History of Lighting – Part 1

Published on LA Confidential, Winter 2014

Mel Brooks’ 2000-year old man used torches in his cave, but today’s lighting is a tad more sophisticated, having gone through half a dozen stages, each producing more light out of less energy i.e., efficacy, than its predecessor.

Torches made of moss and animal fat, and crude oil lamps were the mainstay for indoor lighting until the processing of animal fat into wax gave us the candle in roughly 3000 BCE (efficacy: about .1 lumens/watt). That option worked for much of early recorded European history but, during the Islamic Golden Age (roughly 900-1000 AD), a Persian doctor refined kerosene from crude oil and used it in the first manufactured oil lamp and the first street lights in Andalusian Cordoba, now Spain.

It wasn’t until the late 1700’s that European oil lamps (at ~.3 lumens/watt) became widely available and accepted due to improvements in design and the whaling industry’s ability to produce sperm oil. That refined product burned cleanly, didn’t smell too bad, and was relatively cheap compared to commercially-made candles.

The Industrial Revolution in 19th Century Europe gave us dynamo-based electricity and the first carbon arc lamps (at ~2-4 lumen/watt). They were used mostly in open areas such as parks, street lighting, and large industrial spaces and rail yards. Arc lights didn’t get much traction in the US until much later, due to the lack of an electric distribution system. During that century, gas lighting and kerosene lamps dominated.

In the early 1800s, gas light (initially at less than 1 lumen/watt), used coal or natural gas from mines or wells, and was relatively common in urban England. Its use expanded rapidly after the development of the incandescent gas mantle around 1890. That device more than doubled the efficacy (to 2 lumens/watt) of gas lighting, using a filament containing thorium and cerium, which converted more of the gas flame’s heat into white light. Gas lighting then cost about a quarter of candles and less than kerosene. It created a great demand that eventually led to the creation of the first utilities. Gas light dominated for several decades (through World War I, for those having access to it), until electricity became available, gradually supplanting both gas and kerosene lighting.

As the 20th century approached, Edison created his electric incandescent lamp (1.4 lumens/watt), and the power industry adopted it as a standard. During this period, Nikola Tesla, and others,    demonstrated forms of fluorescent lighting where gases in a tube glowed when charged by electricity. Efficacy began to exceed 20 lumens/watt. It wasn’t until 1926, however, that the first fluorescent lamp received a patent and not until after World War II when commercial fluorescents (~60-70 lumens/watt) began to supplant incandescent.

With the rapid development of technologies in the postwar era, other forms of lighting were developed, producing the first diodes emitting visible red light, the Light Emitting Diode (LED) in 1962. Mercury vapor and metal halides came to market in the late ‘60s (at 50-100 lumens/watt), and sodium vapor lamps in the ‘70s, with efficacies up to 180. Although more efficient, several of these light sources, especially sodium, provided light that distorted colors and was  far from white.

Compact fluorescents with built-in ballasts came to market in 1981 (50 lumens/watt), the first white high-pressure sodium lamp (the Philips white SON) in 1985, and the first commercially available electronic ballast appeared in 1987. The early ‘90s saw an explosion of new and ever more efficient light sources: electrode-less fluorescent aka induction lighting in 1991; the first ceramic metal halide (CMH) in 1992; and, in 1994, the first T5 fluorescents and sulfur-based lamps. Efficacies for white sources were now routinely exceeding 70 and approaching 100.

But the lighting revolution was just getting started. After almost 30 years in development, the first phosphor-based blue (and later white) LEDs were seen in 1995. As cost dropped, and efficacy rapidly improved, they began to supplant sources that had ruled some types of fixtures for decades. In only a few years, tritium lighting which used a slightly radioactive glowing gas, common in exit signs, essentially bit the dust even though it used no power for illumination.

In the 21st century, competition among sources continues. Fluorescent lamps, powered by electronic ballasts operating in computer-designed fixtures, are producing greater efficiency and longevity, giving LEDs a run for their money. High-output fluorescents are pushing metal halide aside, just as induction lighting competes with LED for street lighting. Efficacies near 100 are now considered the norm with LED having a theoretical potential exceeding 200.

The biggest loser in all this competition is the incandescent lamp, which ruled downlights and table lamps for almost a century. Screw-in LED units, while still relatively expensive, are quickly supplanting Edison’s venerable invention. Through international competition, government regulation, and various financial incentives, incandescence is fast becoming the “whale oil” of our time. During its heyday, however, many billions of Edison’s bulbs were produced, and used by billions of people, making it the king of indoor lighting for decades. To paraphrase King Louis XVI in the History of the World – Part 1, “It was good to be the King”.

Click here for full LA Confidential Winter 2014


Don’t Lose Your “Head” Over Line Loss

Published on LA Confidential, Winter 2013

During the process of delivering electricity, some power is lost through wire resistance, transformers, and other physical causes. Losses may also occur due to theft of service and meter error. Such line losses occur at two levels: while transmitting power at high voltage to a utility’s zonal boundary, and again when distributing it at lower voltage across a utility’s territory to customer meters. Transmission losses to a zone are typically included in a retail zonal price. On the utility side of that boundary, losses are calculated as the difference between the total kWh received at the boundary (plus whatever the utility generates inside its territory), and the total received at all customer meters, is expressed as a percentage. The generic term for that latter difference is Unaccounted For Energy (UFE), though many know it simply as ‘line loss’. Utilities charge power suppliers for the line loss by requiring that each supplier provide that much extra power to the utility boundary based on the losses seen in the prior day (or other time period). Any after-the-fact adjustments are then made up in the following time period.

UFE is built into tariff supply rates charged by utilities and in fixed rate commodity prices from third party suppliers. Under indexed pricing, however, one’s power price floats based on the hourly wholesale market price (the “index”). A supplier then adds a fixed dollar per kWh adder that may include capacity, ancillary services (provided by the grid operator), and the supplier’s management fees. Line loss may then be seen as an extra line item charge, or added to the index price.

Up to mid-2009, Con Edison provided monthly line loss factors for each of its 3 zones (H, I, and J). It then began publishing zonal hourly line loss factors through its ESCo News bulletins, typically three months after the fact (find them at http://www.coned.com/escos/news/index.asp.) Those hourly factors, however, are initially based on standard system-wide load profiles. They are then corrected in the next day’s (or month’s) line loss percentage for the same hour. Such corrections may occasionally result in negative line loss.

Several Ways To Charge For It

Most suppliers use a fixed line loss rate for the term of a contract, but that number varies among suppliers. In recent indexed bids, some used Con Edison’s standard average of 7.9%, but others offered factors as low as 5.5%. Based on discussions with the latter, they developed their loss numbers in roughly the same way they generate fixed priced bids. They calculate annually weighted averages based on a customer’s monthly usage. It remains unclear why, using the same data, one offered 6.3% while another offered 5.5%.

Some ESCos use more complex methods involving floating factors that cannot be compared in advance, such as:

– Con Edison’s posted average monthly line loss numbers

– Creating a rolling 12-month average weighted rate based on a customer’s particular on and off-peak usages

– Con Edison’s line loss numbers multiplied by a customer’s hourly kWh usage; this can be especially tricky because Con Edison’s hourly numbers – during a given month – have ranged from +26.1% to -12.8%. How can a line loss be negative, you ask? Because each hour’s loss is initially based on the standard load profile for each rate class, and then corrected in the following hour based on real data.

Several Ways To Invoice For It

In a supplier’s bill, line loss may appear by:

– bumping up the customer’s hourly kWh consumption by the line loss percentage, yielding different kWh from the supplier and that delivered by the utility

– bumping up the hourly pricing by the line loss factor; it then won’t match that posted by the grid operator

– a separate line item that sums the costs into one monthly dollar number.

Is One Way Better Than Another?

To answer that question, both dollar impact and price risk need to be considered.

To see the difference between two fixed line loss offerings, let’s assume an average floating supply rate of $0.08 per kWh wherein capacity is responsible (on average) for $0.01 per kWh of that total. That leaves $0.07 per kWh to be impacted by line loss. If supplier A charges an annual fixed 7.9% and supplier B charges 5.5%, that’s a difference of about 1.7 mills (i.e., tenth of a cent) per kWh. That could be the difference between the winner and the next closest bidder, and may be more than the fee paid to the broker or consultant handling the deal. For a large customer (e.g., 35,000 MWhr per year), that number translates to $61,000 annually. (About one full-time staff salary for a typical customer.)

In an indexed bid, the supplier offering a lower fixed adder and a lower fixed line loss rate should be the winner. But if the line loss rate is not the lowest, and the fixed adder is, some number juggling is needed to see which is offering the lower total rate. But what do we do with the more sophisticated UFE calculation methods?

The weighted rolling 12-month average introduces additional price uncertainty because the line loss is now floating and can’t be compared to the fixed rate offerings. How consistently a facility uses energy month-to-month may now become important, e.g., an industrial with highly variable (or seasonal) production, or those with wide year-to-year variations. Accurately forecasting a comparable annual rate may not be possible.

Trying to project the impact of an hourly varying line loss factor, however, requires taking into account a customer’s hourly kWh usage (from a prior year), and may thus carry even more price risk. Because Con Edison’s hourly line loss numbers flop across such wide ranges, a facility’s hourly load profile – and its year-to-year consistency – become crucial to calculating this method’s net impact. But there’s an additional wrinkle: Con Edison’s hourly line loss numbers are published 3 months after the fact. What happens in the first 3 months of a contract, before the corrected line loss numbers are available? Is it even contractually legal to use data from a time period that precedes the start date of a contract by 3 months? The supplier answered that. “If accepted by the client, then it’s contractually OK.” His contract, however, contained no description of his line loss process, and it took several emails to elicit that information.

Click here for full LA Confidential Winter 2013.

Solar PV Cuts Peak Demand Charges

Published on LA Confidential, Winter 2013

Solar power from photovoltaic (PV) panels, when secured through long-term contracts, has become competitive with power supplied by utilities. Under a Power Purchase Agreement (PPA), a developer designs, installs, owns, maintains and operates the PV system at a customer’s site, and sells the power generated by the system to the customer at a competitive rate. Through such arrangements, many facilities are now securing years of lower cost and pollution-free electricity.

And the incentives to cover part of the installation cost just keep getting better. NYSERDA recently announced its NY-Sun Initiative (http://ny-sun.ny.gov/) to expand PV installations at NY customers’ sites.

The initial rate that the developer charges the customer depends on a variety of factors, including the installed cost of the system; any rebates, incentives, or renewable energy certificates associated with the system; and the developer’s required return on investment.  PPA’s are especially attractive to non-profit institutions who cannot otherwise take advantage of tax incentives. The initial rate per kWh can be up to 15% lower than the utility rate, or it can be at or even slightly higher than the utility rate.  In the latter case, the system can still be economically beneficial to the customer if the annual PV rate escalation is less than the projected utility rate escalation over the 15 to 20 year term of the PPA.

But care is needed before signing on to a flat electric rate for PV power, especially over a contract that may run 15 to 20 years. While a flat price may be fine for a residential customer whose utility rate is based solely on his kWh consumption, many commercial rates include a separate charge for monthly peak kilowatt demand (kW). In such cases, a flat PV rate may assume that both kWh and kW will be saved proportionately, but some analyses have challenged that assumption.

A typical flat panel PV system orients the units to maximize kWh production. With such fixed positioning, output varies as the sun moves across the horizon, being greatest when the sun’s rays are perpendicular to a panel at noon. But earlier or later in the day, output drops off sharply as those rays strike the panel at shallower angles. To reduce a building’s peak demand by the kW capacity of the panel, the building’s load would therefore need to peak close to noon. If the load peaked much earlier or later in the day, the PV impact on peak kW could be marginal, or even zero.

In such cases, the dollar value of the PV power may be closer to the marginal daytime kWh price – without peak demand – rather than the overall average kWh rate, which has the full cost of monthly demand built into it.

How a customer’s tariff defines peak demand may also impact demand savings. Many tariffs calculate a customer’s monthly peak demand based on the highest peak seen during a specified weekday period (e.g., 8 am to 6 pm) even when demand may be significantly lower the rest of the month. Under such tariffs, even if a building’s load peaked at noon, a month’s potential demand savings may be reduced if – only once in that billing period – the sky at that time is heavily overcast or if it’s raining, either of which may noticeably reduce hourly PV output.

To properly determine the value of a flat PV electric rate thus depends on understanding a building’s load profile and its electric tariff.

In a recent analysis of an industrial facility considering installation of a large PV system, hourly interval data was used to diagram how that system would affect its load profile. We tracked demand usage versus power generation from the PV. We noted peak demands occurred in the morning between 8 AM and 10 AM. The load drops slightly in the middle of the day and afternoon. We stimulated the impact of PV. The results showed that midday loads would have been suppressed by the PV system’s output but the morning peak demands, while shortened, were essentially unchanged in magnitude. All days were assumed to be bright and sunny, with none overcast or raining.

In this case, the PV system would have generated millions of kWh savings, but its peak demand charges would have remained the same. The PV developer wanted to charge a flat $.15/kWh (plus an annual escalation) for his electricity for 20 years. That $0.15 was promoted as a 12% discount off the customer’s present $0.17/kWh average utility rate – which included peak demand charges. When that rate was calculated without any demand charges, it averaged about $0.10/kWh. In light of that difference, the developer was asked to revise his proposal, but refused to do so.

Another study found similar results with other types of buildings, due mainly to the non-coincidence of PV output and time of peak demand. Sampled apartment buildings, for example, peaked most often after 5:30 pm, when PV output was quite low. In all cases, electric bills would have shown significant reductions in kWh but, to a lesser degree, reductions in actual cost.

To avoid surprises, those considering large PV installations should have a competent consultant review their annual hourly load profile and their utility tariff before signing a long-term contract.

Click here for full LA Confidential Winter 2013.

All LED Lights May Be Dimmed

Published on LA Confidential, Winter 2013

The advent of light emitting diode (LED) lamps and fixtures for area and spot lighting has created both opportunities and pitfalls. Dimming, in particular, has presented some new challenges.

Many LED product vendors have labeled their units as “dimmable” using standard incandescent dimmers. If the dimming level is relatively minor (e.g., 20%), that may indeed occur successfully. Dimming for energy savings during off-hours, cleaning, or daylighting may, however, involve reductions of 50 to 70%, while architectural dimming for presentations or events may require dimming by 95%. In such cases, dimmed LEDs have demonstrated unacceptable flickering or fluttering.

Unlike incandescent lamps, which have filaments that continue to glow briefly after power is shut off, LEDs fully extinguish very quickly once power to them ceases. Many incandescent dimmers (called “triacs”) chop out parts of the normal alternating sinusoidal current flow (called “phase cutting”). Those interruptions are close enough together in time so that an incandescent filament continues to glow between them, thus avoiding flicker. When an LED lamp experiences those chopped waves, however, it may rapidly flicker.

For some people, that may be simply irritating. For those few with a particular vision disorder, it may induce headaches, fatigue, and possibly convulsions.

In some cases, an LED unit may simply not dim at all: it may just go off, even at a high dimmer setting. In others, LEDs may experience non-linear dimming: a visible reduction in output requires a large adjustment of the dimmer, only to be followed by a big drop in output when the dimmer is turned down a bit further.

To avoid such situations, facility personnel often test sample screw-in LED lamps with existing dimmers across their full range. But note this important point: screwing in only one LED lamp, while the rest remain incandescent, or installing only one LED even when the remaining sockets are empty, may yield falsely acceptable results. Due to the impact that some inexpensive LED units may have on the power wave shape even when not dimmed, a proper test requires re-lamping all sockets controlled on a dimmer circuit with the same type and brand of LEDs. If no flickering occurs, no mixing of brand or type on the same circuit should be allowed thereafter.

Because LED vendors are unlikely to loan their products, such a test may require buying a batch of non-returnable units, only to find that they all flicker on an existing dimmer. To help clarify the situation, the lighting controls industry has developed tables of what types of dimmers being sold today are compatible with various LED products. To deal with incompatibilities, new standards for LED dimming are being developed.

For many applications, it may make more sense to instead replace an existing dimmer with a new unit designed to work with LEDs. Several manufacturers now offer “universal” dimmers designed to work with any lamp labeled “dimmable” (including compact fluorescents).

But be ready for sticker shock. While an incandescent dimmer can be purchased for $15, a fully compatible “universal” dimmer may cost several times that price. Note also that some new dimmers may require a neutral wire at the switch box which may not be present in some older or residential buildings.

Click here for full LA Confidential Winter 2013.