FMEA RPN Calculator

Calculate Risk Priority Number from Severity, Occurrence, and Detection ratings. Includes AIAG-VDA Action Priority classification.

Single Failure Mode
Multiple Failure Modes

Risk Ratings

5
Moderate — customer dissatisfied
5
Moderate — occasional failures
5
Moderate — controls may detect
125
Risk Priority Number (S × O × D)
MEDIUM PRIORITY
Severity
5
Moderate
Occurrence
5
Moderate
Detection
5
Moderate
Criticality (S×O)
25
Severity × Occurrence

Rating Reference

RatingSeverityOccurrenceDetection
1No effectFailure eliminatedAlmost certain detection
2-3Minor — slight inconvenienceRemote — rare failuresHigh — likely detected
4-6Moderate — customer dissatisfiedModerate — occasionalModerate — may detect
7-8High — significant impactHigh — repeated failuresLow — unlikely detected
9-10Very high — safety/regulatoryVery high — near certainVery low — no detection

What Is RPN (Risk Priority Number)?

Risk Priority Number is the core metric of Failure Mode and Effects Analysis (FMEA). It quantifies the risk of each potential failure mode by multiplying three factors: Severity (impact on the customer), Occurrence (how often the cause happens), and Detection (how well current controls catch it before it reaches the customer). RPN = S × O × D, producing a score from 1 to 1,000.

FMEA originated in the U.S. military (MIL-P-1629, 1949) and was adopted by the automotive industry through AIAG. Today it is required by IATF 16949, AS9100 (aerospace), and ISO 14971 (medical devices). RPN provides a repeatable, team-based method for prioritizing corrective actions on the failure modes that matter most.

RPN vs. Action Priority (AIAG-VDA)

Traditional RPN has a well-known weakness: it treats S, O, and D as interchangeable. A failure with S=10, O=1, D=1 (catastrophic but rare, easily detected) gets RPN=10 — the same as S=1, O=2, D=5 (trivial). The 2019 AIAG-VDA FMEA Handbook addressed this by introducing Action Priority (AP), a lookup-based classification that weighs Severity more heavily. Any failure with S ≥ 9 and O ≥ 4 is automatically High priority regardless of Detection. This calculator shows both RPN and AP so you can compare the traditional and modern approaches.

How to Reduce a High RPN

Severity is usually fixed by the product design, so focus on the two factors you can control:

FMEA in Practice

Effective FMEA is a living document, not a one-time exercise. Update it when designs change, new failure modes are discovered, or corrective actions are implemented. For persistent FMEA tracking with Bayesian evidence linking, automatic RPN recalculation, and PDF export, use the full FMEA tool in Svend.

Need a full FMEA?

Build persistent FMEAs with S/O/D scoring, RPN tracking, Bayesian evidence linking, and automatic risk recalculation. Export to PDF.

Start Free

Frequently Asked Questions

What is RPN in FMEA?
RPN (Risk Priority Number) is calculated as Severity × Occurrence × Detection, where each factor is rated 1–10. It ranges from 1 (lowest risk) to 1,000 (highest risk) and helps prioritize which failure modes need corrective action first.
What is a good RPN threshold?
Traditional FMEA uses RPN > 100 or > 125 as a threshold for required action. However, the AIAG-VDA FMEA Handbook (2019) recommends Action Priority (AP) levels instead, which weight Severity more heavily and prevent masking of high-severity, low-occurrence risks.
What is Action Priority (AP)?
Action Priority is the modern replacement for fixed RPN thresholds. It classifies each failure mode as High (H), Medium (M), or Low (L) based on the specific combination of S, O, and D values rather than their product. This prevents a high-severity failure from being hidden by low occurrence or detection scores.
How do I reduce a high RPN?
Focus on reducing Occurrence (prevent the cause with poka-yoke, design changes) or Detection (improve inspection, add SPC, automated testing). Severity is usually fixed by the design. After implementing actions, re-score to verify the RPN decreased.
Why is Detection scored in reverse?
Detection rates how well your controls can catch a failure before it reaches the customer. A score of 1 means almost certain detection (best), while 10 means no detection capability (worst). This is inverse because better detection reduces risk.