Key takeaways
Short answer: DFMEA and PFMEA are two applications of the same FMEA method to two different questions. Design FMEA (DFMEA) examines how a product's design could fail — its failure modes, effects, and risks — so the design can be improved before anything is made. Process FMEA (PFMEA) examines how the manufacturing process could fail to produce the product correctly. DFMEA asks will the design fail in use; PFMEA asks will we make it right. One protects the product, the other the process. For the parent method, see FMEA vs fault tree analysis.
Design FMEA analyses the ways a product's design could fail and the consequences of those failures, so the design itself can be improved before it ever reaches production. Working from the product's functions, a DFMEA team identifies potential design failure modes — a component that could fracture under load, a material that could degrade, a geometry that could allow misassembly — assesses their effects and severity, and prioritises by risk, then drives design changes to remove or mitigate the worst. Its focus is the product as designed: will this design perform reliably and safely in use? DFMEA happens early, during design, because the cheapest place to eliminate a failure is on the drawing board, before tooling, process, and production are committed.
Process FMEA analyses the ways a manufacturing process could fail to make the product correctly. Working from the process steps, a PFMEA team identifies potential process failure modes — a step that could be done out of sequence, a setting that could drift, a fixture that could allow a part in backwards — assesses their effects (often a defect or nonconformity) and severity, and prioritises by risk, then drives process controls and improvements to prevent or detect them. Its focus is the process as planned: will we actually build this product right, consistently? PFMEA happens during process planning, after the design is reasonably settled, and it is where mistake-proofing, controls, and inspection points are designed into the process to catch or prevent each failure mode.
The clean distinction is the subject of the analysis. DFMEA analyses the product design — could the thing itself, as designed, fail? PFMEA analyses the manufacturing process — could we fail to make the thing correctly? A design can be perfect yet badly producible, and a process can be capable yet building a flawed design — which is why both analyses are needed, on different subjects. They also sit at different points in time: DFMEA in design, PFMEA in process planning. And they inform each other — a design failure mode identified in DFMEA may drive a process control in PFMEA, and a hard-to-produce feature flagged in PFMEA may send feedback to the design. Same method, two targets, two moments.
Consider a bracket. The DFMEA, during design, asks how the bracket itself could fail: could it fatigue and crack under cyclic load? The team identifies that risk, judges it severe, and changes the design — a thicker section, a fillet to reduce stress — eliminating the failure mode in the product. Later, the PFMEA, during process planning, asks how making the bracket could go wrong: could the operator fit it backwards, could the press setting drift and produce an undersize part? The team adds a mistake-proofing fixture that only accepts the correct orientation and a control on the press setting. DFMEA made the design robust; PFMEA made the process incapable of building it wrong. Both were needed — a great design poorly produced still fails.
DFMEA and PFMEA are sequential and linked. DFMEA comes first, in design, and its outputs flow into the PFMEA: design failure modes and critical characteristics identified in the DFMEA become things the process must control, so the PFMEA inherits and acts on them. The PFMEA, in turn, can feed back to design — if a feature proves genuinely impossible to produce reliably, that is a signal to revisit the design. Done well, the two form a chain from product to process: design out what failures you can, then process-proof against the rest. Skipping the DFMEA leaves the process trying to compensate for a fragile design; skipping the PFMEA leaves a good design at the mercy of an uncontrolled process. The handoff between them is where robust products and capable processes meet.
Both FMEAs prevent, upstream, the failures that later show up as OEE losses on the floor. PFMEA is the more directly connected: by identifying and controlling process failure modes, it prevents the defects and stoppages that hit the quality and availability factors — the same source-quality logic as poka-yoke and jidoka. DFMEA contributes by ensuring the product design itself is robust and producible, so the process is not fighting an inherently fragile design. Every failure mode designed out or controlled before production is a loss that never reaches OEE — the cheapest kind of improvement, made before the first part is built. Both are upstream cousins of FMEA and fault tree analysis.
Fabrico closes the loop between these upstream analyses and downstream reality. The real downtime and defect data it captures shows which failure modes actually occur in production — feedback that makes the next revision of a PFMEA, and sometimes the DFMEA, grounded in what really happens rather than what was anticipated. A failure mode that keeps appearing in the OEE loss data is a signal that a control is missing or a design is marginal. Book a demo to feed real production data back into your FMEAs.
DFMEA (Design FMEA) analyses how a product's design could fail, to improve the design before production. PFMEA (Process FMEA) analyses how the manufacturing process could fail to make the product correctly. DFMEA targets the product; PFMEA targets the process.
DFMEA comes first, during design, because the cheapest place to eliminate a failure is before tooling and process are committed. PFMEA follows during process planning, and inherits the critical characteristics and failure modes the DFMEA identified.
Usually yes. A robust design can still be built wrong by an uncontrolled process, and a capable process can still build a fragile design. The two analyse different subjects — product and process — and a great result needs both.
They are linked: DFMEA outputs, such as critical characteristics and design failure modes, flow into the PFMEA as things the process must control. The PFMEA can also feed back to design if a feature proves impossible to produce reliably.
Both prevent failures upstream that would otherwise become OEE losses. PFMEA directly prevents process defects and stoppages that hit the quality and availability factors, while DFMEA ensures the design is robust and producible so the process is not fighting it.
Вход
Блог
