Versatile AHU fault detection - design, field validation and practical application
This dataset, related to the eponymous paper in journal Energy and Buildings, contains data of 6 air handling units with manually indiced faults. The paper is needed, it containts deiails about the fault induction and more detailed specifications of the AHU. Dataset contains classified "non-fault data" as well as "faulty data", in order to allow validation of fault detection and diagnostic tools.
Steps to reproduce
The AHU data has been recorded by BMS. Names of the datapoints has been changed for clarity. The names consist of three tags (original metadata), as: 1) "place", such as "chiller_valve", "heat_recovery_bypass", "supply_air_outlet" 2) "quantity", such as "air_temperature", "signal_0-100", etc. 3) "type", such as "real_value" (sensor), "required_value" (setpoint) etc. The dataset also contains graphical explanation of the "place" tags, and schedule of the faults with basic description. The faults in the schedule corresponds with the fault classification in the AHU data record.