Ontologies¶

An ontology is a sort of specification or data model of how data about something should be constructed. The building automation industry has created many different ontologies, which have substantial overlap although also some relevant differences. Some you may have heard of include Brick, Haystack, ASHRAE 223P, Google's Digital Buildings Ontology (DBO), as well as several others.

Normal's perspective is that the ontology in use should be chose to be well-aligned with an application's data requirements; for instance, an application performing advanced diagnostics and optimization about air handlers may require very detailed modeling of different air handler components; but cursory (or no) coverage of other systems in a building. Therefore, it is difficult to choose a single ontology which works well for all use cases but is also not so complex as to be hard to understand and use correctly.

Out of the box, Normal ships with an ontology derived from the Haystack. Haystack is commonly used in the buildings industry, and provides flexible support for modeling equipment as a composition of features, which is a useful concept.

Furthermore, the ontology used by Normal is defined using a configuration file making it possible to customize for specific use cases. In particular, you may either extend or simplify all elements of the ontology, including

• Point definitions
• Equipment types
• Standard points lists
• Relationships
• Trending defaults

This allows a semi-custom ontology easily be deployed, guaranteeing consistency between individual sites and matching the structure of the data with the application.