8.2 Failure Mode and Effects Analysis (FMEA/FMECA)

8.2 Failure Mode and Effects Analysis (FMEA/FMECA)

Standardized under IEC 60812 and referenced in IEC 31010, Failure Mode and Effects Analysis (FMEA) is a rigorous, inductive, bottom-up quantitative and qualitative risk assessment technique. Originally developed by the military (MIL-STD-1629A) and extensively adopted across aerospace, automotive, manufacturing, and healthcare sectors, FMEA provides a systematic process for identifying how equipment, software, or business processes can fail, evaluating the operational impacts of those failures, and prioritizing corrective actions.

When quantitative failure rate data is integrated with severity ratings, the technique expands into Failure Mode, Effects, and Criticality Analysis (FMECA).

---

1. System Decomposition and Methodology

FMEA begins with system decomposition. Analysts break a complex process or system down into subsystems, assemblies, and individual functional components. For each component, the team systematically evaluates four core dimensions:

1. Item / Function: The intended operation of the component within the system.
2. Failure Mode: The specific physical or operational manner in which the component fails to perform its intended function (e.g., failure to open, short circuit, software memory leak, seal leakage).
3. Failure Mechanism & Cause: The underlying physical, chemical, or operational driver causing the failure (e.g., thermal fatigue, oxidation, input validation failure).
4. Failure Effects: The consequences of the failure mode across three hierarchical tiers:
   • Local Effect: Impact restricted directly to the specific component.
   • Next Higher Level Effect: Impact on the immediate assembly or sub-system.
   • End Effect: Impact on the total system, end-user, environment, or enterprise safety.

---

2. The FMEA Scoring Framework: Severity, Occurrence, and Detection

To quantify risk, FMEA assigns standardized numerical scores ranging from 1 to 10 across three distinct risk parameters:

Severity ($S$)

Measures the intensity of the worst-case End Effect of a failure mode on the system or user.

• 1 (Negligible): No discernible effect on system performance or safety.
• 4 to 6 (Moderate): System performance degraded; minor downtime or customer dissatisfaction.
• 9 to 10 (Catastrophic): High safety hazard, regulatory non-compliance, or total system destruction occurring without prior warning.

Occurrence ($O$)

Measures the estimated frequency or probability that a specific failure mode will occur during the planned operational lifespan of the system.

• 1 (Extremely Remote): Probability $< 1 \text{ in } 1,500,000$ operating cycles ($< 10^{-6}$).
• 5 (Moderate): Occasional failures; approximately $1 \text{ in } 2,000$ operating cycles.
• 10 (Almost Inevitable): High failure rate; $> 1 \text{ in } 2$ operating cycles.

Detection ($D$)

Measures the capability of existing internal controls, automated diagnostic monitors, or inspection procedures to detect the failure mode before the component is deployed or before the End Effect manifests.

• 1 (Certain Detection): Automated, 100% fail-safe inline diagnostic detection.
• 5 (Moderate Detection): Manual inspection or periodic testing with moderate detection likelihood.
• 10 (Absolute Uncertainty): No control mechanism exists; failure mode is completely undetectable until catastrophic failure occurs.

---

3. Risk Priority Number (RPN) Calculation and Limitations

The fundamental metric generated by an FMEA is the Risk Priority Number (RPN), calculated as the mathematical product of Severity, Occurrence, and Detection:

$\text{RPN} = S \times O \times D$

Because each factor ranges from 1 to 10, the RPN spans a dynamic scale from 1 to 1,000.

Mathematical Comparison and RPN Pitfalls

A common error in enterprise risk management is using an arbitrary RPN threshold (e.g., "remediate only items with $\text{RPN} > 120$") as the sole criterion for corrective action. Relying solely on RPN can mask critical safety hazards due to mathematical non-uniqueness:

Consider two distinct component failure modes:

• Failure Mode Alpha: $S = 10, O = 2, D = 6 \implies \text{RPN} = 10 \times 2 \times 6 = 120$
• Failure Mode Beta: $S = 4, O = 5, D = 6 \implies \text{RPN} = 4 \times 5 \times 6 = 120$

Although both yield an identical RPN of 120, Failure Mode Alpha represents a potential catastrophic fatality ($S = 10$), whereas Failure Mode Beta represents a moderate operational inconvenience ($S = 4$). ISO 31000 best practices mandate that any item with a high Severity score ($S \ge 8$) must receive mandatory risk treatment and engineering re-design regardless of how low its Occurrence or Detection scores might be.

---

4. FMECA: Quantitative Criticality Analysis

Failure Mode, Effects, and Criticality Analysis (FMECA) enhances basic FMEA by adding quantitative failure rate calculations to build a rigorous Criticality Index.

Failure Mode Criticality ($C_m$)

The criticality metric for a specific failure mode ($C_m$) is calculated using component failure rate data:

$C_m = \alpha \times \beta \times \lambda_p \times t$

where:

• $\alpha$ = Failure Mode Ratio: The fraction of the total component failure rate attributed to this specific failure mode (where $\sum \alpha_i = 1.0$).
• $\beta$ = Conditional Probability of Loss: The conditional probability that the failure mode will actually result in the identified End Effect (ranging from 0 to 1.0).
• $\lambda_p$ = Part Failure Rate: The total failure rate of the component (typically failures per million operating hours, $10^{-6} \text{ hrs}$).
• $t$ = Operating Time: The total operational mission duration (in hours).

Worked FMECA Calculation

A critical cooling pump seal operates continuously for one year ($t = 8,760 \text{ hours}$).

• Total Part Failure Rate ($\lambda_p$): $20 \text{ failures per } 10^6 \text{ hours} = 2.0 \times 10^{-5} \text{ failures/hour}$.
• Failure Mode under evaluation: Fluid Seal Leakage.
• Failure Mode Ratio ($\alpha$): $0.35$ (35% of pump failures are seal leaks).
• Conditional Loss Probability ($\beta$): $0.80$ (80% of seal leaks lead to total pump shutdown).

Calculating $C_m$: $C_m = 0.35 \times 0.80 \times (2.0 \times 10^{-5}) \times 8,760$ $C_m = 0.28 \times 0.00002 \times 8760 = 0.049056$

This quantified criticality index ($C_m = 4.91 \times 10^{-2}$) allows risk engineers to plot the failure mode precisely on a quantitative Criticality Matrix (plotting Probability/Frequency against Severity) to justify safety investments to executive leadership.

Representative FMEA Worksheet

The table below illustrates a standard FMEA operational worksheet tracking baseline scoring, recommended engineering controls, and post-treatment re-evaluations: