The Language of Modern Aerospace.

Core meaning

In industrial and manufacturing contexts, a **defect** is any nonconformance in a product, material, or process output relative to defined requirements. These requirements are typically documented in specifications, drawings, standards, control plans, or work instructions.

A defect may involve:
– Dimensions or physical properties outside tolerance
– Missing, extra, or incorrect components or features
– Cosmetic or surface flaws beyond acceptable limits
– Functional failures during testing or inspection
– Incorrect labeling, documentation, or identification linked to the product

Defects are usually detected through inspections, in-process checks, automated vision systems, functional tests, or customer feedback.

Use in manufacturing workflows

In regulated and high-reliability manufacturing, the term “defect” commonly appears in:

– **Quality management systems (QMS):** Recorded as nonconformances or defects in logs, CAPA records, and deviation reports.
– **Manufacturing execution systems (MES):** Captured as scrap, rework, or defect codes at work centers, often tied to batches, lots, or serial numbers.
– **Statistical process control (SPC):** Counted for defect rates, DPMO/PPM, yield, and process capability analyses.
– **Traceability workflows:** Linked to specific equipment, operators, materials, and process parameters to enable root cause analysis.

Defects can lead to scrap, rework, hold/review status, or controlled release decisions, depending on severity and impact.

Types and classification

Organizations often classify defects to standardize reporting and analysis. Common distinctions include:

– **By severity**
– Critical: May pose safety risk or regulatory impact
– Major: Affects fit, function, or performance
– Minor: Limited to appearance or non-critical features

– **By disposition**
– Scrap: Cannot be economically or acceptably repaired
– Rework: Can be brought into conformance by additional processing
– Use-as-is / concession: May be accepted under controlled decision and documentation

– **By origin**
– Design-related: Arising from specifications or design choices
– Process-related: Caused by manufacturing methods, equipment, or setup
– Material-related: Due to incoming material or component issues

Boundaries and exclusions

In this context, a defect:

– **Is:** A specific instance of nonconformance affecting a unit, lot, or process output.
– **Is not:**
– A generic “issue” or “problem” unless it violates a defined requirement.
– A process risk or failure mode that has not occurred yet (those are typically addressed in FMEA or risk assessments).
– A regulatory term by itself; some regulations define more specific defect-related categories that must be handled under their own rules.

Defects are distinct from **process variation** in general; only the portion of variation that drives outputs outside defined limits is considered defective.

Common confusion and related terms

– **Defect vs. nonconformance:** In many quality systems, a defect is one type of nonconformance. Nonconformance can also refer to process, documentation, or system deviations that do not attach directly to a product unit.
– **Defect vs. deviation:** A deviation is often a documented, controlled departure from a requirement (planned or unplanned). A defect is an actual failure to meet a requirement in a product or output.
– **Defect vs. failure:** A failure is typically functional (the item does not perform as intended). A defect may be functional or cosmetic, as long as it violates defined acceptance criteria.

Site context: defects and data-driven improvement

In data-driven manufacturing and AI/analytics use cases, defects are treated as **labels** linked to production data. Systems often:

– Capture defect type, location, and timestamp in MES or QMS
– Associate defects with process parameters, equipment conditions, and material lots
– Use these labeled records to train models for predicting defect risk, identifying root causes, or optimizing process windows

The completeness, accuracy, and traceability of defect data significantly influence the usefulness of analytics and AI in reducing scrap and rework.