Process FMEA Step by Step: The AIAG-VDA 7-Step Method With a Worked Example

A process FMEA that actually reduces risk follows a structured sequence. The AIAG & VDA FMEA Handbook defines seven steps, each building on the previous one. Skip a step or rush through it, and the downstream analysis will be shallow. This guide walks through all seven steps with a running example from an injection molding process so you can see how each step produces inputs for the next.

Before You Start: Inputs for a Process FMEA

The PFMEA does not start from a blank sheet. Gather these inputs before the first team session:

• Process flow diagram (PFD) — The sequence of manufacturing operations. Every operation number in your PFD becomes a row in the PFMEA’s structure analysis.
• DFMEA outputs — Design failure modes that become potential process causes. The AIAG-VDA handbook documents twelve linkage points between DFMEA and PFMEA.
• Lessons learned — Warranty returns, 8D reports, scrap/rework data, and prior FMEAs from similar products.
• Customer requirements — Special characteristics (Critical and Significant) identified on drawings and specifications.
• Cross-functional team — Typically the quality engineer (facilitator), process engineer, tooling engineer, maintenance technician, and machine operator. Four to eight people per session.

Step 1: Planning and Preparation

Define the scope of the PFMEA. This means identifying:

• Which product or product family the PFMEA covers
• Which manufacturing process (or processes) are in scope
• Whether you are starting from scratch or updating a foundation FMEA from a similar product
• The timeline—when must the PFMEA be complete relative to PPAP submission?

Running example: A Tier 2 automotive supplier is launching a new ABS sensor housing produced by injection molding. The PFMEA covers the molding, trimming, and assembly operations. The team starts with a foundation FMEA from a similar housing launched two years ago.

Step 2: Structure Analysis

Build the process structure by decomposing the manufacturing process into three levels:

1. Process item (top level) — The product being manufactured. Example: ABS sensor housing.
2. Process step (middle level) — Each operation from the PFD. Example: Op 10 Injection Molding, Op 20 Gate Trimming, Op 30 Insert Assembly.
3. Process work element (lower level) — Sub-steps within each operation. Example: under Op 10: material drying, melt injection, packing/holding, cooling, ejection.

This three-level structure is a key AIAG-VDA addition. Older FMEA approaches listed process steps in a flat table; the structure analysis forces you to think about sub-step interactions that flat lists miss.

Step 3: Function Analysis

For each element in the structure, document its intended function using verb-noun format:

Functions at the work element level are where most teams cut corners. Document them anyway—they drive the failure mode identification in Step 4. If you cannot state the function, you cannot systematically identify how it fails.

Step 4: Failure Analysis

This is the core of the PFMEA and the step that consumes the most team time. For each function, ask three questions:

1. How can this function fail? (failure mode)
2. What happens when it fails? (effect, evaluated at the process item level and the end-user level)
3. Why does it fail? (cause, at the work element level)

Running example (Op 10, packing/holding phase):

Notice the cause is always at the work element level (the mechanism), not a surface-level label like “operator error.” Getting to root-level causes is one of the most common FMEA quality issues. If the team writes “operator error” as a cause, ask “why did the operator make the error?” until you reach a process mechanism.

Step 5: Risk Analysis

Rate each failure chain on three scales (1–10):

• Severity (S) — Based on the worst-case effect. The short shot causing field failure rates severity 8–9 (loss of primary function / safety concern depending on application).
• Occurrence (O) — Based on the likelihood of the cause given current prevention controls. If check ring replacement is on a fixed PM schedule and historical data shows 2 occurrences per 100,000 shots, occurrence is around 3–4.
• Detection (D) — Based on the effectiveness of current detection controls. If 100% parts pass through a vision system checking for short shots, detection is 2–3.

Then determine risk priority using either RPN or Action Priority (AP). For a detailed walkthrough of how to assign these ratings and calculate the result, see our guide on calculating RPN in FMEA. For a comparison of the two methods and when to use which, see Action Priority vs. RPN.

Step 6: Optimization

For every failure chain rated High (AP) or above your RPN threshold, assign a recommended action with four elements:

1. What — The specific action (e.g., “Add cavity pressure sensor with automatic reject at < 85% pack pressure target”)
2. Who — Responsible individual (name, not department)
3. When — Target completion date
4. Type — Prevention (reduces occurrence) or detection (reduces detection rating). The ASQ and AIAG-VDA standard both emphasize: prioritize prevention over detection. Error-proofing (poka-yoke) that eliminates the cause is always preferred over inspection that catches the defect after the fact.

Running example: For the short shot failure mode (S=9, O=4, D=3, RPN=108, AP=High), the team assigns two actions:

• Prevention: Add cavity pressure monitoring with automatic shot reject if pressure drops below 85% of target. Owner: Process Engineer (J. Martinez). Due: Week 24. Expected effect: occurrence drops to 2.
• Detection: Upgrade vision system to include volumetric fill analysis. Owner: Quality Engineer (R. Chen). Due: Week 26. Expected effect: detection drops to 1.

Step 7: Results Documentation

After recommended actions are implemented:

1. Verify implementation. Confirm the action was actually completed, not just planned. Check the cavity pressure sensor is installed and the reject logic is validated.
2. Re-rate S, O, D. Severity stays at 9 (the effect has not changed). Occurrence drops from 4 to 2 (cavity pressure monitoring prevents most short shots). Detection drops from 3 to 1 (volumetric fill analysis catches virtually all escapes).
3. Recalculate risk. New RPN = 9 × 2 × 1 = 18. New AP = Medium (down from High). Risk reduction confirmed.
4. Update the control plan. Add the cavity pressure monitoring and vision system upgrade to the control plan for this operation. This is the linkage auditors will check—the PFMEA and control plan must tell the same story.

Use our RPN and Action Priority calculator to quickly compare before-and-after risk ratings as you document results.

Common Mistakes That Undermine Process FMEAs

Three patterns that consistently produce weak PFMEAs:

• Skipping the structure and function analysis. Teams jump straight to failure modes without documenting what each process step is supposed to do. This produces incomplete failure mode lists because you cannot systematically identify how something fails if you have not defined what it is supposed to do.
• Writing “operator error” as a cause. This is a symptom, not a root cause. Push to the mechanism: “Operator loads part in wrong orientation because fixture does not have asymmetric locating features (no poka-yoke).”
• Never re-rating after actions. If you assign recommended actions but never close the loop with revised ratings, the FMEA becomes a list of intentions rather than a record of risk reduction. Auditors per IATF 16949 will flag this.

Keeping the PFMEA Alive

A PFMEA completed during APQP Phase 3 and never updated is a compliance artifact, not a risk management tool. Trigger a PFMEA review whenever: a process change occurs (new equipment, material, or supplier), a customer complaint traces back to a process failure mode, an engineering change notice (ECN) modifies the product, or a new failure mode is discovered during production. The teams that get the most value from FMEA are the ones that treat it as a living document—updated continuously, not filed away after PPAP.