An innovative fault analysis framework to enhance building operations

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
University of Alabama Libraries

This study proposes a failure analysis framework to enhance the building operations. A literature review was conducted for fault modeling and the fault mode and effect analysis (FMEA) applications to the building technology with a conclusion that few FMEA studies were applied in the building Heating, Ventilation, and Air-Conditioning (HVAC) systems. This study aims to fill this gap by investigating fault impacts through an FMEA analysis integrated with the whole building energy performance simulation (i.e., EnergyPlus). The primary objective of this study is to rank the fault based on impacts in terms of building energy consumptions and/or occupant thermal comfort under multiple faults. To achieve this goal, an extensible fault model library was established, including the building envelope, HVAC systems, lightings, etc. An FMEA framework was built to inject the possible fault models into EnergyPlus to evaluate their impacts on building energy consumption and thermal comfort. A parametric sensitivity analysis was used to determine and rank the criticality of the faults considering the fault concurrence frequency. Benefits and drawbacks of the response surface model through a deep learning algorithm (i.e. the multilayer perceptron regression) were explored during the sensitivity analysis for the fault rankings. The proposed fault analysis framework with rankings were tested and demonstrated for two DOE reference buildings (i.e., the medium office and the secondary school) in four different climate zones (i.e., Atlanta, Chicago, Miami, and San Francisco) with 24000 EnergyPlus fault simulations. Each fault mode is one fault model, or combination of multiple fault models, depending on the building energy systems. A total of 129 fault modes from 41 groups of fault models were implemented and simulated for the medium office case. The 129 fault modes are corresponding with 129 faults’ associated input parameters. For the secondary school, a total of 553 fault modes from 64 fault models were implemented during this study. The 553 fault modes are mapped into 553 fault associated input parameters. Those fault modes were injected into EnergyPlus simultaneously to study the fault impacts under multiple faults. The results demonstrate the proposed FMEA framework is robust and scalable for the fault impact analysis. The top critical faults for the medium office is the HVAC-Left-ON for the packaged rooftop unit, for the site energy, source energy, HVAC energy. Excluding the HVAC-Left-ON, the top critical faults vary greatly among the 4 climate zones. For the secondary school, the top critical faults are the Chiller-Fouling and Boiler-Fouling, for the air-cooled chiller system plus the gas-fired boiler.

Electronic Thesis or Dissertation
Mechanical engineering