Fmea

Spring, 2005

FAILURE MODES AND EFFECTS ANALYSIS

(FMEA)

Julie VanLaanen

Failure Modes and Effects Analysis (FMEA)

Failure Mode and Effect Analysis is an analytical technique used by a product design team as a means to identify, define, and eliminate, to the extent possible, known or potential failure modes of a product or system.

How The FMEA Fits Into Your Final ProposalProposal Sections:

Letter of Intent ----------------->> TransmittalTitle Page

Executive Summary Table of Contents

List of FiguresLetter of Intent ------------------>> Introduction

QFD/House of Quality ------>> Design ObjectivesFMEA Analysis ----------------->> Product, System, or Process

Division of ResponsibilityProject Milestones------------- >> Project Schedule

Budget Resume -------------------------->> Qualifications BibliographyQFD/House of Quality ------>> Appendices

Detailed Project Schedule- >> Appendix Glossary

Was Failure a design requirement in your QFD?

Should it have been?

Can you design for failure?

Failure is ALWAYS a Design Constraint

All Products fail!

Determining how they fail, when they will fail, and why they are failing will allow a designer to incorporate failure as an acceptable design constraint.

Failure as an acceptable design constraint = Customer Satisfaction = Design Quality

From: “Inviting Disaster – Lessons From the Edge of Technology”

By James R. Chiles

Regarding the continuing failure of rear cargo door on a DC-10:

“ The design was originally going to use hydraulics, but under pressure from its client, American Airlines, to simplify and lighten the DC-10 equipment; McDonnell Douglas shifted to an electric door closer instead. This worried engineers working for the builder of the door assembly, Convair Division of General Dynamics. Convair engineers even sent McDonnell Douglas a formal document, called a “failure modes effects analysis,” describing the problem and the disastrous consequences.

How could your design fail?

Identify potential failure areas in your designDetermined analytically through models and

analysisDetermined empirically through stressed life

testingDetermined after product release through

customer warranties and failure reporting

When will my design fail?

Determining how my design might fail requires a through understanding of my design.

Part of answering this question is determining when my customer believes failure is acceptable.

When will my design fail?

Design failure is acceptable if customers perceive it as acceptable. Design Failure has a direct effect on Customer perceptions of Product Quality.

Periodic Maintenance (oil changes)Periodic Replacements (change tires)Acceptable End of Life (customer expectations –

car replacement every 8-10 years)

Reliability Curve

Time

FailureFrequency

ChanceInfantMortality

Wearout

Useful Life

Assuming a part/product is designed correctly, statistically it should follow a standard bath-tub curve

Infant Mortality – failures don’t usually affect the customer, except DOAUseful life – failures here are problemsWear Out – failures here are acceptable

Causes of Reliability Curve Failure

Early Failure(Infant Mortality)

Chance Failure Wear Out Failure

- Poor Quality Control

- Poor workmanship

- Insufficient Burn In

- Substandard Materials

- Contamination

- Human Error

- Incorrect Packaging

- Design Errors

- Insufficient SafetyMargins

- Undetected Defects

- Misapplication

- Abuse of Equipment

- Lack of PreventiveMaintenance

- Acts of God

- Stress Concentration

- Corrosion

- Fatigue

- Creep

- Abrasive Wear

- Wrong Environment

- Poor Maintenance

Tools used to Design for acceptable Product Failure

Failure Modes and Effects Analysis (FMEA): Method to identify and understand potential product failures and their severity

Mean Time Between Failures (MTBF) of individual parts. Determined by:

MIL Standards Testing Mathematical Predictions

Design of Experiments (DOE): Experimentally optimizing our design using physical models and optimized testing (Taguchi)

Tools to used to Design for acceptable Product Failure (cont.)

Statistical Process Control (SPC): Monitoring process failures through tight statistical control of process variables

Six-Sigma (6): Statistical repeatability to six standard deviations (in other words, 99.9999998% within design specifications!)

QUALITY MAPProcess Control to

Identify Possible Causes

SPC

6

Quantify Processesand Rank Suppliers

DOE

Model Variablesand InteractionsReduce Variance

Reliability Engr.

Evaluate ProductPerformance Predict

Product Life

FMEA

Prevent Failures fromReaching the Customer

Failure Modes and Effects Analysis (FMEA)

Begins at the initiation of a product design cycle

Seeks to identify potential failure modes before a actual failure occurs

Continues throughout product design cycle

Three Key Questions Posed by FMEA

What could fail in each component of my product?

To what extent might it fail and what are the potential hazards produced by the failure?

What steps should be implemented to prevent failures?

Design FMEA AnalysisItem andFunction

PotentialFailureMode

PotentialEffects of

Failure

SEV

PotentialCause(s)of Failure

OCC

DetectionMethod &

QualityControls

DET

RPN

RecommendedActions

List PartName,NumberandFunction

List thepossiblemodes offailure

List theconsequencesof failure onpart functionand on thenext higherassembly

List thosesuch as:inadequatedesign,impropermaterials,etc.

List thesemeasuresavailable todetectfailuresbefore theyreach thecustomer

List them foreach of thefailure modesidentified asbeingsignificant bythe RPN

= Critical characteristic which may effect safety, compliance with Gov. regulations, or require special controls.

SEV = Severity rating (1 to 10)

OCC = Occurrence frequency (1 to 10)

DET = Detection Rating (1 to 10)

RPN = Risk Priority Number (1 to 1000)

Rifle Bolt ExamplePart and Function

Potential Failure Mode

Potential Effects of Failure

S

E

V

Potential Cause(s) of Failure

O

C

C

Detection Method & Quality Controls

D

E

T

R

P

N

Recommended Actions

Item and FunctionList each subassembly and component, along with it’s basic

function. Example: Rifle Bolt

Part and Function



S

E

V


O

C

C


D

E

T

R

P

N

Recommended Actions

Rifle BoltChambers bulletLocks into receiverFires RoundSustains firing pressure on lugsProvides extraction of spent case

Identify and List Potential Failure Modes

For any failure mode, the idea is to identify what could occur, not that it will necessarily occur. Identify through:

BrainstormingAnalysisTestExperience

Example Failure Modes

Identify and List Potential Failure Modes

Part and Function



S

E

V


O

C

C


R

P

N

Recommended Actions


Fracture

Jamming

List Potential Effects from Failure

Impact on EnvironmentHazard to humansExcessive noisePoor AppearanceUnpleasant OdorErratic Operation

Potential Effects from Failure

Part and Function



S

E

V


O

C

C


R

P

N

Recommended Actions


Fracture Catastrophic failure with destruction of weapon and injury to personnel

Jamming Failure of weapon to function

Critical Potential Effects from Failure

Are the potential effects critical? Will they effect: Personal Safety Compliance with a Government Standard Compliance with a Government Regulation Require special controls

Identifying Critical Failures

Part and Function



S

E

V


O

C

C


R

P

N

Recommended Actions




Specifying Severity of Failure (SEV)Estimate potential severity (SEV) of failure and its effect

Ranking Effect Criteria: Severity of Effect1 None No effect2 Very Minor Very minor effect on product or system

performance.3 Minor Minor effect on product or system performance.4 Low Small effect on product performance.

The product does not require repair.5 Moderate Moderate effect on product performance.

The product requires repair.6 Significant Product performance is degraded. Comfort or

convenience functions may not operate.7 Major Product performance is severely affected but

functions. The system may not be operable.8 Extreme Product is inoperable with loss of primary function.

The system is inoperable.9 Serious Failure involves hazardous outcomes and / or

noncompliance with govt. regulations or standards.10 Hazardous Failure is hazardous, and occurs without warning. It

suspends operation of the system an/or involvesnoncompliance with govt. regulations

Specifying Severity of Failure (SEV)

Part and Function



S

E

V


O

C

C


R

P

N

Recommended Actions



10


8

Potential Cause of Failure

List possible potential causes of failure Tolerance stack-up Assembly errors Poor maintenance Impact loading Overstressing Poor Design

These causes should indicate appropriate preventive measures that might be adopted

Potential Cause of Failure

Part and Function



S

E

V


O

C

C


R

P

N

Recommended Actions



10

ShrinkagePorosity caused by improper feed


8 Out of spec. DimensionChange in shell refractory

Likelihood of Occurrence (OCC)Estimate the potential occurrence of failure

Ranking PossibleFailure Rates

Probability of Failure

1 1 x 10-6 Nearly Impossible2 1 x 10-5 Remote3 1 x 10-4 Low4 4 x 10-4 Relatively Low5 2 x 10-3 Moderate6 1 x 10-2 Moderately High7 4 x 10-2 High8 0.2 Repeated Failures9 0.33 Very High10 0.55 Extremely High: Failure Almost

Inevitable

Likelihood of Occurrence (OCC)

Part and Function



S

E

V


O

C

C


R

P

N

Recommended Actions



10


6



5

Detection Methods & Quality Controls

Develop Methods to detect and control the potential causes of failure Manufacturing tests Incoming Inspection SPC Methods Vendor/Supply Controls Dimension Inspection

Detection Methods & Quality Controls

Part and Function



S

E

V


O

C

C


R

P

N

Recommended Actions



10


6 Incoming Part InspectionDye penetrant testing



5 Measure patternsConfirm finished casting dimensions

Likelihood of Detection (DET)Estimate the likelihood that the detection method

in place will actually detect the fault

Ranking Detection Probability1 Almost Certain Detection2 Very High Chance of Detection3 High Probability of Detection4 Moderately High Chance of Detection5 Moderate Chance of Detection6 Low Probability of Detection7 Very Low Probability of Detection8 Remote Chance of Detection9 Very Remote Chance of Detection

10 Absolute Uncertainty – No Control

Likelihood of Detection (DET)

Part and Function



S

E

V


O

C

C


D

E

T

R

P

N

Recommended Actions



10 ShrinkagePorosity caused by improper feed


5




3

Calculate a Risk Priority Number (RPN)

The RPN prioritizes the relative importance of each failure mode and effect on a scale of 1 – 1000. It is calculated as follows:

RPN = (SEV) x (OCC) x (DET)A 1000 rating implies a certain failure that is

hazardous and harmfulA 1 rating is a failure that is highly unlikely and

unimportantRatings above 100 will occurRating below 30 are reasonable for typical applications

Calculate a Risk Priority Number (RPN)

Part and Function



S

E

V


O

C

C


D

E

T

R

P

N

Recommended Actions





5 300




3 1

2

0

Develop Recommended Actions

Based on your RPN number, develop recommended actions to solve failure modes Assign responsibilities Outline corrective actions Revise test plans, material specifications

These actions should be specific, not general action items

Develop Recommended ActionsPart and Function



S

E

V


O

C

C


D

E

T

R

P

N

Recommended Actions





5 300

Initiate radiographic testing of all rifle bolts




3 1

2

0

Initiate SPC program to check and maintain bolt dimensions

Spring, 2005

Acknowledgement

Lecture taken from:Dr. Bob Frost of the Metallurgical Engineering Dept. from his course on Statistical Process Control (SPC)Dr. Munoz from the Mechanical Engineering Department.Product Design: Techniques in Reverse Engineering and New Product Development by Kevin Otto & Kristin Wood, pp 564-571.

Career

Fmea