Failure Reporting, Analysis,

Corrective Action System

Presented by: Ricky Smith, CMRP

Copyright 2009 GPAllied©

What is FRACAS?
• A Failure Reporting, Analysis, and Corrective Action
System (FRACAS)
– Process by which failures can be reported in a timely manner
– Analyzed so that a Corrective Action System can be put in place
and eliminate or mitigate the recurrence of a failure.
• The goal of FRACAS
– Help an organization better understand failures
– The reporting required to identify failures
– Actions required to eliminate or mitigate failures

Copyright 2009 GPAllied©

What is FRACAS?

Copyright 2009 GPAllied©

Maintenance’s Objective
“Mitigate Failures with an effective Maintenance Strategy”

- Preventive Maintenance (quantitative inspection, restoration,

replace, replenish, etc.)
- Condition Monitoring (ID of a Failure early enough a failure
can be mitigated) also known as PdM
- Run-to-Failure

Copyright 2009 GPAllied©

PF Curve
Priority 4
Priority 5 P-F Curve
Failure Ultrasonic Energy
Vibration Analysis
Initiated Detected Fault Detection Oil Analysis
Detected

Audible Noise

Priority 2
Equipment Condition

Hot to Touch

Mechanically
Condition

Loose

Ancillary

Priority 1
Damage

Catastrophic
Failure
PRECISION PREDICTIVE PREVENTIVE RUN TO FAILURE

Time

Copyright 2009 GPAllied©

Proactive Maintenance

Copyright 2009 GPAllied©

FRACAS? Issues?

“A Proactive Reliability Process is a supply chain. If a step

in the process is skipped, or performed at a substandard
level, the process creates defects known as Failures.

The output of a healthy reliability process is optimal asset

reliability at optimal cost.”

― Ron Thomas, former Reliability Director at

Dofasco Steel, Hamilton, ON

Copyright 2009 GPAllied©

Causes of Failures
• No repeatable procedures
• Not executing PM to specification
• Not performing Corrective Maintenance to specification
• Personnel not trained (skills, maintenance, reliability)
• Parts not stored to specifications
• Not using the right tools
• Best Lubrication Practices unknown
• RCA is not applied or not known
• No auditing of Maintenance Activity or Work

Copyright 2009 GPAllied©

Known Best Practice Data
• 15% of Maintenance Work is Execution of PM
• 15% is the results of the PM Execution

PM / PdM is a controlled Experiment

• 15% of Maintenance is Execution of PdM

• 35% is the results of PdM
• Emergency Work – Less than 2%
• Planned Work – 90%
• Scheduled Compliance – 80% (by day by week)

Copyright 2009 GPAllied©

Failure Patterns
Initial Break-in period
Bathtub Pattern D = 7%
Pattern A = 4%

Random
Pattern E = 14%
Wear Out
Pattern B = 2%
“Why is infant mortality so high?”
Maybe because we do not applyTime
Time the PF Curve Philosophy?

Fatigue
Pattern C = 5% Infant Mortality
Pattern F = 68%

Age Related = 11% Random = 89%

Source: John Moubray, Nolan & Heap

Copyright 2009 GPAllied©
PF Curve with Priorities
Identify defect, Plan and Schedule Work Zone

100% Reactive Work

Copyright 2009 GPAllied©
Work Priority Distribution

PM PdM CPM CPdM REQ

Priority 1 and 2 Work is Reactive

Priority 3, 4, and 5 Work is Proactive

Copyright 2009 GPAllied©

PF Curve and Managing Proactive Maintenance

Point P

Point F
Copyright 2009 GPAllied©
Work Priority Distribution – a Few Simple Rules

EMERGENT work is anything

that is done as a Priority 1 or 2

CORRECTIVE work is
anything done as a result of
an inspection

CORRECTIVE work should

never be done as a P1 or P2
ROUTINE work is anything
done as a PRIORITY 3

CORRECTIVE work can be

done as Priority 3

CORRECTIVE work should be

done as a Priority 4 and 5

Most CORRECTIVE work

should be done as P4 or P5 if
you embrace P-F mentality

Copyright 2009 GPAllied©

FRACAS Enables Success of an Organization

Copyright 2009 GPAllied©

“It isn’t what you know that will kill you, it is
what you don’t know that will”

Copyright 2009 GPAllied©

It is all about the Data

Copyright 2009 GPAllied©

How do you know where are?
How do you know the Direction to reach you objective.?

Copyright 2009 GPAllied©

Without good data we are lost!

Copyright 2009 GPAllied©

“Bad Data” resulting in High Variation
- Variation in your Maintenance Process

- PM Program not effective

- Variation in PM Compliance
- PM program not focused on specific Failure Modes
- Repeatable Corrective Maintenance Procedures not
available or if available not used
- Storeroom is a deathtrap for parts, little or no PM
program on critical spares
Copyright 2009 GPAllied©
Definition of Data Quality – ISO 14224
Must have Confidence in the collected Reliability and
Maintenance data, and hence any analysis, is strongly
dependent on the quality of the data collected. High-quality
data is characterized by:
a) completeness of data in relation to specification;
b) compliance with definitions of reliability parameters,
data types and formats;
c) accurate input, transfer, handling and storage of
data (manually or electronic);
d) sufficient population and adequate surveillance
period to give statistical confidence;
e) relevance to the data users need.

Copyright 2009 GPAllied©

Copyright 2009 GPAllied©
The International Standard for Failure Data

Copyright 2009 GPAllied©

Failure Data – ISO 14224

Copyright 2009 GPAllied©

ISO 14224 Maintenance Taxonomy Standard

Copyright 2009 GPAllied©

ISO 14224 Maintenance Taxonomy Applied

Copyright 2009 GPAllied©

Effects of Good Data Quality

Copyright 2009 GPAllied©

KPI #1 – Asset Health Report

Asset Health Metric - The percent of assets with no identifiable “Defect”

Copyright 2009 GPAllied©

KPI #2 – Mean Time Between Failure
• Mean Time Between Failures (MTBF) is the average length of
operating time between failures for an asset or component.
(Definition extracted from Published SMRP Metrics)

• MTBF is used primarily for repairable assets and components of

similar type. (Definition extracted from Published SMRP Metrics)

Copyright 2009 GPAllied©

KPI #2 - MTBF Example / Reporting by Taxonomy

Area Level MTBF Component Level MTBF

Equipment Taxonomy (ISO 14224)

Systematic classification of equipment into generic groups based on factors
possibly common to several of the items

Copyright 2009 GPAllied©

KPI #3 – MTBF / “Equipment Condition Report”

Asset Health Metric - The percent of assets with no identifiable “Defect”

Copyright 2009 GPAllied©

KPI #4 – “Equipment Condition Report / Equipment Detail”

Asset Health Metric - The percent of assets with no identifiable “Defect”

Copyright 2009 GPAllied©

KPI #5 – Route Compliance

Copyright 2009 GPAllied©

Route Compliance Impact on Asset Health Report

Asset Health Metric - The percent of assets with no identifiable “Defect”

Copyright 2009 GPAllied©

KPI #6 – Monthly Maintenance Cost % of RAV
• Maintenance Cost
- Labor Cost
- Material Cost
- Contract Maintenance Cost
- Overtime Cost
• Correlate Cost to Other KPIs

Copyright 2009 GPAllied©

Data Accuracy Requirements
• Process Map
• Roles and Responsibilities Identified

Copyright 2009 GPAllied©

The 7 Steps to a Successful FRACAS

Copyright 2009 GPAllied©

Step 1 – Determine your end goal.
• The beginning of every journey starts with a destination
• The beginning of the journey to an effective FRACAS is the same as
any other
• You must have an end goal in mind
• The goal of a FRACAS is not to gather data, but to eliminate failures
from the organization
• Knowing this helps ensure that every policy, procedure, and activity in
the system is goal oriented
• The roles, goals, and responsibilities of everyone involved with the
system can be focused toward that goal
• Having the goal in mind allows you to build the shared vision and
values that will make the system work successfully for you

Copyright 2009 GPAllied©

Step 2 – Create the Data Collection Plan
1. Determine the measures you will use
– MTBF
Mean Time Between Failures (MTBF) is the average length of operating time
between failures for an asset or component. MTBF is used primarily for
repairable assets and components of similar type.

– MTTF
Mean Time to Failure (MTTF) is the average length of operating time to failure
of a non-repairable asset or component, i.e., light bulbs, rocket engines. A
related term, Mean Time Between Failures (MTBF), is the average length of
operating time between failures for an asset or component. Both terms are a
measure of asset reliability and are also known as Mean Life.

Copyright 2009 GPAllied©

Continued
2. Determine what data needs to be collected to create the desired
measures

Copyright 2009 GPAllied©

Continued
3. Determine how the data will be collected
- There are different types of data that need to be collected
- Determine how the data will be collected
- Failure data can be collected through the EAM/CMMS, automated process
data systems, or by using checklists

4. Determine how data will be analyzed

– The best bet is probably to start with Pareto
– Just make sure to remember that data analysis - where to apply methods
JDI, RCA and RCM
– Do not fall victim to analysis paralysis. Analytical reports are nice, but no
statistic ever solved a problem

Copyright 2009 GPAllied©

Keep it simple in the beginning

# of Failure by
1st area

# of Failures by
2nd
Equipment

3rd # of Failure by
Component or Part

Copyright 2009 GPAllied©

Examples – Dominant Failure Pattern

Cause Code
=
Failure
MTBF –
Object Type # LOL = 5%
(Electric Motor) # MAERR = 10%
# LOP = 20%
Pattern
# OPER = 65%

Cause Code
=
# LOL
# MAERR
#LOP
# OPER

New Failure Pattern

Copyright 2009 GPAllied©
Example Continues
• % of Assets with No Identifiable Defect

• Failure Rate for Specific Components

Drive Belt-Broken-Ageing Failure Rate

0.0010696

2-parameter
Weibull
0.00096266 Median rank

η: 5212
0.0008557
β: 3.745
γ: 0
0.00074874
ρ: 0.9781
ε: 0.05664
0.00064178
Failure Rate

B10: 2858
0.00053481 B15: 3208
B20: 3492

0.00042785
P0: 0%

0.00032089
Distribution rate
Reg ionalised rate

0.00021393

0.00010696

0
0 602.49 1205 1807.5 2410 3012.5 3614.9 4217.4 4819.9 5422.4 6024.9

Time

Copyright 2009 GPAllied©

Step 3 – Determine Organizational Roles, Goals,
and Responsibilities (RACI)
• Who collects the data?
• Who analyzes the data?
• Who takes what action based on analysis results?

Copyright 2009 GPAllied©

Develop Process Map / Maps

Copyright 2009 GPAllied©

Develop a RACI Chart for each Process Map
Decisions/ Maint. Maint. Reliability Maint. Maint.
Functions Manager Supervisor Engineer Planner Tech
Failure Data Entry I A C I R
Data Accuracy A R C C C
RCA – Invalid Data I A R C R
Analysis of Data A I R C C
Actions Identified A C R I C
Actions Taken A R C C I
FRACAS Activated A R C C I

Copyright 2009 GPAllied©

Step 4 – Create the FRACAS Policies and
Procedures Manual
• Policies and Procedures Manual that will serve as the basis for
managing and administering the FRACAS system
• This is a tedious step, but should not be skipped
• This Manual will serve as the basis for developing:
– Initial FRACAS Training for all key personnel
– Ensuring that new employees understand and participate in
FRACAS effectively
– Best to develop a FRACAS Management Manual and a pocket
size book that is easy for people to carry around and refer to as
required

Copyright 2009 GPAllied©

What should be in the Manual?
• Let’s develop this together to insure we both agree on the content
• The manual should have at the minimum
– Definition of FRACAS
– Benefits of FRACAS
– Managements’ Guidance
– Managements’ Expectations
– Roles and Responsibilities by position
– FRACAS reports
– Key data which must be input to generate these reports
– What these reports will do for the mine
– Training Outline for each person based on RACI Charts

Copyright 2009 GPAllied©

Step 5 – Develop FRACAS Training Plan

• Each person in the organization will need to be trained according to

their level of participation in the FRACAS
• Use the Tasks Listed on all RACI Charts

Decision Maint. Maint. Reliability Maint. Maint.

Function Manager Supervisor Specialist Planner Tech
Failure Data Entry
Data Accuracy
RCA – Invalid Data
Analysis of Data
Actions Taken
FRACAS Activated

Copyright 2009 GPAllied©

Training Plan Requirements
- Training Plan should consist of the following
 Position : Maintenance Technician
 Task: Close out a Work Order
 Condition: Given 5 Completed Emergency Work Orders
 Standard: Close out Work Orders to 100% Compliance
 Method of Training: Lecture, Web, Reading, etc.
 Method of Validation: Written Test, Web Test, Verbal Recall, etc.

Copyright 2009 GPAllied©

Step 6 – Implement FRACAS
• Let’s turn it on
– Publish FRACAS Policies and Procedures Manual
– Train required personnel
– Hold required informational meetings
– Begin data collection on highest priority area
– Analyze data and report results on Public FRACAS Information
Board
– Create corrective actions based on results.
• Mitigation of Human Error
• Changes to the current maintenance strategy
• Changes to how production operates equipment
• Resign Equipment

Copyright 2009 GPAllied©

Step 7 – Monitor, Show Success, and Adjust
• Monitor data quality and results

Decision Maint. Maint. Reliability Maint. Maint.

Function Manager Supervisor Specialist Planner Tech

Failure Data Entry I A C C R

Data Accuracy A R C C I

RCA – Invalid Data A C R I C

Analysis of Data A I R C I

Actions Taken A R C R I

FRACAS Activated A R R C I

Copyright 2009 GPAllied©

Monitor and Adjust
• Good data is the backbone of good decision making
• It is important to monitor data quality
• Make adjustments to either the data collection plan or the training
program
• Insure data is consistent and informative
• Many organizations believe they have good data only to find out
their data collection is inconsistent

Copyright 2009 GPAllied©

Questions?

Copyright 2009 GPAllied©