8.1 Failure Mode and Effects Analysis (FMEA)

8.1 Failure Mode and Effects Analysis (FMEA)

Exam Tip: ASQ CMQ/OE exam questions frequently test the distinction between Design FMEA (DFMEA) and Process FMEA (PFMEA), the mathematical calculation of the Risk Priority Number (RPN), and the fundamental rule that Severity can only be reduced by redesigning the product or process, never by adding inspection.

Overview and Historical Context of FMEA

Failure Mode and Effects Analysis (FMEA) is a risk assessment methodology designed to identify potential failures in a product, service, or process before they occur. Rather than reacting to quality failures after production or field deployment, FMEA provides a structured, quantitative framework for evaluating risk during the design and manufacturing planning phases.

FMEA originated in the United States Military with the publication of MIL-STD-1629 ("Procedures for Performing a Failure Mode, Effects and Criticality Analysis") in November 1949. The procedure was subsequently adopted by NASA during the Apollo space program in the 1960s to minimize single-point failure modes in manned spaceflight. In the late 1970s, Ford Motor Company introduced FMEA to the automotive industry to improve vehicle safety and reliability. Today, FMEA is governed globally by standards maintained by the Automotive Industry Action Group (AIAG) and the German Verband der Automobilindustrie (VDA), which published their harmonized AIAG-VDA FMEA Handbook in 2019.

---

DFMEA vs. PFMEA: Core Structural Differences

Although both variants employ similar rating scales, Design FMEA and Process FMEA address distinct stages of the product lifecycle and evaluate different failure mechanisms.

Design FMEA (DFMEA)

DFMEA assumes the product will be manufactured according to drawing specifications. It evaluates potential failure modes caused by design deficiencies—such as choosing an alloy prone to stress corrosion cracking or underestimating thermal dissipation requirements in electronic enclosures.

Process FMEA (PFMEA)

PFMEA assumes the product design is feasible and valid. It focuses exclusively on how process steps could fail to execute the design intent—such as an automated torque wrench applying insufficient clamping force due to air line pressure drops or an assembly line technician misinterpreting a visual work instruction.

---

The 1 to 10 Rating Scales: Severity, Occurrence, and Detection

Risk assessment in FMEA relies on three discrete criteria, each rated on an ordinal scale from 1 (lowest risk) to 10 (highest risk).

1. Severity (S)

Severity estimates the seriousness of the effect of a potential failure mode on the end customer or downstream process. Severity is assessed independently of how likely the failure is to occur.

• 10 (Hazardous without Warning): Failure affects safe operation or involves non-compliance with governmental regulations without prior warning.
• 9 (Hazardous with Warning): Safety hazard or regulatory non-compliance accompanied by prior warning signs.
• 7 - 8 (High / Major System Loss): Primary function rendered inoperable (e.g., vehicle engine stalls; medical device fails to start).
• 4 - 6 (Moderate / Minor System Loss): Secondary function degraded or customer experiences significant dissatisfaction.
• 2 - 3 (Low / Slight Effect): Minor cosmetic defect, slight fit/finish flaw, or minor noise that is noticed by picky customers.
• 1 (None): No discernible effect on product function, safety, or customer perception.

2. Occurrence (O)

Occurrence measures the likelihood or frequency that a specific cause will generate the failure mode during the planned operational lifecycle or production volume.

• 10 (Very High / Persistent): Failure is almost inevitable (>= 1 in 2 occurrences; process capability Cpk < 0.33).
• 7 - 8 (High / Frequent): Repeated process failures (approx. 1 in 20 to 1 in 80; Cpk approx. 0.83 to 1.00).
• 4 - 6 (Moderate / Occasional): Moderate failure rates (approx. 1 in 400 to 1 in 2,000; Cpk approx. 1.17 to 1.33).
• 2 - 3 (Low / Infrequent): Isolated failures (approx. 1 in 15,000 to 1 in 150,000; Cpk approx. 1.50 to 1.67).
• 1 (Extremely Unlikely / Remote): Failure is virtually eliminated through automated error-proofing (< 1 in 1,500,000; Cpk >= 2.00).

3. Detection (D)

Detection evaluates the capability of current design verification controls or manufacturing inspection systems to detect the cause or failure mode before the product leaves the design stage or manufacturing facility.

• 10 (Absolute Uncertainty): No inspection mechanism exists, or defect is fundamentally undetectable prior to customer receipt.
• 8 - 9 (Very Low / Manual Visual Inspection): Relies on operator visual inspection or manual gauging after processing.
• 5 - 6 (Moderate / Post-Process Gauging): Automated gauging or 100% manual inspection at end of line.
• 2 - 3 (High / In-Process Automatic Detection): Automatic detection in-station that stops line or prevents part transfer.
• 1 (Certain Detection / Error-Proofing): Poka-yoke device makes defect generation physically impossible (e.g., asymmetrical keying).

---

Calculating and Interpreting the Risk Priority Number (RPN)

The traditional metric for prioritizing risk in FMEA is the Risk Priority Number (RPN), computed as the product of the three individual scores:

$\text{RPN} = \text{Severity (S)} \times \text{Occurrence (O)} \times \text{Detection (D)}$

Because each score ranges from 1 to 10, the resulting RPN ranges from a minimum of 1 to a theoretical maximum of 1,000.

Mathematical Limitations of RPN

Historically, organizations established arbitrary cutoff thresholds (e.g., "any RPN greater than 100 requires mandatory action"). Quality professionals and ASQ standards emphasize that arbitrary RPN thresholds are mathematically flawed because identical RPN values can conceal radically different risk profiles:

• Scenario A: S = 10, O = 2, D = 2 => RPN = 40 (Hazardous safety risk with low occurrence).
• Scenario B: S = 2, O = 4, D = 5 => RPN = 40 (Minor cosmetic flaw with moderate occurrence).

In Scenario A, a Severity of 10 demands immediate corrective action despite the low RPN of 40. Relying solely on RPN thresholding would treat both scenarios equally.

AIAG-VDA Action Priority (AP) Framework

To overcome RPN limitations, the 2019 AIAG-VDA harmonized standard introduced Action Priority (AP) logic tables. AP prioritizes risk based on qualitative combinations of S, O, and D into three discrete levels:

1. High Priority (H): Highest risk; mandatory requirement to identify effective mitigation or document engineering justification.
2. Medium Priority (M): Moderate risk; team should identify action plans to reduce risk or improve controls.
3. Low Priority (L): Low risk; current controls are deemed adequate, though continuous improvement actions are optional.

---

Developing and Executing Effective Mitigation Action Plans

When a failure mode warrants corrective action (via high RPN or High Action Priority), the multidisciplinary FMEA team must implement targeted interventions:

1. Reducing Severity (S): Requires physical redesign of the product or process requirement. Examples include substituting a toxic chemical with a non-hazardous compound or adding a redundant braking system.
2. Reducing Occurrence (O): Focuses on eliminating failure causes. Examples include improving process capability (Cpk expansion), optimizing machine tooling parameters, or implementing preventative maintenance cycles.
3. Reducing Detection (D): Focuses on improving control mechanisms. Examples include implementing machine vision automated optical inspection (AOI) or mechanical poka-yoke guide pins.

After implementing mitigation actions, the team must re-assess Severity, Occurrence, and Detection to calculate a Revised RPN. The FMEA is a living document that must be updated continuously throughout the product lifecycle.