PFMEA Example: Automotive Stamping Process Step-by-Step
This is a complete PFMEA (Process FMEA) example for an automotive metal stamping and sub-assembly process. We follow the AIAG-VDA 7-step method, use 4M cause analysis (Man, Machine, Material, Method), score with Action Priority tables, and show how each PFMEA row links to the control plan. For how PFMEA differs from DFMEA, see our DFMEA vs PFMEA comparison. For a design FMEA example, see our automotive DFMEA example.
About This Example
Product: Structural B-pillar reinforcement bracket, high-strength steel (HSLA 590).
Process: Blanking → Progressive stamping → Piercing → Weld nut projection welding → E-coat → Final inspection → Pack/ship.
Standard: AIAG-VDA FMEA Handbook, 1st Edition (2019).
FMEA type: PFMEA (Process FMEA).
Customer: OEM, critical structural part (PPAP Level 3).
Step 1: Planning and Preparation
| 5T Element | This Example |
|---|---|
| inTent | PFMEA for B-pillar bracket manufacturing to ensure process controls meet DFMEA severity flow-down and PPAP requirements |
| Timing | Process planning phase, before pre-launch control plan finalization |
| Team | Process engineer (lead), tooling engineer, quality engineer, production supervisor, maintenance technician |
| Task | Identify process failure modes at each operation, evaluate risk, define process controls |
| Tool | FMEA software with AIAG-VDA PFMEA AP tables |
Step 2: Structure Analysis (Process Flow)
In PFMEA, the “structure” is your process flow. Each operation is an element in the hierarchy.
| Op # | Operation | Equipment |
|---|---|---|
| 10 | Receive coil / Incoming inspection | Coil storage, spectrometer |
| 20 | Blanking | Blanking press, 400T |
| 30 | Progressive stamping (form) | Progressive die, 600T press |
| 40 | Piercing (holes and locators) | Station within progressive die |
| 50 | Weld nut projection welding | Projection welder, fixture |
| 60 | E-coat (outsourced) | Third-party e-coat line |
| 70 | Final inspection and pack | CMM, go/no-go gauges, packaging line |
The process flow diagram (PFD) is a required input. This table is a simplified version. A real PFD would show material flow, decision points, and rework loops.
Step 3: Function Analysis
For each operation, define what the process step must achieve.
| Op # | Process Function | Product Characteristic |
|---|---|---|
| 10 | Verify incoming material meets spec | Material grade HSLA 590, thickness 1.8 mm ± 0.05 |
| 20 | Shear blank to size | Blank dimensions 420 × 180 mm ± 0.5 |
| 30 | Form bracket profile to drawing | Bend angles 90° ± 0.5°, no cracks, profile per GD&T |
| 40 | Pierce mounting holes and locators | Hole diameter 10.5 mm ± 0.1, true position ± 0.2 |
| 50 | Weld M8 nut to bracket | Weld strength ≥ 15 kN push-out, nut perpendicularity ± 1.5° |
| 60 | Apply e-coat for corrosion protection | Coating thickness 20-30 μm, adhesion per cross-hatch test |
| 70 | Verify dimensions and pack per customer spec | All critical dimensions per control plan, packaging per DUN standard |
Step 4: Failure Analysis with 4M Causes
PFMEA uses the 4M framework to categorize causes: Man (operator), Machine (equipment), Material (input material), Method (procedure/parameters). This ensures you consider all cause categories, not just the obvious one.
Example: Op 30 - Progressive Stamping
Function: Form bracket profile to drawing spec.
Failure mode: Part cracked at bend radius.
Effect (local): Part structurally compromised at bend.
Effect (end user): B-pillar reinforcement fails in crash. Safety: potential occupant injury.
4M cause analysis:
| 4M Category | Cause |
|---|---|
| Man | Operator loads blank with grain direction rotated 90° (grain-sensitive forming) |
| Machine | Die radius worn below minimum (polishing during maintenance reduced radius) |
| Material | Coil hardness above spec limit (supplier variation in HSLA 590 batch) |
| Method | Forming simulation not updated for current material lot. Press tonnage not validated against material variation |
Each 4M cause gets its own row in the PFMEA with independent scoring. The same failure mode can have four separate causes with different occurrence and detection profiles.
Example: Op 50 - Weld Nut Projection Welding
Function: Weld M8 nut with ≥15 kN push-out strength.
Failure mode: Weld strength below 15 kN.
Effect: Nut pulls out during vehicle assembly torquing. Line stop at OEM. In field: fastener loosening, structural joint failure.
| 4M Category | Cause |
|---|---|
| Man | Operator places nut off-center in fixture |
| Machine | Electrode cap worn, current density reduced below weld window |
| Material | Nut projection height out of spec (supplier). E-coat residue on weld zone |
| Method | Weld schedule not revalidated after electrode dress cycle change |
Step 5: Risk Analysis and Action Priority
Score each cause row independently. Use the AIAG-VDA PFMEA AP table.
Op 30 - Cracking, Material Cause (Coil hardness above spec)
| Factor | Score | Rationale |
|---|---|---|
| Severity (S) | 10 | End effect: crash performance compromised. Safety. |
| Occurrence (O) | 4 | HSLA 590 batch variation documented. 2-3 hardness exceedances per year from supplier data. |
| Detection (D) | 5 | Current control: Visual inspection for cracks at end of press line. Catches obvious cracks but misses micro-cracks in radius. |
Action Priority: S=10, O=4, D=5 → High.
Op 50 - Weak Weld, Machine Cause (Electrode wear)
| Factor | Score | Rationale |
|---|---|---|
| Severity (S) | 8 | End effect: OEM line stop + field joint failure. Major but not immediately life-threatening. |
| Occurrence (O) | 5 | Electrode wear is normal. Exceeds limit every 5,000-8,000 welds without monitoring. |
| Detection (D) | 6 | Current control: Destructive push-out test every 200 parts. Catches trends but can miss individual escapes. |
Action Priority: S=8, O=5, D=6 → High.
Step 6: Optimization
| Failure / Cause | AP | Recommended Action | Owner |
|---|---|---|---|
| Cracking / Coil hardness | High | Add incoming hardness testing per coil (Rockwell B, accept ≤ HRB 88). Reject lots above spec before blanking | Quality Eng. |
| Cracking / Coil hardness | High | Add eddy-current crack detection after forming (100% inline) | Process Eng. |
| Cracking / Die radius worn | High | Define minimum die radius wear limit in PM checklist. Add gauge check at each die maintenance | Tooling Eng. |
| Weak weld / Electrode wear | High | Install weld monitor with current/voltage feedback. Alarm on out-of-window welds | Process Eng. |
| Weak weld / Electrode wear | High | Reduce electrode dress interval from 5,000 to 3,000 welds based on wear curve data | Maintenance |
Full Example: 10 PFMEA Rows
| Op | Failure Mode | End Effect | 4M Cause | S | O | D | AP |
|---|---|---|---|---|---|---|---|
| 10 | Wrong material grade received | Part fails crash test | Material: supplier ships wrong grade | 10 | 2 | 3 | High |
| 20 | Blank dimension out of spec | Part cannot form correctly | Machine: blanking die wear | 6 | 4 | 3 | Medium |
| 30 | Crack at bend radius | Crash performance failure | Material: coil hardness above spec | 10 | 4 | 5 | High |
| 30 | Crack at bend radius | Crash performance failure | Machine: die radius worn | 10 | 3 | 5 | High |
| 30 | Bend angle out of spec | Fit issue at OEM assembly | Method: press tonnage not validated | 6 | 3 | 4 | Medium |
| 40 | Hole position out of tolerance | Bolt pattern misalignment at OEM | Machine: progressive die locator pin wear | 7 | 4 | 3 | Medium |
| 50 | Weld strength below 15 kN | OEM line stop, field joint failure | Machine: electrode cap worn | 8 | 5 | 6 | High |
| 50 | Weld strength below 15 kN | OEM line stop, field joint failure | Man: nut placed off-center | 8 | 3 | 4 | Medium |
| 60 | E-coat thickness below 20 μm | Corrosion in field, warranty claim | Method: e-coat supplier process drift | 5 | 4 | 5 | Medium |
| 70 | Out-of-spec part shipped | OEM assembly issue, sort/rework | Man: inspector misses defect under fatigue | 7 | 3 | 5 | Medium |
Linking PFMEA to the Control Plan
Every PFMEA row with a detection control must map to a control plan entry. The control plan specifies the how, how often, and what to do if out of spec.
| PFMEA Row | Control Plan Entry | Method | Frequency | Reaction Plan |
|---|---|---|---|---|
| Op 10: Wrong material | Incoming material verification | Spectrometer check + cert review | Every coil | Quarantine, notify supplier |
| Op 30: Crack at bend | Inline crack detection | Eddy-current sensor | 100% | Auto-reject, stop if 3 consecutive |
| Op 40: Hole position | Hole position check | CMM or go/no-go gauge | 5 pc/hr + SPC | Stop, adjust, re-verify Cpk |
| Op 50: Weld strength | Weld monitor + destructive test | Inline current/voltage + push-out every 200 pc | 100% monitor, 200 pc destruct | Quarantine lot, investigate |
| Op 60: E-coat thickness | Coating thickness check | Magnetic gauge per batch cert | Per incoming batch from supplier | Reject batch, escalate to supplier |
The PFD, PFMEA, and Control Plan form a linked triad. OEM auditors check that every critical PFMEA row has a corresponding control plan entry with a defined reaction plan. For what Ford specifically checks, see our Ford PPAP FMEA post.
How DFMEA Severity Flows In
The DFMEA for this bracket identified design failure modes with S=10 (crash performance). That severity carries into the PFMEA:
- DFMEA: “Bracket fails to maintain structural integrity under side-impact load” - S=10
- PFMEA: “Crack at bend radius” has S=10 because the process failure leads to the same design failure effect
- Any process step that could compromise the crash-critical characteristic inherits the DFMEA severity
This is the DFMEA → PFMEA linkage that AIAG-VDA requires and that auditors verify. If your DFMEA has S=10 for a failure mode but the corresponding PFMEA rows show S=6, the disconnect will be questioned. See our automotive DFMEA example for the design side.
Common PFMEA Mistakes
Listing only one 4M cause per failure mode. Most process failures have multiple contributing causes across Man, Machine, Material, and Method. Analyzing only the obvious one (usually Machine) misses the others.
Scoring detection for the wrong control. Detection should reflect what you have today, not what you plan to add. Score current state, then add actions to improve.
No link to the process flow diagram. Every PFMEA operation should map to a PFD step. Operations in the PFMEA that do not appear on the PFD, or PFD steps with no PFMEA analysis, are audit findings.
Generic controls. “100% inspection” tells the auditor nothing. What inspection? By what method? At what capability? “CMM check of hole true position, 5 pc/hr, Cpk ≥ 1.67” is defensible.
Forgetting the reaction plan. The control plan must specify what happens when a control fails. “Stop and adjust” is a start. “Quarantine all parts since last good check, 100% sort, root cause investigation per 8D” is complete.
How Tacit AI Approaches This
Tacit AI generates PFMEA drafts from your process flow, work order history, and engineering documents.
Process flow to PFMEA in hours. Provide your PFD and process specs. Tacit AI suggests failure modes for each operation using 4M cause analysis, informed by your work order data for similar processes. Engineers review and refine.
DFMEA severity flows automatically. When the DFMEA identifies a high-severity design failure mode, the corresponding PFMEA operations inherit that severity. Gaps are flagged: if a safety-critical design characteristic has no process control in the PFMEA, it surfaces.
Control plan generates from PFMEA. Detection controls defined in the PFMEA populate the control plan with methods, frequencies, and reaction plans. The PFD → PFMEA → Control Plan triad stays in sync.
Work orders close the loop. When production work orders reveal process failures the PFMEA did not predict, they are flagged as gaps. When a controlled process step still produces defects, the control effectiveness is questioned. Your PFMEA stays current with production reality.
Book a working session with your process flow diagram. See what Tacit AI generates for your manufacturing process and compare against your current PFMEA. For PFMEA fundamentals, see our DFMEA vs PFMEA comparison.