A typical problem is when the same behavior may be due to various sources.
In the example of pressure decrease in a container, the decrease may be due to leakage from the container or from one of the valves, unexpected state of a valve, and more. This makes the troubleshooting procedure very tedious and too slow (as was the case in the PORV of the TMI accident).
To proceed with the recovery procedure, we need to find a model that matches the actual system behavior uniquely. The mapping from failure source to system behavior should be isomorphic.
Updated on 19 Apr 2016.