Knowledge overview

Reliability assurance

Assurance connects risk, evidence and release decisions.

Reliability assurance provides a traceable technical justification that requirements are met under relevant conditions.

Component failure probabilities are combined through a series system to obtain the system failure probability

Assurance starts in development

Reliability cannot be tested into a finished product. Concept and design decisions offer the greatest leverage to avoid or control failure mechanisms.

Qualitative methods identify critical paths; quantitative methods assess lifetime, failure probability or availability. A robust strategy connects both.

Match methods to mechanics, electronics and software

Mechanical fatigue, electronic ageing and software-related failure patterns require different evidence. Complex systems also demand attention to interfaces and common causes.

The method must fit the product and the failure physics rather than follow a generic checklist.

Test, simulate and document

Stochastic and numerical simulation can complement physical testing when models and inputs are sufficiently validated.

Evidence documentation records targets, methods, data, assumptions, uncertainty and results so that decisions remain traceable.

The decision is made at the permitted failure fraction q: the upper one-sided confidence bound of the estimated failure probability yields the conservative life Bq,L at that level.

Two Weibull curve plots for the Bq-level demonstration decision: Bq,L lies to the right of Bq,req in the demonstrated example and to the left in the example that is not demonstrated.
The requirement is demonstrated when Bq,L ≥ Bq,req. The light-blue one-sided confidence region extends from F(t) = 0 to the upper confidence bound Fᵤ(t) across the full time range; the dark-blue curve shows the field behaviour estimated from data.

Related knowledge

Continue the technical perspective

Risk and assurance services