Reliability Centered Maintenance

Applied to Achieve Best Life Cycle Costs

Dr Bill Wessels PE CRE

Principal Research Scientist
bill.wessels@uah.edu

Reliability & Failure Analysis Laboratory

http://rfal.uah.edu

Research Institute

Dr Bill Wessels PE CRE

Experience:

1975 1989 Tramp Engineer, mining

mobile machinery & process equipment
life-cycle sustainability
1989 2005 Contractor [Biomedical
Devices, DoD and NASA] design-forreliability, RCM
2005 Research in design-forreliability & life-cycle sustainability for
mechanical and structural components

Credentials

BS USMA, West Point 70

PhD Reliability Systems Engineering, UAHuntsville 96
PE MechE, Pennsylvania, 82
CRE, ASQ, 96
Author Practical Reliability engineering and Analysis for System
Design and Life-Cycle Sustainment, CRC Press, 2010

SRE RAMS 2013

Research Institute

Introduction
Reliability-centered maintenance serves to
preserve system functionality by performing
scheduled maintenance actions that prevent
unscheduled system downing events during
scheduled operations
RCM achieves its objective by developing an
understanding of critical part failure
Understanding part failure is manifested by
the ability to measure the consumed life of a
critical part
Scheduled maintenance is defined by
application of a risk threshold to the
consumed life
SRE RAMS 2013

Research Institute

RCM Roadmap
Critical
Items
List

System
Subsystem

Failure
Mechanisms
Sources

OR

Part

Part

Condition
Indicator

Part

Assembly
Functional
Failure

Part
Failure Mode

Part

Transportation

NO

CBM

YES

Metric: Part Condition Indicator

OR

Storage

Single Fail Mech

Single Fail Mode
f(t)

Function

Y
E
S

Time-to-Failure
Reliability Failure
Math Model

TDM

Metric: Part Time in Service

NO

Operator
Maintainer

Perception of
Part Failure
P-F Interval
Critical
Items
List

RCM
Database

OR

Criticality Analysis

Part Failure
Consequences
Analysis

Operational

Ambient

Multiple Fail Mech

Multiple Fail Mode
f(s)

Y
E
S

Stress-Based
Reliability Failure
Math Model

SDM

Metric: Part Stress Profile

Failure
Mechanisms

Physical
Stress
Failure
Modes

Change in
Shape
(Strain)

OR

RCM
Database

Thermal
Stress

Environmetal
Stress

OR

Change in
Geometry
(Crack)

Part
Failure
Math Model

Part
Repair
Math Model

Logistics
Downtime
Math Model

Preventive
Maintenance
Math Model

Resource
Queuing
Math Model

Spare
Strategy
Math Model

Change in
Material Property
(Corrosion)

Part Cost Objective Function

Labor Costs
Materials Costs
Overhead Costs

Part
Failure
Effects

Part Reliability
Life-Cycle Affordability

Part
Reliability/Cost
Simulation

RCM
Database

Assembly Reliability
Life-Cycle Affordability

Assembly
Reliability/Cost
Simulation

System Reliability
Life-Cycle Affordability

System
Reliability/Cost
Simulation

SRE RAMS 2013

Research Institute

Part ReliabilityMaintainability Cycle

SRE RAMS 2013

Research Institute

System Downtime
SRE RAMS 2013

Research Institute

Source of RCM Cost Benefit

post-Repair Logistics Downtime
Time-to-Repair
pre-Repair Logistics Downtime
W
Re ork
su
me
s

Part fails
System down
Maintenance action initiated
W
St ork
art
s

Time
$Labor
$Overhead

$Labor
$Materials
$Overhead

$Labor
$Overhead

Total Downtime
$Lost Opportunity
post-Repair Logistics Downtime
Time-to-Repair
pre-Repair Logistics Downtime

W
St ork
art
s

W
Re ork
su
me
s

Part fails
System down
Maintenance action initiated

Time
$Labor
$Overhead
$Labor
$Materials
$Overhead

$Labor
$Overhead
$Benefit

$Lost
Opportunity

SRE RAMS 2013

Research Institute

Top Level RCM Path

Develop a rank ordered critical items list CIL

Perform reliability failure analysis to identify:

Sources of failure mechanisms (Conditions of use)

Failure mechanisms action on the critical part (Cause of
part failure)
Failure modes (Damage to part)
Failure effects part (Symptoms of the damage)
Failure effects assembly (Consequences of part
failure)

Select appropriate RCM Path

CBM
TDM
SDM

SRE RAMS 2013

Research Institute

Condition-Based Maintenance
CBM
Consumed life characterized by a condition
that defines the failure mode/damage
Change in shape
Change in geometry
Change in material properties

Method for investigating the damage, NDI

Visual inspection of condition indicator
Manual measurement of condition indicator
Sensor measurement of condition indicator

Metrics that characterize the damage

Risk Threshold
SRE RAMS 2013

Research Institute

Notional Linear CBM Concept

SRE RAMS 2013

10

Research Institute

Notional CBM Concept

SRE RAMS 2013

11

Research Institute

SRE RAMS 2013

12

Research Institute

Decreasing Condition Indicator

SRE RAMS 2013

13

Research Institute

SRE RAMS 2013

14

Research Institute

Diverging Condition Indicator

SRE RAMS 2013

15

Research Institute

SRE RAMS 2013

16

Research Institute

Diverging Condition Indicator

SRE RAMS 2013

17

Research Institute

SRE RAMS 2013

18

Research Institute

Converging Condition Indicator

SRE RAMS 2013

19

Research Institute

SRE RAMS 2013

20

Research Institute

Converging Condition Indicator

SRE RAMS 2013

21

Research Institute

SRE RAMS 2013

22

Research Institute

Fatigue Condition Indicator

SRE RAMS 2013

23

Research Institute

Fatigue Condition Indicator

SRE RAMS 2013

24

Research Institute

SRE RAMS 2013

25

Research Institute

Step Condition Indicator

SRE RAMS 2013

26

Research Institute

SRE RAMS 2013

27

Research Institute

Step Condition Indicator

SRE RAMS 2013

28

Research Institute

SRE RAMS 2013

29

Research Institute

Notional Attribute CBM Concept

SRE RAMS 2013

30

Research Institute

Time Directed Maintenance, TDM

Consumed life characterized by time in
service, but
Single failure mechanism with single failure
mode
Failure mode correlated to operating time

SRE RAMS 2013

31

Research Institute

Time-to-Failure Data

SRE RAMS 2013

32

Research Institute

Time-to-Failure Histogram

SRE RAMS 2013

33

Research Institute

Fit Data to 3-Parameter Weibull

Median Ranks Regression
First: Fit the 2-Parameter Weibull
Then: Select a value for the location
parameter, g
Choose the mid-time between 0 and Tmin
Calculate Xg = Xi g
Calculate r2 for Xg & Y

If: r2 for Xg < r2 for the 2-Parameter Weibull,

use the 2-Parameter Weibull
If: r2 for Xg > r2 for the 2-Parameter Weibull,
fit the 3-Parameter Weibull
SRE RAMS 2013

34

Research Institute

Fit Data to 2-Parameter Weibull Median Ranks Regression

SRE RAMS 2013

35

Research Institute

2-P Weibull Median

Ranks Regression

SRE RAMS 2013

36

Research Institute

2-P Weibull Median Ranks Regression

SRE RAMS 2013

37

Research Institute

TTF Location
Parameter, g

SRE RAMS 2013

38

Research Institute

TTF Location Parameter, g

SRE RAMS 2013

39

Research Institute

3-P Weibull Failure Math Model - TTF

SRE RAMS 2013

40

Research Institute

3-P Weibull Failure Math Model

SRE RAMS 2013

41

Research Institute

Failure Math Model pdf f(t)

SRE RAMS 2013

42

Research Institute

Risk Threshold & Hazard Function

SRE RAMS 2013

43

Research Institute

Risk Threshold Logic

SRE RAMS 2013

44

Research Institute

Part Alternatives Hazard Functions & Risk Threshold

SRE RAMS 2013

45

Research Institute

Parts Alternatives Risk-based

Scheduled Maintenance Interval, SMI

SRE RAMS 2013

46

Research Institute

Exponential v Weibull Failure Math Models

SRE RAMS 2013

47

Research Institute

Exponential v Weibull Hazard

Function w/Risk Threshold

SRE RAMS 2013

48

Research Institute

Time-to-Repair Data

SRE RAMS 2013

49

Research Institute

Time-to-Repair Histogram

SRE RAMS 2013

50

Research Institute

3-P Weibull Failure Math Model

SRE RAMS 2013

51

Research Institute

Repair Math Model

SRE RAMS 2013

52

Research Institute

LDT Math Model - Baseline

SRE RAMS 2013

53

Research Institute

pre-Repair LDT Math Model

Baseline
SRE RAMS 2013

54

Research Institute

post-Repair LDT Math Model

Baseline
SRE RAMS 2013

55

Research Institute

pre-Repair LDT Math Model

SRE RAMS 2013

56

Research Institute

post-Repair LDT Math Model

SRE RAMS 2013

57

Research Institute

Deterministic Reliability,
Maintainability & Availability

SRE RAMS 2013

58

Research Institute

Reliability Simulation Inputs

SRE RAMS 2013

59

Research Institute

Baseline Maintenance Labor Inputs

SRE RAMS 2013

60

Research Institute

PM Maintenance Labor Inputs

SRE RAMS 2013

61

Research Institute

Maintenance Costs Inputs

SRE RAMS 2013

62

Research Institute

Part Life-Cycle Simulation Factors

SRE RAMS 2013

63

Research Institute

Baseline Reliability Simulation

Report
Simulation Report for Part 1 Baseline
Results from 259 runs of sim time 40000-hrs
Part 1
Availability
MTBF

Minimum

Mean

Maximum St.Dev

0.98984 0.99051

95% CL

0.99137 0.00030 0.99002

SEM

95% CL

0.00002 0.99048

628.51

683.68

734.35

19.74

651.09

1.23

681.65

MDT

6.32

6.54

6.80

0.08

6.68

0.005

6.55

MMT

1.67

1.76

1.88

0.03

1.81

0.002

1.76

LDT
System Failures

4.79
54

58.00

4.86
63

SRE RAMS 2013

1.66

4.79
0.10

58.17

64

Research Institute

PM Reliability Simulation Report

Simulation Report for Part 1 - PM

Results from 259 runs of sim time 40000-hrs
Part 1
Availability
MTBF

Minimum

Mean

Maximum St.Dev

95% CL

SEM

95% CL

0.99305 0.99328 0.99342 0.00006 0.99318 0.000004 0.99328

467.32

467.43

467.49

0.03

467.38

0.00

467.42

MDT

3.10

3.16

3.27

0.03

3.73

0.00

3.16

MMT

0.75

1.71

2.38

0.04

1.78

0.00

1.71

LDT

0.83

1.40

2.19

85

85

85

System Failures

SRE RAMS 2013

1.95

1.45

65

Research Institute

Reliability Simulation Summary

SRE RAMS 2013

66

Research Institute

Labor, Materials, Overhead & Lost

Opportunity Costs Table

SRE RAMS 2013

67

Research Institute

Engineering Economics Basics

n

NPV PVi ,t
t 0

1 i 1
PVi ,t Ai ,t
t
i 1 i
t

i 1 i
NPV

n
1 i 1
n

Aeqv
SRE RAMS 2013

68

Research Institute

Economic Analysis
Expensed Costs &
Lost Opportunity Costs

SRE RAMS 2013

69

Research Institute

Part Uniform Recurring Equivalent Cost

DetMean Det95%CL BSimMean BSim95%CL PMSimMean PMSim95%CL
A(CICC+Event)

-$335

-$490

-$332

-$348

-$425

-$432

A(CExp+LostOp) -$2,118 -$3,816 -$2,090

-$2,239

-$1,452

-$1,747

SRE RAMS 2013

70

Research Institute

Part Operational Availability

Baseline

RCM

Part AO 0.99051 0.99328

SRE RAMS 2013

71

Research Institute

Stress-Directed Maintenance,
SDM
Consumed life characterized by applied stress
profile
Measures failure mechanisms acting on part
Not well developed for use by practioner

SRE RAMS 2013

72

Research Institute

SDM - The Present Static Loading

h(x)

x g
h( x)

r
xr

f s

f (s)

1
s e

s
sop sdesign

g
95%
LCLStrength

SRE RAMS 2013

73

Research Institute

SDM - The Future - Wearout

SRE RAMS 2013

74

Research Institute

Summary
RCM yields definition of scheduled
maintenance interval for critical parts
RCM measures part consumed life constrined
by a risk threshold
RCM metrics:
CBM Condition indicator
TDM Time in service
SDM Stress load profile

SRE RAMS 2013

75

Research Institute