May, 2026
Journal
Field-based AFDD for refrigerant undercharge in residential HVAC systems: enhancing reliability through false alarm mitigation
This study evaluated rule-based and machine learning (ML) based automated fault detection and diagnostics (AFDD) algorithms for detecting refrigerant undercharge faults in residential heating, ventilation, and air conditioning (HVAC) systems, using actual building data and a minimal set of features. The ML-based algorithms included Decision Tree (DT) and K-Nearest Neighbor…