Intelligent Office Lighting Control Using Natural Light and a GA-BP Neural Network-Based System

Abstract

Intelligent lighting control systems are essential for regulating office illumination. Both illuminance levels and uniformity are important factors influencing the comfort of the office lighting environment. Thus, designing automatic control systems to regulate lighting is essential. This study addresses the issue of natural glare by proposing a method that uses a genetic algorithm (GA) to optimize a backpropagation (BP) neural network model. The model predicts the angle of window slats, with the Sun altitude and azimuth angles as inputs, and the slat angle as the output. For artificial lighting control, a linear function is proposed to manage the relationship between work plane illuminance, natural light intensity, occupancy rates, adjacent luminaire illuminance, and the dimming factor (K). The optimal K value for each luminaire is determined using the least squares method in MATLAB. The intelligent lighting system transmits dimming factors via a ZigBee tree network structure to achieve target illuminance levels. The system’s effectiveness is validated through simulations in DIAlux software, demonstrating that the workplace illuminance in occupied areas reaches 500 lx, while, in unoccupied areas, it reaches 300 lx, with an illuminance uniformity greater than 0.7. This addresses the issue of low illuminance uniformity during daytime. Additionally, the lighting power densities (LPDs) of 1.53 W/m2 and 3.8 W/m2 are well below the specified threshold of 6 W/m2, indicating significant energy savings while maintaining compliance with office lighting standards.

Keywords

• GA-BP
• feedback adjustment
• illumination uniformity
• lighting control
• simulation
• wireless sensor