A fault is a condition that prevents the system from working as designed. A failure is the event that causes the system to stop working. The distinction matters for maintenance because faults can be detected and corrected before they become failures.
Faults. A thermocouple that reads 5°C low is a fault. The system still works. The parts may still pass inspection. But the process is not running at the designed setpoint. The fault is hidden. It does not trigger an alarm. It does not stop production. It slowly degrades quality. Other faults include a connector that is slightly loose, a shield that is partially grounded, or a probe that has lost contact pressure but still reads.
Failures. A thermocouple that reads open circuit is a failure. The controller alarms. The zone shuts down. Production stops. The failure is obvious. It demands immediate attention.
Why the distinction matters. Most maintenance resources go to failures. But failures are often the end result of faults that went undetected. A connector that was slightly loose for weeks eventually vibrates free. A probe that was reading low for months eventually fails completely. If the fault had been detected early, the failure could have been prevented.
How to detect faults. Faults are detected by measurement, not by alarms. Calibration checks detect drift. Insulation resistance tests detect moisture ingress. Temperature curve analysis detects slow response. Visual inspection detects loose connectors and worn springs. These activities are planned. They happen during scheduled maintenance, not during production.
The bottom line. The goal of a maintenance program is not to fix failures. It is to find faults before they become failures. A thermocouple that reads 2°C low is not a failure. But it is a fault. Find it. Fix it. The failure never happens.
