Artificial versus natural light source identification with light intensity sensors for energy monitoring

摘要

Studies have shown that increasing the granularity of energy consumption information in buildings could facilitate achieving energy saving objectives. Lighting systems in buildings are one of the main consumers of electricity with a share of approximately (direct and indirect) 23 percent of the total electricity in the US. Direct monitoring of lighting systems is not feasible due to challenges for instrumentation. In this study, we are proposing an approach for monitoring of lighting systems energy consumption by using single light intensity sensors. Load monitoring through light sensors is challenging due to the contribution of natural light in variation of the signal magnitude in the captured time series. Accordingly, we have proposed a new feature in frequency domain that helps us identify artificial light (AL) source type and improve AL event detection. Field experimental study showed the effectiveness of the proposed feature by eliminating all false positives in event detection.