Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
Studies have shown that the exposure to daylight can have substantial visual/non-visual benefits for building occupants. To optimise daylighting provision while maintaining a comfortable visual environment, daylighting control systems have been investigated by architectural design and academic research for a number of years. However, real-time regulation of daylighting in buildings requires transient daylighting simulation with high accuracy and, daylighting simulation, as performed to date, is significantly impacted by the employed sky luminance distribution models which only crudely reproduce the real sky characteristics due to oversimplification and limited luminance sampling inputs. In this paper, an embedded photometric device is proposed to combine high dynamic range (HDR) imaging based high-resolution sky luminance monitoring with quasi real-time on-board daylighting computing, composed of a low-cost image sensor and a field programmable gate array (FPGA) micro-processor. A deliberate calibration procedure of the whole imaging system, regarding its spectral response (spectral correction error f’1 = 8.89%), vignetting effect and signal response, was formulated and validated experimentally. The device was made to measure a wide luminance range (150 dB) including that of the direct solar disk, sky vault, and landscape simultaneously. Finally, experiments during predominant clear and overcast sky conditions were conducted respectively to assess its performance in daylighting simulation, both qualitatively and quantitatively. The experimental results demonstrated its quality in solar tracking as well as its capability to reduce daylighting simulation error to one seventh to one third of that of a common practice using the conventional Perez all-weather sky model for workplane illuminance calculation in office buildings.
Mihai Adrian Ionescu, Shokoofeh Sheibani
Jan Wienold, Geraldine Cai Ting Quek, Dong Hyun Kim
, ,