Glossary · Maintenance and diagnostics
Fault detection
Also known as: Error detection
German: Fehlererkennung
In automation and dependability engineering, fault detection is the recognition that a fault or abnormal condition is present in a system, for example through plausibility checks, limit monitoring, test pulses, watchdogs or comparison of redundant channels.
- Maintenance
In one sentence
Fault detection is recognizing that a fault or abnormal condition is present, using checks such as limits, plausibility tests, watchdogs or redundancy.
Example
The controller detects a wire break because the 4–20 mA signal of a pressure transmitter drops below 3.6 mA.
How it applies
- Engineering: Detection methods are chosen per Failure mode: signal range checks, discrepancy monitoring between two channels, feedback monitoring of contactors, cyclic tests and communication watchdogs.
- Functional safety: In safety-related parts of a control system, the effectiveness of fault detection is quantified as Diagnostic coverage (DC), and the detection must be followed by a defined Fault reaction.
- Operation: A detected fault produces a message or alarm that tells the operator what is affected.
- Documentation: Each detectable fault should map to a message text, a cause and a remedy in the manual. The documentation team needs this list from engineering early, because it drives HMI texts and translations.
Fault detection vs. fault diagnosis
Fault detection answers "is something wrong?". Fault diagnosis answers "what is wrong, where and why?". Detection without diagnosis leaves the technician with a generic alarm.