The metadata inference function classifies BAS labels into groups of the relevant sensor and actuator types, and then associates the labels to the corresponding AHU or zone.

A building automation system (BAS) in a large commercial building contains a large number of data points which can be extracted for performance analytics. Everything from airflow damper positions to fan speed, and from room air temperature to temperature setpoints, is captured within the BAS. BAS metadata is used to provide context to the type of data point, its relation to other points, and its taxonomy. As with any analysis, this toolkit depends upon this context in order for the functions to perform. However, metadata labels are often inconsistent, varying from building to building and from technicians and vendors. Though initiatives like Project Haystack and BrickSchema aim to establish standardized ontologies for BAS metadata labels, such initiatives are seldom implemented. Thus, a method to quickly classify and associate metadata labels if needed to facilitate performance analytics.

The metadata inference function classifies BAS metadata labels into groups of the relevant sensor and actuator types, and then associates the labels to the corresponding AHU or zone. The data point which are relevant for this toolkit are listed in Table 1 and make up the required AHU- and zone-level HVAC control network trend data for the AHU anomaly, zone anomaly, end-use disaggregation, and complaint analytics functions. Note that this function does not produce any KPIs or visuals intended to identify energy deficiencies or opportunities for energy optimization like the other functions in the toolkit. This function serves purely as a pre-processing tool to aid identify the relevant data points for AHU- and zone-level HVAC control network trend data.