Energy Performance Prediction Of Lighting Systems
David Caicedo, Philips Research
Ashish Pandharipande, Philips Research

We consider the problem of estimating lighting energy consumption in a lighting system upgrade scenario. In this scenario, a building has an existing lighting control system, denoted LCS". We are interested in estimating the energy consumption if a more energy efficient lighting control system, denoted LCS2, had been installed. Lighting data is available from LCS2 for a specific monitored area and from LCS" for the target upgrade area. A conventional method extrapolates the energy consumption from the monitoring area to the target area using surface area information. This has limited accuracy since differences in occupancy and daylight distribution across the two areas are not accounted for. To address this problem, we construct an energy model using support vector regression using lighting data from the monitored area. Lighting data from the target area is then used in the model to estimate the energy consumption. We show that the proposed method provides a better estimate of the energy consumption compared to the simple extrapolation method using data from an indoor office lighting system.