자연채광 모델의 불확실성을 고려한 실시간 전기조명 제어

다국어 초록 (Multilingual Abstract)

It has been widely acknowledged that daylit illuminance at work plane is of strong stochastic nature and thus hard to accurately predict. In this study, the authors present a daylighting prediction model that can account for the aforementioned stochastic nature using Gaussian process model. The model can quantify uncertainty in daylit illuminance at 16 work planes of a given target building. The daylighting prediction model is then integrated to automatic electric lighting control that can switch on/off six lighting groups in the building. It was found that the integrated control could save lighting energy consumption by 15.3% over four days (Aug 9 – Aug 12).