A Data-driven Fault Detection and Diagnostics Impact Analysis Framework for Enhancing Building Indoor Air Quality

 Abnormal operation of HVAC systems can lead to higher energy consumption, poor indoor air  quality, thermal discomfort, and reduced productivity. However, current HVAC fault detection  & diagnostics (FDD) metrics have been primarily developed for energy savings, with little  consideration of the impacts on indoor air quality (IAQ). This thesis proposes a data-driven fault  impact analysis framework for enhancing building IAQ. After a systematic review of fault rules  from three leading vendors, a curated list of faults that influence IAQ are selected. Based on two  years of time series BAS and FDD data for a federal office building, Machine Learning tools  help to quantify the relationship between building operation faults and indoor carbon dioxide  (CO2) levels, to help generate FDD priorities for IAQ.