Mastering FMEA: Reduce Risks & Improve Quality | Welle Inspection

A successful FMEA is not an informal meeting; it is a rigorous process executed by a dedicated cross-functional team. This team must include representatives from design, manufacturing, materials, quality, and maintenance to ensure all perspectives are considered.

1.Define the Scope and Boundaries

The first step is to establish a clear focus. For a Design FMEA (DFMEA), this involves using a Boundary Diagram to visually define what is included in the analysis and what is not. For a Process FMEA (PFMEA), a detailed Process Flow Diagram is the starting point. A clearly defined scope is critical to prevent the analysis from becoming unmanageably broad.

2.Analyze the Structure and Functions:

• Structure Analysis: Break down the system or process into its core components. For a DFMEA, this could be a system, subsystem, and individual components. For a PFMEA, it’s the sequence of process steps.
• Function Analysis: For each element in the structure, precisely define its intended functions and requirements. A vague function like “seal the container” is insufficient. A better function is: “Maintain a hermetic seal against an internal pressure of 2 bar at temperatures from -20°C to 80°C for a service life of 10 years.”

3.Identify Potential Failure Modes:

 For each function, the team identifies how it could fail to perform its intended duty. A failure mode is the specific manner of failure. Examples include:

• Fracture (of a bracket)
• Corrosion (on a terminal)
• Incorrect Algorithm Logic (in software)
• Torque Below Specification (in an assembly step)

4.Analyze Effects and Causes of Failure

• Effects (The Consequences): What happens if the failure occurs? Analyze effects at three levels: the local effect (on the component itself), the system effect (on the next level up), and the end-user effect (impact on safety, performance, or regulatory compliance). The end-user effect determines the Severity rating.
• Causes (The Root Why): For each failure mode, brainstorm all plausible root causes. Don’t stop at the first “why.” A cause like “operator error” is a starting point, not a conclusion. Why did the operator make the error? Was it poor training, confusing work instructions, or faulty tooling?

5.Analyze Current Controls (Prevention & Detection)

• Prevention Controls: What existing systems or practices are in place to prevent the cause from occurring? Examples include robust design verification, error-proofing (Poka-Yoke), or stringent material specifications.
• Detection Controls: If the cause does occur, what systems are in place to detect it before the product leaves the facility? This includes inspections, tests, and monitoring systems.

6.Assign Risk Ratings (S, O, D)

 Using a standardized 1-10 rating scale defined by your organization, the team rates each failure chain:

• Severity (S): Rates the seriousness of the end-user effect. A 10 is reserved for failures that could lead to injury or violate safety regulations. Severity can only be reduced by a design change.
• Occurrence (O): Rates the likelihood that the root cause will occur, considering your current prevention controls. This rating should be based on data (e.g., Cpk, field data) whenever possible. To provide context, organizations often create a rating table that links scores to specific failure rates. For example:
• Low Occurrence (e.g., Score 1-2): Corresponds to extremely low failure rates, such as 1 in 150,000. These are isolated incidents that a customer would likely never notice.
• Moderate Occurrence (e.g., Score 3-4)

 Represents low-level failure rates, like 1 in 15,000. These might cause minor performance degradation but are generally managed by robust process controls.

• Medium Occurrence (e.g., Score 5-6): Indicates occasional failures, perhaps around 1 in 2,000 to 1 in 500. At this level, performance may be noticeably affected, but safety is typically maintained. These are often the “nuisance” failures that drive customer complaints.
• High Occurrence (e.g., Score 7-8): Suggests frequent failures, such as 1 in 100. This level indicates that process controls are insufficient, and performance is significantly impacted.
• Very High Occurrence (e.g., Score 9-10): Represents almost certain or inevitable failures, like 1 in 20 or higher. This signifies a fundamental flaw in the process or design that must be addressed immediately.
• Detection (D): Rates the effectiveness of your detection controls at finding the cause or failure mode before it leaves your facility. A 10 means you have no reliable way to detect it.

7.Prioritize Actions with Action Priority (AP)

The traditional Risk Priority Number (RPN = S×O×D) is no longer the primary tool for prioritization under the AIAG-VDA FMEA standard. RPN has known mathematical flaws (e.g., 6×6×1 = 36 and 9×2×2 = 36 represent vastly different risks).
The industry has moved to Action Priority (AP). Based on the specific combinations of S, O, and D ratings, the AP table classifies the risk as High, Medium, or Low Priority for action. This provides a much more logical and safety-focused method for directing resources to where they are needed most.

8.Develop and Implement an Action Plan

 For all High and Medium AP items, the team must define concrete actions. The plan must specify what will be done, who is responsible, and the target completion date. The goal is to drive down the Occurrence or Detection ratings. After actions are completed, the team re-scores O and D and recalculates the AP to confirm the risk has been effectively mitigated. This creates a closed-loop corrective action system.