Data Mining Approach to Characterize the Operating Conditions of AHU for a Small Office Building in Different Climate Zones

The operating conditions inside the air handling unit (AHU) affect the performance of in-duct accessories, such as air purification devices. Proper design and sizing of in-duct accessories to meet performance targets require hour-by-hour investigation of the operating conditions of the AHU. However, there are technical barriers to designers of in-duct accessories to conduct such investigation. This paper presents a data mining approach with k-means clustering to create surrogate models to represent the hourly data. Hourly operating conditions of AHU for small office building in two different climate zones are clustered into multiple groups. The approach empowers AHU accessories designers to design and size their devices properly without conducting simulation.