1

Failure Mode and Effect Analysis

(FMEA)

2

What is Failure Mode and Effect Analysis (FMEA)?

FMEA is a quality audit procedure which aims to anticipate failures in a product’s functional design.

– “Failure” may be the result of a design, manufacturing process, or use or the malfunctions of a product or service.

3

FMEA

The aim of FMEA is to anticipate in advance:

– What might fail

– What effect this failure would have on the overall function, &

– What might cause the failure

4

FMEA

The significance of the failure is assessed against:

– The probability of failure occurrence

– An assessment of the severity and the effect of that failure

– The probability of existing quality systems spotting the failure before it occurs

5

Where Does FMEA Occur in the product life cycle?

Planning Concept

Development

System-

Level

Design

Detail

Design

Testing

and

Refinement

Production

Ramp-Up

Concept

FMEA

Design

FMEA

Process

FMEA

6

Design Project FMEA

Design FMEA’s should cover:

– all new components

– carried over components in a new environment

– any modified components

Mandatory on all control and load carrying parts

7

Design Project FMEA

“Failure” - a component or system not meeting or not functioning to the design intent

Design intent - may be stated in terms of MTBF, load or deflection, coat thickness, finish quality, etc.

“Failure” need not be readily detectable by a customer

8

FMEA Process

Identify a failure mode

Determine the possible effects or consequences of the failure

Assess the potential severity of the effect

Identify the cause of failure (to take action!)

Estimate the probability of occurrence

Assess the likelihood of detecting the failure

9

Failure Mode

Failure mode - the manner in which a component or a system, where the failure occurs (doesn’t meet design intent)

Potential failure modes

– Complete failure

– Partial failure

– Intermittent failure

– Failure over time

– Over-performance failure

10

Failure Mode

Question to be asked: “How could the component or system fail?”

Examples: Consider failure modes of a penlight’s function that is defined as “Provide light at 3 0.5 candela.” – No light

– Dim light

– Erratic blinking light

– Gradual dimming light

– Too bright

11

Failure Mode - Identification

List potential failure modes for the particular part or function – assume the failure could occur, however

unlikely

Sketch free-body diagrams (if applicable), showing applied/reaction loads. Indicate location of failure under this condition.

List conceivable potential causes of failure for each failure mode

12

Failure Mode – Effects

For each failure mode, identify the potential downstream consequences of each failure mode (the Effects)

Procedure for Potential Consequences – Beginning with a failure mode (FM-1) – list all

its potential consequences

– Separate the consequences that can result when FM-1 occurs: “Effects of FM-1”

– Write additional failure modes for remaining, depending on circumstances

Add these to list of failure modes

13

Failure Mode – Effects

Team brainstorms failure modes and effects

Example: Analyzing penlight bulb

– Premature burnout – user could trip, fall, be injured

– While used in eye examination, bulb might explode, resulting in injury

14

Failure – Severity

To analyze risk, one must first quantify the Severity of the Effects

Assume that all Effects will result if the Failure Mode occurs

Most serious Effect takes precedence when evaluating risk potential

Design and process changes can reduce severity ratings

15

Severity of Failure Rank

Hazardous – No warning: Unsafe operation,

without warning

10

Very high: Product inoperable; loss of primary

function

8, 9

High: Product operable, but at a reduced level 6, 7

Low: Product operable; comfort or convenience

items at reduced level

4, 5

Minor: Fit/finish, squeak/rattle don’t conform;

average customer notices

2, 3

No effect 1

DFMEA Severity Table

16

Failure Mode – Causes After the Effects and Severity are

addressed, the Causes of Failure Modes must be identified

In Design FMEA (DFMEA), design deficiencies that result in a Failure Mode are Causes of failure – Assumes manufacturing and assembly

specifications are met

Process FMEA (PFMEA) has similar investigation

Causes are rated in terms of Occurrence – Likelihood that a given Cause will occur AND

result in the Failure Mode

17

Failure Mode - Occurrence Estimate the probability of occurrence

on a scale of 1 -10 (consider any fail-safe controls intended to prevent cause of failure)

Consider the following two probabilities:

– probability the potential cause of failure will occur

– probability that once the cause of failure occurs, it will result in the indicated failure mode

18

Failure Occurrence - Ranking

Occurrence Criteria Ranking – Remote: unreasonable to expect failure (1)

– Low: similar designs have low failure rates (2,3)

– Moderate: similar designs have occasional moderate failure rates (4, 5, 6)

– High: similar designs have failed in the past (7,8,9)

– Very high: almost certain failure, in major way (10)

Probability of Failure Failure Rates Rank

Very High: Failure almost

inevitable

> 1 in 2 10

1 in 3 9

High: Repeated failures 1 in 8 8

1 in 20 7

Moderate: Occasional failures 1 in 80 6

1 in 400 5

1 in 2000 4

Low: Relatively few failures 1 in 15,000 3

1 in 150,000 2

Remote: Failure unlikely < 1 in 1,500,000 1

Example DFMEA Occurrence Table

20

Current Controls

Design controls grouped according to purpose

– Type 1: Controls prevent Cause or Failure

Mode from occurring, or reduce rate of occurrence

– Type 2: Controls detect Cause of Failure Mode and lead to corrective action

– Type 3: Controls detect Failure Mode before product reaches “customer”

21

Detection

Detection values are associated with Current Controls

Detection is a measure of Type 2 Controls to detect Causes of Failure, or ability of Type 3 Controls to detect subsequent Failure Modes

High values indicate a Lack of Detection

Value of 1 does not imply 100% detection

Detection Criteria: Likelihood of Detection Rank

Absolute

Uncertainty

Design Control does not detect, or there

is no Design Control

10

Very Remote Very remote chance Control will detect 9

Remote Remote chance Control will detect 8

Very Low Very low chance Control will detect 7

Low Low chance Control will detect 6

Moderate Moderate chance Control will detect 5

Moderately High Mod. High chance Control will detect 4

High High chance Control will detect 3

Very High Very high chance Control will detect 2

Almost Certain Control almost certain to detect 1

DFMEA Detection Table

23

Reducing Risk

The fundamental purpose of the FMEA is to recommend and take actions that reduce risk

Adding validation or verification can reduce Detection scoring

Design revision may result in lower Severity and Occurrence ratings

Revised ratings should be documented with originals in Design History File

24

Design Project FMEA - Results

Risk Priority Number (RPN) – RPN = Severity x Occurrence x Detection

– Mathematical product of the seriousness of a group of Effects (Severity), the likelihood that a Cause will create the failure associated with the Effects (Occurrence), and an ability to detect the failure before it gets to the customer (Detection)

– Note: S, O, and D are not equally weighted in terms of risk, and individual scales are not linear

25

Non-Intuitive Statistical Properties of the RPN

Scale

Incorrect Assumption Actual Statistical Data

The average of all RPN

values is roughly 500

The average RPN value

is 166

Roughly 50% of RPN

values are above 500

(median is near 500)

6% of all RPN values are

above 500

(median is 105)

There are 1000 possible

RPN values

There are 120 unique

RPN values

26

Criticality

Criticality = Severity x Occurrence

High Severity values, coupled with high Occurrence values merit special attention

Although neither RPN nor Criticality are perfect measures, they are widely used for risk assessment

27

Interpreting the RPN

No physical meaning to RPN Used to “bucket problems” Don’t spend a lot of time worrying

about what a measure of “42” means Rank order according to RPN Note that two failure modes may

have the same RPN for far different reasons: – S=10, O=1, D=2: RPN = 20 – S=1, O=5, D=4: RPN = 20

28

Actions

Actions taken are the important part of FMEA

Change design to reduce – Severity (redundancy?)

– Occurrence (change in design, or processes)

– Detection (improve ability to identify the problem before it becomes critical)

Assign responsibility for action

Follow up and assess result with new RPN

29

FMEA

Benefits

Systematic way to manage risk

Comprehensive

Prioritizes

Problems

Based on qualitative assessment

Unwieldy

Hard to trace through levels

Not always followed up

30

FMEA Levels

CFMEA – 1 (Concept) – Failures in the

concept (inability to achieve performance)

– Detection Ability to find the

failures (i.e., use of historical data, early models, etc.)

DFMEA – 2 (Design) – Failures in current

design (performance)

– Detection Highlighting failures

during the detail design phase

PFMEA – 3 (Process) – Failures in

production process

– Detection Finding the errors in

the production line

31

Relationships (CFMEA, DFMEA, PFMEA)

32

FMEA

33

FMEA

34

END