Prolonged refrigeration door open.

Trigger, action plan, and escalation criteria.

Related metric

refrigeration_door_exposure

Trigger

door_open_duration > configured_open_ceiling, outside approved_restocking_window (with door-sensor health confirmed)

Value drivers

Energy savings Product loss prevention Sustainability

Preconditions

Asset is door-equipped with valid door telemetry; prolonged-open ceiling and approved restocking windows configured per asset type and operating context; implausible-duration screen enabled so suspected sensor faults are held out until validated.

Human role

Primary for store communication and escalation when the door stays open; secondary review after AI on sensor validation and context check.

Action plan

(1) Validate the door-sensor health — screen implausibly long open events as possible sensor faults and exclude them from operational alarms until validated; never read a binary closed signal as proof the gasket is sealing. (2) Check the restocking window and operating context — a burst of openings or a longer open event during an approved window is expected activity, not an anomaly. (3) Correlate with temperature movement — is the asset's temperature beginning to move toward its band edge as the door stays open? (4) Notify the store so the door can be closed when the open event is genuine and outside an approved window. (5) Escalate if the door remains open after notification. (6) Recommend a product check on the affected category if temperature risk develops; record the outcome. (7) Run recurring-event analysis on the asset — repeated prolonged-open events point to a restocking-process or door-hardware issue; capture closure evidence.

Escalation

(1) Door open well past its ceiling with temperature already moving toward the band edge → out-of-cycle store notification with a product-check recommendation. (2) Door remains open after the first store notification → escalate to the site's operational contact. (3) A door-sensor fault pattern across many assets makes door exposure unreliable → pause door alarms for the affected assets until the sensors are validated. (4) Recurring prolonged-open events on the same asset after a hardware repair → flag the door or its sensor for a deeper look at the MBR.

Prevention

(1) Configure the prolonged-open ceiling and the approved restocking windows per asset type and operating context so normal activity is not flagged and real anomalies are. (2) Keep the door-equipped asset list and the implausible-duration screen current so doorless assets are excluded and suspected sensor faults are held out until validated. (3) Review per-asset open-time exposure at cadence so a slow erosion of door discipline is caught before it drives excursions.