Definition
In reliability and maintenance analysis, a function is described as what the equipment must do and at what performance. A functional failure is the state in which it stops doing so. It is a condition seen from the point of view of the user or the process: the pump no longer delivers the required flow, the sensor no longer measures within its accuracy, the motor no longer moves the load.
Difference from physical failure and failure mode
- Functional failure: the function is not fulfilled. Example: the pump does not maintain the necessary pressure.
- Failure mode: the specific way in which the loss occurs. Example: worn impeller, leaking seal, seized bearing.
- Cause: the origin of the failure mode. Example: cavitation, misalignment, poor lubrication.
The same functional failure can have several failure modes, and the same failure mode can produce different functional failures depending on the function considered.
Why the distinction matters
Equipment can be physically complete and still have failed functionally, for example an instrument whose drift takes it outside its permitted uncertainty. It can also have a broken part and not have failed functionally if there is redundancy or if performance is still sufficient. Defining the function first with measurable criteria avoids arguments over whether something is damaged or not.
Total and partial failures
The loss can be total, when the function disappears, or partial, when it degrades below the acceptable threshold. Partial failure is usually harder to detect, because the equipment appears to be operating.
An example
For a cooling fan, the function can be defined as moving an airflow that keeps the temperature below a certain limit. If the motor runs but the blade is dirty and the flow drops, there is a partial functional failure, even though nothing is broken. The failure mode is fouling, and the cause is lack of cleaning or an unsuitable filter.
How to use it in practice
When analyzing failure modes and effects, start from a list of functions with their criteria, list the ways in which each function can be lost, and only then look for failure modes and causes. This order avoids focusing on components and overlooking performance losses that the user does perceive.