Straightforward troubleshooting of a network of connected units requires testing all the units, until the faulty one is identified. This
procedure is time consuming, and might be inefficient, especially in emergency operation. After a fault was detected, the operators need to identify the source, namely, the particular component that failed.
Straightforward troubleshooting of a network of connected units requires testing all the units, until the faulty one is identified. This
procedure is time consuming, and might be inefficient, especially in emergency operation.
The problem
is that sometimes the
operators cannot identify the failed component based on
the attributes of the
alarm, because they were not trained to recognize
components by associated
alarms, or because they did not experience the
alarm
often enough to remember the particular
system behavior. This is especially important to enable safe behavior
under stress.
The pilots at the AF 447 accidents were not trained to recognize the exceptional maneuver by the captain in charge.
The TMI team in the control room was not trained to resolve conflicting information obtained when the PORV was stuck open.
Troubleshooting may be facilitated by segmentation, as demonstrated here ....
Operators should be trained to recognize the
alarm
situation and to identify the source of the
alarms.
To enable training the
operation in exceptional condition, the
system should be equipped with mean to bypass the sensors, and to simulate their behavior in exceptional conditions.
Updated on 11 Feb 2017.