Analysis of sensor offset characteristics in building energy systems based on redundant sensors: A case study on variable air volume system

Sensor offsets have a significantly negative impact on control optimization, operation and maintenance management, and energy consumption in building energy systems. Previous studies have employed various methods to detect, diagnose, and calibrate these sensor offsets. However, the definition of sensor offsets in these methods has largely relied on experiential knowledge, lacking a comprehensive analysis and discussion of the underlying reasons for sensor offset occurrence and its specific characteristics in terms of magnitude, form, and number. To address this issue, redundant sensors with the same or higher accuracy were installed adjacent to the original sensors in a variable air volume (VAV) system. The same data acquisition module was used to collect measurements from both the original and redundant sensors. Subsequently, the study meticulously analyzed sensor offsets under different operation conditions, including summer, winter, and transition seasons, as well as at distinct locations on the water and air sides, and various working stages, namely stable and dynamic stages. The results demonstrated that the sensor offset was influenced by the change of different operation conditions. However, the average offsets of most sensors were basically the same under different operation conditions. In contrast, the air side, where numerous sensors were deployed, exhibited a non -uniform airflow and rapidly changing measurements, consequently resulting in a larger offset (30.4 %) compared to the water side (9.26 %). Regarding the offset characteristics, temperature sensors were found to be a constant in offset form for stable or large fluctuation by dividing working stages, while differential pressure or flow rate sensors exhibited a normal distribution. Leveraging these analyses, a robust framework for defining sensor offsets was established, enabling the effective application of virtual sensing technology in practical building energy systems.