Fault Diagnosis and Root-Cause Analysis

Hold representative samples, records, photographs, instrument data and raw-material identifiers. Cleaning, reworking or discarding the entire batch can erase the evidence needed to distinguish causes.

Compare affected and unaffected product by raw material, operator, equipment, position, time, size and environmental exposure. Patterns often reveal whether the fault is local, systemic or batch-specific.

Five Whys, fishbone diagrams and fault trees organise hypotheses but do not prove them. Test the most plausible mechanisms against measurements, process history and controlled trials.

A root cause is the controllable mechanism whose removal prevents recurrence. Contributing conditions may increase likelihood without being sufficient alone, and several interacting causes can produce one fault.

Implement corrective action, define the expected response and verify recurrence is reduced. Document rejected hypotheses as well as the final conclusion so later investigations do not repeat unsupported assumptions.