A proactive fault detection and diagnostics method for zone-level auto-commissioning in VAV AHUs

• Auto-commissioning identifies faults at early stages through active tests. • Proactive fault detection reveals hidden issues missed during normal operation. • Active testing shifts fault detection from passive observation to proactive action. • Diagnosis remains complex due to overlapping sensor, control, and hardware faults. • Automated detection is a foundational step toward future auto-correction. Fault detection and diagnostics (FDD) in HVAC systems is critical for maintaining optimal indoor environmental quality and ensuring energy efficiency. Undetected faults in HVAC components can lead to increased energy consumption, occupant discomfort, and higher maintenance costs. Traditional FDD requires symptoms to appear during the normal operation of a building, which may require waiting for a considerable amount of time until enough evidence is naturally collected for the correct diagnosis. To address this challenge, the paper presents an automated commissioning approach for detecting zone-level faults in variable air volume (VAV) systems served by an air handling unit (AHU). The proposed method consists of four on-demand automated tests designed to isolate specific faults, including airflow faults, heating faults, and cooling and heating mode control errors. Each test involves an actuation step, where a setpoint value is adjusted or a data point is overwritten in the building automation system (BAS), followed by a waiting period and a fault condition check. The method was applied to an academic office building in Ottawa, Canada, where it successfully identified five naturally occurring zone-level faults across different test categories in a total of 11 zones. These findings demonstrate the potential of this automated commissioning approach for early fault detection in HVAC systems. As a next step, this method will be expanded to detect system-level faults, while future work should also explore the development of automated fault correction sequences.