Glossary term
Validation
The process of confirming that a model, product, or system satisfies its intended use.
Definition
methodValidation is the process of confirming that a model, product, process, or system is suitable for its intended use in its intended context.
Validation answers whether the right thing has been built or modelled for the intended need. It differs from verification, which checks whether specified requirements have been met correctly. In engineering practice, validation connects user needs, operating context, acceptance criteria, test evidence, uncertainty, and residual risk.
Validation is evidence-based confirmation that an engineered result is fit for its intended purpose. A model can be mathematically correct and still not valid for a decision if it omits the relevant physics, operating range, user behavior, uncertainty, or failure modes. A product can pass individual specifications and still fail validation if it does not solve the real use case.
Validation requires a defined intended use. That includes users, environment, load cases, interfaces, operating limits, acceptance criteria, and consequences of failure. The evidence may include physical tests, simulation comparison, field trials, inspection data, usability studies, calibration records, reliability testing, or statistical analysis.
Engineering use
In product development, validation checks that the design meets customer and system needs. In simulation, validation compares model predictions with independent experimental or field data. In manufacturing or quality systems, validation may confirm that a process consistently produces acceptable output.
Validation should be planned before testing so that sample size, measurement uncertainty, boundary conditions, pass/fail rules, and data quality are adequate. Post-hoc interpretation of convenient tests rarely provides strong evidence.
Common mistakes
A common mistake is treating verification and validation as interchangeable. Verification asks whether requirements or model equations were implemented correctly; validation asks whether those requirements or equations are adequate for the intended use. Another is validating only nominal conditions while ignoring edge cases, degradation, misuse, and uncertainty. A strong validation review states intended use, acceptance criteria, evidence source, independence of test data, uncertainty, sample rationale, deviations, residual risk, and decision outcome.