Control of an Adaptive Light Shelf Using Multi-Objective Optimization

Users are increasingly expecting intelligent behavior from automated building systems. However, incorporating and integrating user preferences in an automatic control algorithm is a complex task. In most commercial systems, control is done through local search techniques such as Proportional Integral Derivative (PID) control. Multiple criteria cannot be accommodated in such techniques. The objective of this paper is to evaluate the use of multi-criteria optimization in the control of a lighting and shading device, an adaptive light shelf. This study is done in two stages. In the first stage, the performance of the light shelf is evaluated using simulations. EnergyPus and Radiance are used to simulate thermal load and daylighting. These are used to determine the optimal control actions using a multi-objective optimization algorithm. The energy savings for electrical lights is compared with a traditional control strategy. In the second stage, a prototype is constructed to evaluate the actual performance. Results from simulation show that significant energy savings can be achieved through multi-objective optimization control strategy.