Satisfying peak demand can be very expensive for utilities, which pass this cost onto their customers in the form of time-variable pricing or demand charges. Utilities, therefore, share a common interest with their customers to reduce peak demand. This is because shaving the peak enables them to satisfy customer demand while avoiding the high cost of building new capacity or having to buy expensive power from other markets during an emergency or demand spike. Besides charging more for power used during peak demand periods, a number of utilities offer financial incentives to building owners to curtail load on request, usually during an emergency grid event.

To reduce peak demand, we can turn equipment off, turn it down or use it more efficiently. Lighting, while being a major energy consumer in a typical commercial building, has a small role to play—at first glance. While we can use it more efficiently, it is difficult to turn off in routinely occupied spaces for obvious reasons and cannot be turned down in many spaces without installing dimmable ballasts.

But suppose we do just that: replace every fluorescent ballast in a commercial building with a dimmable ballast.

A number of solutions are available from manufacturers. They include ballasts that communicate through low-voltage wiring and, therefore, can be integrated into a lighting control system that leverages dimming for other energy--saving purposes. They also include line-voltage ballasts, which are ideal for retrofit and designed to provide continuous or stepped dimming.

How low could we go with lighting levels during peak demand events? Guy Newsham, Ph.D., senior research officer at the National Research Council Canada’s Institute for Research in Construction (NRC-IRC), led a group of researchers to study this question that is at the heart of dimming’s potential as a demand-response strategy.

“Our aim was to explore how far and how fast one could dim smoothly from normal levels and over what period, without incurring undue hardship on occupants,” he said.

NRC-IRC conducted two laboratory studies in full-scale office mock-ups where various dimming scenarios were studied with typical office workers conducting office tasks. The researchers then took the results and used them to design a field study, conducted during the summer. The study included an open-plan office with 330 dimmable lighting fixtures and a college campus with 1,850 dimmable fixtures in several buildings. Load shedding was performed during afternoon hours over several days. The rate of dimming spanned 1–30 minutes with dimming reductions up to 40 percent. Occupants were warned that an experiment would be conducted over the summer involving afternoon dimming but were not told on which days.

“In the field study, we attained load sheds of 14 to 23 percent of normal lighting load with no complaints from occupants,” Newsham said. “In combination with the lab study data and findings of related work by others, we derived several recommendations. The first stage of demand response should be dimming by amounts that the majority of occupants don’t even notice. The second stage of demand response, when more load reduction is required, may involve dimming to light levels that are noticeably lower but are still acceptable to the large majority of occupants.”

In stage 1, dimming can occur rapidly, over as little as 10 seconds, by 20 percent with no daylight, 40 percent with low prevailing daylight, and 60 percent with high prevailing daylight. If dimming occurs slowly, during 30 minutes or more, and with no immediate expectation of dimming occurring, levels may drop by 30 percent with no daylight and 60 percent with high prevailing daylight.

In stage 2, dimming can occur rapidly, over as little as 10 seconds, by 40 percent with no low daylight and 80 percent with high prevailing daylight. If dimming occurs slowly, during 30 minutes or more, and with no immediate expectation of dimming occurring, levels may drop by 50 percent with no daylight and 80 percent with high prevailing daylight.

“Nevertheless, it is imperative to recognize that demand-response dimming should only be enacted to alleviate temporary grid stress problems that occur infrequently and is intended to prevail for a few hours at most, and light levels should be returned to normal levels thereafter,” Newsham said. “Our studies do not provide support for these lower light levels becoming the ‘new normal.’”

The challenge is making the capital investment in dimmable ballasts worthwhile during infrequent reductions in load, particularly if daily load shedding is off the table. Another challenge is liability: Managers may feel at risk if they dim for demand-response purposes and somebody trips and falls and blames low lighting levels. This may require some leniency in IES-recommended lighting practices to accommodate certain temporary deviations.

However, it is highly likely that dimmable lighting will be a component of demand-responsive commercial buildings in the future. According to the Department of Energy, about 281 gigawatts of new generating capacity will be needed by 2025 to satisfy growing demand for energy, much of which will be allocated solely to satisfy peak demand.

DILOUIE, a lighting industry journalist, analyst and marketing consultant, is principal of ZING Communications. He can be reached at