Efficient photodetector placement for daylight-responsive smart indoor lighting control systems

Abstract

With continuous increase in energy consumption worldwide, design of smart lighting system to reduce electricity consumption have been a pressing issue. This research focuses on design of smart indoor lighting system analyzing multiple zones with variable desired lighting conditions, to enhance user comfort and energy consumption. The accuracy of such system significantly benefits from large number of photodetectors, which increases the energy consumption and complexity. In this paper, a novel efficient control mechanism which reduces the number of photodetectors with no significant loss in accuracy is presented. A linear optimization method and a clustering algorithm are applied to select the least number and appropriate locations of photodetectors. A neural network is also employed to set the dimming level of luminaires. Furthermore, few auxiliary photodetectors are mounted and additional neural network is trained to find the relation between the output (measured illuminance) of these auxiliary photodetectors and that of zone-mount ones. In order to maintain user comfort, all zone-mount photodetectors are then removed. The added auxiliary photodetectors are used by the controller to analyze the daylight variation. The experimental results show that up to 82% of the maximum number of photodetectors can be removed with zero error in the calculated maintained illuminance. Additionally, the Mean Absolute Error (ERMeanAb) of providing the desired maintained illuminance for the entire control system is lower than 23.6lx, when there are no photodetectors on zone surfaces. This research provides a novel platform of designing a smart indoor lighting system with application to different working environments.