Share

Superheat prediction & fault diagnostics of HVAC from simple temperature measurements using big data approach.

Number: 3683

Author(s) : TESFAY M. K., RAFAIE M., SINKAR K., ARUNASALAM P., BESARLA D., ALSALEEM F.

Summary

New advancements in data & algorithms have pushed new techniques and methods to the forefront in optimizing energy efficiency as well as keeping the thermal comfort of residents in intelligent buildings research. HVAC elements, being ubiquitous and fundamental elements in buildings today, their diagnostics maintenance, operational functionalities, and control are essential aspects in this regard. The tremendous amount of data generated from buildings every day and recent developments with data science tools have changed the control and monitoring of these units from exhausting physical modeling and operation to data-driven techniques that are more reliable and efficient. The massive streaming data generated by smart building sensors have inspired new ways of controlling and diagnosing faults in comfort systems using machine learning and big data analytics. In this work, we present a big-data driven approach to model the dynamic of two similar HVAC (but, healthy and faulty) systems from simple temperature measurements collected over an extended period. The model showed good accuracy in predicting the system superheat for both systems. This demonstrates the potential of big data approach to substitute the need for having the expensive pressure sensor to measure superheat.

Available documents

Format PDF

Pages: 10

Available

  • Public price

    20 €

  • Member price*

    15 €

* Best rate depending on membership category (see the detailed benefits of individual and corporate memberships).

Details

  • Original title: Superheat prediction & fault diagnostics of HVAC from simple temperature measurements using big data approach.
  • Record ID : 30028670
  • Languages: English
  • Subject: Technology
  • Source: 2021 Purdue Conferences. 6th International High Performance Buildings Conference at Purdue.
  • Publication date: 2021/05/24
  • Document available for consultation in the library of the IIR headquarters only.

Links


See other articles from the proceedings (52)
See the conference proceedings

Indexing