Fault detection & diagnosis at Montréal-Trudeau Airport

Working with real operating data

The first case study used five-minute data from one air-handling unit serving four zones over summer 2025. The framework checked how outdoor and return air were mixed, whether dampers and cooling equipment responded as expected, and whether temperatures and airflow followed their control targets.

What the first checks found

Most mixing and cooling checks behaved as expected. The framework also flagged one zone where measured airflow stayed below its target and periods when discharge-air temperature did not follow the reference sequence. Those patterns identify where an operator should look next; they are not treated as confirmed faults until the actual control sequence and equipment are checked.

From detection to action

The value of fault detection is not another stream of alarms; it is a shorter, better-supported list of conditions for operators to investigate. The next phase will verify the flagged conditions, extend the checks across more equipment and seasons, and examine how a diagnosis can lead to an appropriate maintenance or corrective action.