Smart Lighting on the Smart Grid

User-Centric Model-Predictive Lighting Retrofit System : 2012-2013

Mechanical Engineering
Berkeley, CA 94720
Energy End-Use Category:
Lighting and Cooling

Research Partner: NASA Ames Sustainability Base

Project Overview:


Prior research in the Berkeley Energy and Sustainable Technologies Laboratory (BEST Lab) has demonstrated that retrofitting zonal lighting control systems with wirelessly actuated, user-controlled, individually dimmable luminaires in offices can save 50% of the office’s electricity usage; with additional daylight controls, it is theoretically possible to save 70% of the energy use. The current best-on-market technology was developed as a parallel effort to our research and reports similar energy savings. As of 2010, 70% of the U.S. stock of commercial buildings does not have intelligent lighting (including new buildings). Fifty percent of the intelligent lighting control systems have been deactivated by the users, and the remaining 50% operate at 50% of target. However, adoption and operational efficiency of smart lighting is adversely affected by high retrofit costs and lack of a closed-loop monitoring system.

Prototype/Technology Description

“User-centric model-predictive lighting retrofit system” is an innovative approach to address the current deficiencies in intelligent lighting retrofits with a comprehensive systemic approach that also adds novel capabilities to the retrofitted system. Our primary contribution is in the development of a new framework/method for retrofitting that includes a combination of new hardware and software components. The package will consist of a wireless illuminance sensor network, wireless luminary controllers, shades with built-in wireless shading controllers, and a server loaded with automated model calibration algorithms, inverse model generation, and predictive control and energy evaluation software with optimization capabilities. The energy and cost savings prediction algorithm will account for operational uncertainties. Users will initially input their lighting preferences manually and will undergo occasional, short satisfaction surveys to adjust dynamically the lighting preferences. The system will use the predictive inverse model to automatically optimize lighting schemes for energy efficiency while maintaining user preferences.

Potential Energy and/or Cost Savings

Our system will offer savings of more than 1.32 quads of commercial primary energy in the United States. Of the total average estimated energy savings, approximately 20% is from lighting and 10% from cooling with some extended energy savings, assuming that better control over heat gain will facilitate adoption of more energy-efficient cooling technologies in moderate climates in the country. A streamlined and less resource intensive commissioning and retrofitting process coupled with accurate model-based predictions of expected energy and cost savings could enhance the implementation of lighting retrofits in 21% of the target commercial buildings (assuming an increase in the commercial building stock by the time of product co

mmercialization and licensing). A further 15%-49% of indirect cooling energy savings is possible, depending on the climate zone, due to better adoption of thermally active systems. Cost savings are from a reduction of 50% or more in person-hours required for installation and commissioning.

Future Commercialization

The expected energy and cost savings should win recognition for our system in the current intelligent lighting market as a stand-alone solution or as an add-on to existing best-in-the-market intelligent lighting systems. This integrated solution to intelligent and energy-efficient lighting retrofit systems will work well with the most promising, emerging smart grid market technologies, as well.