University of California Energy Institute

Energy efficiency has recently come to the forefront of energy debates, especially in the state of California. This focus on efficiency has been driven by the deregulation of electrical-energy distribution, the increasing price of electricity, and the implementation of rolling blackouts. Currently, buildings consume over 1/3 of primary energy, and 2/3 of all electricity produced in the U.S. Commercial buildings consume roughly half of this, and lighting is responsible for approximately 40% of commercial building energy use. These numbers indicate that research in lighting efficiency has great potential to positively impact energy efficiency.

Efficient lighting controls proven to save up to 45% in electricity consumption are commercially available. In practice however, these systems are poorly received and greatly under-leveraged, resulting in a missed opportunity for impressive energy savings. Accordingly, we proposed the three-phase extension of an intelligent decision framework that addresses two major shortcomings of today’s energy-efficient lighting controls – user satisfaction, and lost energy savings stemming from naïve decision algorithms. The first phase of research was directed at enhancement of an existing preference-balancing control algorithm, in order that it accommodate demand-responsive control as well as the desires of the building manager. The second phase was devoted to identifying user preferences through empirical occupant testing. In the third phase, the resulting algorithms were simulated and evaluated.

Several facilities managers were interviewed and surveyed in order to identify appropriate variables and control policies to represent their desires within the decision algorithm. The preferred demand response strategy was found to be specific to the particular manager. Across all managers, energy was the most commonly selected indicator of the quality of lighting decisions. Automated occupant preference testing was conducted to demonstrate the feasibility of collecting such data in office environments, and to provide realistic occupant perceptions for use in simulation.

Simulated results indicate that the intelligent decision algorithm and framework present a promising control paradigm, and should be further expanded for the explicit inclusion of solar variables. Preliminary assessment showed that energy pricing can be factored into the control algorithm without significantly compromising occupant perceptions of lighting quality. Further energy savings are garnered by curtailing consumption during times of elevated pricing. Provided that curtailment is implemented with a slow enough dimming rate, reductions of up to 30% in illuminance are detected by roughly half of all occupants. Leveraging this research, the intelligent controller implements the specific demand response policy chosen by the facilities manager.