’s recommendations or retain current
practice.
 For the top 5-10% quantities risk and performance analysis may be warranted:
ANALYSIS/SIMULATION SOFTWARE
** SIMLOX
Powerful and versatile simulation tool for event based
simulation and analysis of complex operational and logistic
support operations. Provides graphs on system availability,
resources utilisation, actual versus requested mission time,
etc.
www.systecon.se
** LOGAN Monte Carlo simulation tool
Evaluates availability of complex systems. Can include
effect of manning levels, spares holdings, redundant and
standby systems, etc.
www.reliability-safety-software.com
 For the top 5-10% quantities risk and performance analysis may be warranted:
AvSim+
•
AvSim+ is an availability simulation and system optimisation tool
developed specifically to handle the complexity of large plants
and systems.
•
It uses Monte-Carlo simulation to calculate the expected system
reliability, availability and capacity performance over a chosen
life cycle.
www.reliability.com.au
Maintenance PLANS:
Horses for Courses
 Dependant upon criticality, differing levels of analysis effort
should be applied.
 For the top 5-10% quantities risk and performance
analysis may be warranted
 For the next 40-50% use template a rule-based
methods (such as RCM).
RCM SOFTWARE
** RCM Turbo
Australian owned and developed expert methodology for the
implementation of reliability centred maintenance strategies.
A unique decision support platform leading directly to new
schedules. www.strategicorp.com
DISPLAYING MAINTENANCE
AND
RELIABILITY DATA
Log Scatterplot of mean repair times versus number of failures
with limit values.
100.0
12
Mean Time To Repair
ACUTE
28.9
8 2
7
15
17
16
6
14
13
ACUTE &CHRONIC
10
3
5
1
11
4
10.0
9
CHRONIC
1.0
1
10
15.9
100
Num ber of Failures
“Downtime Priorities, Jack-Knife Diagrams and the Business Cycle”
Peter Knights, May 2004 issue The Maintenance Journal
100.0
12
Mean Time To Repair
ACUTE
28.9
16
17
8
15
6
14
13
10.0
72
4
ACUTE & CHRONIC
10 1
3
11
5
9
RELIABILITY
PROBLEMS
CHRONIC
1.0
1
10
15.9
Num ber of Failures
100
100.0
12
Mean Time To Repair
ACUTE
28.9
16
17
8
15
6
14
13
10.0
72
4
AVAILABILITY
PROBLEMS
10 1
3
11
5
9
CHRONIC
1.0
1
10
15.9
Num ber of Failures
100
100.0
Mean Time To Repair
MAINTAINABILITY
PROBLEMS
28.9
16
12
17
8
15
72
6
14
13
4
10.0
ACUTE &CHRONIC
10
3
5
1
11
9
CHRONIC
1.0
1
10
15.9
Num ber of Failures
100
CASE STUDY
USING
MAINTENANCE DATA
Casing seal
Bearing 1
Bearing 2
coupling
Shaft
seal
ELECTRIC MOTOR
GEAR BOX
Casing seal
Bearing 1
Bearing 2
coupling
Shaft
seal
ELECTRIC MOTOR
GEAR BOX
Failures over a 48 month period :MTBF
900 hrs
DOWNTIME
450 hrs
No
FAILURES
36
FAILURES
>3hrs DT
14
Casing seal
Bearing 1
Bearing 2
coupling
Shaft
seal
ELECTRIC MOTOR
GEAR BOX
MTBF
DOWNTIME
No
FAILURES
FAILURES
>3hrs DT
Bearing 1
4000hrs
9hrs
8
2
Bearing 2
2000hrs
427hrs
16
12
Coupling
8000hrs
2hrs
4
0
Shaft seal
6000hrs
5hrs
6
0
Casing seal
16000hrs
2hrs
2
0
HISTORY OF FAILURES BEARING 2
Clock hrs
Hrs to fail
History of failure
1.
190
190
Bearing incorrectly………
2.
205
15
Lack of………
3.
3705
3500
4.
3785
80
5.
4105
320
6.
8305
4200
7.
8330
25
8.
8570
240
9.
13570
5000
10.
17570
4000
11.
17630
60
12.
22130
4500
13.
22570
440
14.
26670
4100
15.
26710
40
16.
29510
2800
1st failure at 190 hrs
2nd failure at 205 hrs

0hrs
3rd failure at 3705 hrs

4000hrs
1st failure at 190hrs
2nd failure at 205hrs
0hrs
3rd failure at 3705hrs

4000hrs

8000hrs

12000hrs
 
16000hrs

20000hrs


24000hrs



28000hrs
32000hrs
BEARING 2
WORST
1
15.00
2
25.00
3
40.00
4
60.00
5
80.00
6
190.00
7
240.00
8
320.00
9
440.00
10
2800.00
11
3500.00
12
4000.00
13
4100.00
14
4200.00
15
4500.00
16
5000.00
BEST
*RANKED ORDER*
HISTOGRAM OF BEARING No.2 FAILURES
10
9
8
NUMBER
OF
7
FAILURES
6
5
4
3
2
1
0
500 1000 1500 2000 2500 3000 3500 4000 4500
WEIBULL ANALYSIS
• It
allows
analysis
COMPONENTS FAIL.
of
HOW
• If it is found that failure is AGE
RELATED then you can also identify
the
optimum
REPLACEMENT
INTERVAL
BEARING 2
1
15.00
2
25.00
3
40.00
4
60.00
5
80.00
6
190.00
7
240.00
8
320.00
9
440.00
10
2800.00
11
3500.00
12
4000.00
13
4100.00
14
4200.00
15
4500.00
16
5000.00
*RANKED ORDER*
WEIBULL ANALYSIS
RELCODE Data Analysis Software
RELCODE is a software package for Weibull
analysis of Failure Data.
www.albanyint.com.au
MONTE CARLO SIMULATION
6
NUMBER
OF
FAILURES
5
1
2
3
4
7
0 2500 3000 3500 4000 4500
• Use random numbers to recreate the failures
• Number range 1 to 7
• If No. 1 then this is a failure at 2500 hrs
1st failure at
2500 hrs

0hrs
$1000
6
20 hrs
1
NUMBER
OF
FAILURES
5
2
3
4
7
0 2500 3000 3500 4000 4500
Add repair hrs and cost data
1st failure at 2500 hrs
Repaired at 2520 hrs
cost $1000
 
0hrs
• Using more random numbers you can continue
generating failures and repair/cost data.
• You may test a number of fixed time replacement
events, test which failures they will prevent
and see which replacement time is cost effective.
• You can run the simulation for 100yrs for 1000
similar machines operating in parallel and
establish number of man hrs per week
required to maintain the machines.