Field measurement-based validation of fault diagnostics for commercial building HVAC systems.

Number: 2339

Author(s) : EBRAHIMIFAKHAR A., CHEN Y., YUILL D. P., CROWE E.

Summary

An ongoing project is collecting data from multiple vendors of fault detection and diagnosis (FDD) software. We have a dataset of over 12 months from each FDD partner. The project is harmonizing the FDD outputs and aggregating these results for all of the vendors, to have a dataset on which analyses can be conducted to shed light on questions like: which faults are most commonly reported; how frequently each fault type is reported; whether faults correlate with drivers such as geographical region, building type, and others. In order to better understand the error of fault reporting by FDD, field validation has been carried out on a small subset of those buildings for which FDD outputs were gathered. This allows verification of the presence of flagged faults, and checking for faults that were not flagged by the FDD tools. More importantly, it provides insight into some of the challenges of using FDD data in a study of this kind. This paper describes the process of field verification, and preliminary results from the site visits. These site visits were conducted on two retail buildings that are each served by multiple rooftop units (RTUs). Examples of the types of faults that were encountered in the field include: stuck/disabled economizer dampers, disconnected or non-communicating sensors, non-condensable gas in the refrigerant, sensor drift or bias, and loose fan belts.
In the field study there were cases of faults being reported when they weren’t actually present, faults that were present that went unreported, and faults that were present, but are not a type of fault that the FDD tool attempts to diagnose. The sample size for the results in this paper – 49 RTU in two buildings monitored by one FDD tool – is not sufficient to be able to make any generalizations about the commonality of a given fault. However, the results show that the FDD results for some faults are less useful than others. For example, in one building all 27 CO2 sensors are classified as being faulted, but inspection showed that there are no CO2 sensors installed. Thus, these field results play into the broader objectives of the project. The work is ongoing, so this paper is intended to share early results with the buildings community and gather feedback.

Available documents

Format PDF

Pages: 10 p.

Available

Free

Details

  • Original title: Field measurement-based validation of fault diagnostics for commercial building HVAC systems.
  • Record ID : 30030665
  • Languages: English
  • Subject: Technology
  • Source: 2022 Purdue Conferences. 19th International Refrigeration and Air-Conditioning Conference at Purdue.
  • Publication date: 2022

Links


See other articles from the proceedings (224)
See the conference proceedings