Street Lighting Optimal Dimming with Model Predictive Control

Street lighting dimming is adjustable to current micro-location conditions such as weather, road and pavement traffic type and density. With the trade-off goals of energy savings and required lighting quality, it is suitable for optimization problem formulation. The paper proposes a model predictive control for optimal dimming of street lighting adjustable to micro-location conditions and multiple spatial points of interest. Simplified mathematical model of street lighting ray tracing is utilized to capture expected illuminance in various points of interest in a three-dimensional space. Power percentage of luminaires is optimized based on predicted micro-location data and with imposed gradual rate of change. A joint street-wise dimming profile is adjusted to several points of interest for each luminaire as a reference tracking problem for optimizing the light demand trade-off in critical points from safety aspect, user comfort from social aspect and minimizing the overall consumption. The algorithm is verified on the realistic annual simulation for a case study of Kralja Tomislava street in City of Sisak, Croatia. The results show the potential of 31.94% less consumption compared with the currently operating street lighting.