On certain streets in New York City these days, the past and the future appear in different colors. Among the many changes set in motion by former mayor Michael Bloomberg and transportation commissioner Janette Sadik-Khan, one was to swap out the city’s 250,000 streetlights from the current mix of mainly high-pressure sodium (HPS) lamps, with metal halide (MH) and mercury-vapor lamps in some areas, to light-emitting-diode (LED) lamps. Night owls and motorists are getting used to a streetscape where sepia-colored blocks are in retreat and blue-white light is on the march.
The LED revolution is spreading from commercial and residential uses to the municipal realm, as New York, Los Angeles, Portland, Seattle, Boston, Detroit, and other jurisdictions have begun implementing the recommendations of numerous organizations including the U.S. Department of Energy, the Climate Group, and Rensselaer Polytechnic Institute’s Lighting Research Center (LRC) in favor of LED conversion. If this technology continues to advance at its current trajectory, with costs dropping enough that long-range savings increasingly outweigh upfront investment, solid-state lighting (SSL) will logically replace HPS as the dominant form of streetlighting nationwide. Yet SSL’s properties are best suited to certain settings, says John D. Bullough, senior research scientist and adjunct assistant professor at the LRC. Municipalities under budget constraints should heed what the science says about LEDs’ optimal applications.
Calculations of luminous efficacy are only the beginning of a practical analysis of how different lamp types perform. HPS lamps produce approximately 100 to 120 lumens per watt, and contemporary LEDs are now competitive with HPS in that metric. “What’s more important is really getting light where it needs to go,” Bullough points out—illuminating roadways, streets, and sidewalks while avoiding or minimizing upward light pollution and light trespass onto adjacent properties, particularly residential windows. Illuminance measurements are meaningful only in their purposeful context.
The Illuminating Engineering Society (IES) has recently updated its recommendations for roadway lighting. Before this, there had been no major changes in standards for light levels and uniformity since the previous recommendations in 2000, which predated the LED era. For specifiers deterred by first costs, the implication might be “don’t fix what ain’t broke.” However, a detailed report prepared for the National Cooperative Highway Research Program’s Transportation Research Board of the National Academies, authored by Bullough and his LRC colleague Leora C. Radetsky in August 2013, makes the case for a new metric, Luminaire System Application Efficacy (LSAE), to guide engineers in selecting luminaires and their spacing and mounting height in ways that not only save energy and money but target LEDs’ unique properties to settings where priorities also include minimizing both light spill and distracting glare. (The LRC defines LSAE as “average luminous flux within a specific solid angle per unit of power, [including] only the light output that falls on the task plane … and meets the photometric requirements of the application’s task.”)
“We’re past the tipping point where there definitely are LEDs that are as good as or better than HPS,” Bullough says. “Unfortunately, what we can’t quite say yet is which ones for which type of roads.” LEDs may never sort out into the standardized wattage levels seen with HPS and MH lamps, making LSAE calculations complex. Still, Bullough envisions an interactive calculation tool, “under development in pieces” and probably available in the near future (perhaps integrated into lighting software), that analyzes luminaires’ features, power levels, and fixture positioning to recommend designs that optimize energy efficiency and costs.
Retrofit or New?
The first point to consider about LEDs, Bullough says, is whether they would replace HPS (or other non-solid-state) lamps in an existing fixture or constitute a new installation. In either setting, they offer advantages in energy cost, light quality, and maintenance, and last 50,000 to 100,000 hours (depending on manufacturer data) if designed with proper thermal control. LEDs’ energy-use reductions vary widely but boil down to at least a 15 percent advantage over HPS in the LRC’s 2013 literature review.
The benefits of LEDs are greater and more diverse, however, when designed from scratch than when retrofitted. With standard cobrahead fixtures, HPS or MH fixtures take a limited number of forms. “Their performance fundamentally is pretty similar in the way they take the light from the [lamp] and distribute it through a reflector or some kind of refractor, but pretty much every LED fixture is unique in the way it’s designed and how it spreads the light,” Bullough says. The tightest design straitjacket, conversely, is a project using incumbent multipurpose utility poles. In these, except for average light levels, “everything else gets tossed out the window, because … the poles are not designed at all for any kind of lighting,” he says. “They’re just designed for carrying utility cables.” Justifying an upgrade to LEDs is generally less clear cut in retrofits, he adds, because “there’s almost no way to devise a rule of thumb to say we can replace a 250W sodium cobrahead fixture with an LED that’s X number of watts.”
New infrastructure makes the decision easier. LEDs are more precise in their optical control, which is conducive to flexible, site-specific fixtures with full overhead cutoff and reduced glare. “That doesn’t mean that every fixture out there currently does that,” Bullough acknowledges. “There are some fixtures where you can see bare LEDs, and they can be pretty bright.” Another advantage in a new design involves mounting hardware: When pole spacing and height are variables rather than given conditions, system designers have more options to optimize coverage. Precision in design, of course, calls for prescience about purpose and power: In areas of the U.S. and the United Kingdom that have installed LEDs (Detroit, Baltimore, and Hastings), some citizens have complained that their operation has left crucial areas underlit. “LED lights may be okay for drivers,” complained Baltimore city councilman Robert Curran in a 2012 city council budget hearing. “But where do people get mugged? Not in the middle of the street. They get mugged on sidewalks.”
Such practices are correctable, not only by overcoming a default windshield perspective and making pedestrians’ well-being a high priority, but by also using the technology itself. LED controls’ dimming potential allows adaptation to changing conditions, such as circadian variation in pedestrian or vehicular activity, requiring different illuminance around the clock. On a longer timetable, systems can also compensate for the lumen depreciation that occurs over time, instead of overlighting at the outset. “In principle, you could have a lighting system that maintains a constant output by actually underdriving the LEDs at the beginning of their life and then, over time, slowly driving them a little bit higher,” he says.
LED streetlighting upgrades represent a substantial market, bringing impressive savings. In New York City alone, the PlaNYC sustainability program will spend $76 million between 2013 and 2017 on what will be the nation’s largest LED retrofit, Forbes reports. The city anticipates an annual savings of $6 million in energy and $8 million in maintenance, which amortizes the whole capital investment in six years (the lifespan of a current standard HPS lamp) and continues to reap savings as the LEDs last up to 20 years. In 2013, Los Angeles completed the conversion of 141,000 streetlights, the nation’s largest retrofit to date, aided by the Clinton Climate Initiative. The savings found in the first seven years will pay off a loan from the city’s Department of Water and Power, and then the $57 million project will save $7 million in electricity and $2.5 million in maintenance each year.
The gains aren’t limited to major cities: San Luis Obispo, Calif., cut energy consumption by 65 percent and service expenses by 60 percent by converting more than 2,000 HPS lamps to LEDs. A Department of Energy survey of about 240 municipal, county, state, and utility organizations released in September indicated plenty of room to expand: 82 percent of the respondents identified HPS as the most prominent lighting technology, with 36 percent mentioning some ongoing use of the even less efficient mercury-vapor lights; 62 percent reported some use of LEDs, but it was the most prominent technology in only 8 percent. Accounting for some 11 million individual luminaires, which averaged 15.3 years in age, this sample suggests that high-performance lighting offers the nation a rich option in the effort to improve energy efficiency.
Perception Affects Reality
Compared with the correlated color temperature of HPS (around 2100K), LEDs offer a higher and broader range (3000K to 8000K), and this whiter light includes more of the blue and green short-wavelength components of the visible spectrum. Because human eyes and light meters respond differently, analyses of effective lighting include subjective components. Comparisons of illuminance per watt also need to consider impressions of overall brightness along with on-axis and peripheral vision; it is in these areas that LEDs shine.
“When people look at a lighted street, their judgment about how safe that street feels and how secure they are actually seems to be directly correlated with how bright it appears,” Bullough says. The whiter light of LEDs gives observers a perception of greater brightness even when light meters indicate they are delivering fewer lumens than their HPS predecessors. “Obviously you’re going to meet the minimum requirements for the street and roadway area, but you wouldn’t necessarily have to increase the light levels just to get that sense of security,” he says.
A key distinction involves mesopic vision: the medium range between photopic or daytime light levels, when the retina’s cones are operating, and the darker scotopic range, when the eye’s rods are activated. Vision in the mesopic range, involving both rods and cones, involves a shift toward the short wavelengths. (Though IES recommendations do not yet consider mesopic photometry, application committees are exploring the question for a future iteration.) A separate metric, Bullough notes, distinguishes peripheral detection from on-axis detection: about 80 percent of the human visual cortex is devoted to the central 2 percent of the visual field, and visual consciousness focuses on details and colors in that area, while peripheral detection for most people is more a matter of general brightness recognition. From what is known about brightness perception (including the hazard of discomfort glare from overlighting) and about reactions to on-axis and off-axis visual stimuli, “for the same measured light level, a white light source like an LED will give you better peripheral detection at those mesopic light levels,” Bullough says. This carries potential implications for matching road types with conversion priorities, though Bullough notes that studies linking crash safety with light levels do not yet include enough data on LEDs to show “meaningful improvement” over HPS.
In urban areas, where pedestrians often outnumber drivers, commercial activities are often nocturnal, ambient light from buildings and signage is more prevalent, and vehicular speeds are usually lower, general brightness perception is a priority for pedestrians’ understanding of safety, but mesopic vision tends to fade away. The case for LEDs in cities rests more on energy and maintenance than on peripheral-detection benefits.
LEDs, sensors, and other advancing technologies may change the whole paradigm, one in which streetlighting is largely provided by private utilities under public regulation, suggesting that LED conversion may represent one early phase of a broader disruption in the streetlight sector. Current “misalignment between utilities’ revenue model and their customers’ own financial needs and service preferences” implies that the efficiency advantage that LEDs enjoy does call for new business models, preferably ones that don’t align utility incentives with high-volume sales of kilowatt-hours.
As desirable as it is to improve the performance of streetlights, Bullough cautions that what’s worth doing isn’t always worth overdoing. “Over the past 50 years,” he says, “as lighting technologies have improved, light pollution has gotten worse. The ability to produce light more cheaply and efficiently seems like a good thing, but then we just end up putting out more outdoor and streetlighting. There’s legitimate concern that LEDs are only going to accelerate that trend.” A key question that Bullough believes we need to ask is: “What is the light here for?” Matching capabilities to purpose—choosing smarter lighting, not simply more lighting—will not only ease the national energy footprint, it may also let more Americans have a look at the stars.