Using the ZandrEA project in research toward better HVAC system reliability in large commercial buildings

Published

July 28, 2026

Author(s)

Daniel Veronica, Amanda Pertzborn

Abstract

The ZandrEA collaborative open-source software development project is introduced by the U.S. National Institute of Standards and Technology (NIST) to better facilitate university and private-sector research directed at improving the reliability of the complex heating, ventilating, and air-conditioning (HVAC) systems in large commercial buildings. The ZandrEA project provides a versatile artificial intelligence software framework for automated fault detection and diagnostics (AFDD), combining benefits from both rules-based and process history-based (i.e., "data-driven") AFDD approaches. With ZandrEA, researchers avoid having to implement programming to distribute data sampled from a building's HVAC system to their experimental algorithms. They also avoid having to implement a custom, web browser-based, research-focused, graphical real-time display of an experimental algorithm's inputs and results. Those capabilities are provided upon instantiation of a ZandrEA application from its framework, freeing researchers to concentrate on the exploration of novel AFDD algorithms. Details of the ZandrEA framework are illustrated through an example of the kind of research it serves to support. The example shows that local control loops ubiquitous in HVAC systems have an inherent propensity to mask a specific fault from a purely rules-based AFDD approach. Yet, hinging a fault surveillance upon expert rules gives it a ready entryway for Bayesian diagnostic methods to help automate diagnosis of the root causes behind the suspected faults it detects. We show that the ZandrEA framework enables a fundamentally rules-based surveillance to be supplemented with automated classification, model-based prediction, or other data-driven machine learning methods. Similar examples would come from other types of HVAC faults where expert rules alone are not effective for detection or diagnosis. In summary, we show that the ZandrEA project software broadly facilitates collaborative research into essential data-driven AFDD methods, while preserving the inherent advantages of a fault surveillance based fundamentally upon expert rules.

Proceedings Title

9th International High-Performance Buildings Conference at Purdue, July 13-16, 2026

Conference Dates

July 13-16, 2026

Conference Location

West Lafayette, IN, US

Conference Title

2026 Herrick Conferences

Pub Type

Conferences

Download Paper

Local Download

Keywords

applied AI, automated fault detection and diagnostics, AFDD, expert system, FDD, HVAC reliability

Citation

Veronica, D. and Pertzborn, A. (2026), Using the ZandrEA project in research toward better HVAC system reliability in large commercial buildings, 9th International High-Performance Buildings Conference at Purdue, July 13-16, 2026, West Lafayette, IN, US, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=961994 (Accessed September 1, 2026)

Additional citation formats