Development of Anomaly Detectors for HVAC Systems using Machine Learning

Published: 28 November 2022| Version 1 | DOI: 10.17632/mjhr46dkj6.1
Davide Borda


The dataset refers to the operating parameters of an HVAC system that controls the environmental comfort of a non-residential building located in Turin (Italy). The dataset contains 11 variables: 1. timestamp; 2. temperatures of return, supply and outdoor air [°C]; 3. relative humidities of return, supply and outdoor air [%]; 4. the temperature setpoint of the return air [°C]; 5. the saturation temperature in the humidifier [°C]; 6. power required by the fans [kW]; 7. energy required by the fans [kWh]. The data refer to the winters 2019-2020 and 2020-2021. The data is acquired every 15 minutes.



Artificial Intelligence, Machine Learning, System Fault Detection, Building Management