Which Comes First, Lighting or Heating?

The conclusion of Anil Parekh’s article “Do CFLs Save Whole-House Energy? (HE Nov/Dec ’08, p. 20) is that while CFLs do save energy, the amount they save depends upon location.  A number of contributors to the lighting magazine LD+A have gone even farther, and have claimed that CFLs don’t save energy, or alternatively, money, in Canada (think cold).  What all the writers agree on is that some of the waste heat given off by incandescent lamps is actually useful during the heating season, as it reduces the load on the heating system. The opposite is true during the cooling season.



For heating, the amount of energy saved depends upon the relative heating efficiency of the incandescent lamp versus that of the heating system, and the fraction of lighting energy that is used during the heating season. For cooling, the amount of energy saved depends on the efficiency of the air conditioning system, and the fraction of lighting energy used during the cooling season. The same arguments also apply to appliances that give off heat.

The electricity to power lightbulbs all ends up as heat eventually, but not all of it necessarily ends up inside conditioned space.

It’s Not Whether, But When

Parekh’s article was based on a narrow range of experimental conditions.  It is important to determine the full range of conditions under which efficiency improvements, such as CFLs, actually do save energy and money.  It is also important to determine what needs to be done when they do not save energy or money.  In the following discussion, we show that the efficiency of the thermal system, as determined by the efficiency of the heating and cooling appliances, and the integrity of the building shell, are as important as climate in determining how much whole-house energy CFLs save. CFLs always save energy if the thermal system in the house is efficient enough.  The question is not whether CFLs (or other efficient lamps or appliances) should be installed, but when they should be installed.  In cold climates, it is heating improvements first, and then everything else.

For any individual house design, site, use pattern, and climate, one needs to use a building energy program to estimate the effect of replacing incandescent lamps with CFLs. However, it is possible to get a reasonable estimate for a mix of conditions as a function of climate by using a simple heating degree-day calculation. Degree-days are calculated by dividing days into heating or cooling groups based on whether their average temperature is less than (heating) or more than (cooling) a base temperature. To get heating degree-days, calculate the base temperature minus the average temperature for all heating days, and sum them. The calculation for cooling degree-days is analogous.

A sample of cities covering a range of climates in Canada and the United States shows—no surprise—that Canadian cities are on average much colder and require much more heat than U.S. cities (see Tables 1 and 2). However, it is the length of the heating season—far more than its severity—that determines how much waste heat from an incandescent light, or any other light or appliance, takes up some of the heating load.

The degree-day methodology was developed at a time when houses were not generally insulated, and the 65°F base temperature is the outside temperature below which people in uninsulated houses on average began to heat, and above which they began to cool. The 65°F outside temperature is below the inside temperatures (68°F–78°F) that most people consider comfortable, because internal heat loads, such as lamps, and solar gain, raise the inside temperature above the outside temperature. Adding insulation to a house is not expected to change the base temperature for cooling, because the occupant can simply open a window and let the heat out if it is cooler outside than in. Conversely, calculations indicate that insulation can lower the appropriate heating base temperature to 55°F or less, as opening the window doesn’t help when it is too cold.

Lighting is used more heavily during the dark winter months than during the light summer months. We used a seasonal load profile to estimate the percentage of annual lighting energy use that occurs during the heating season (see Tables 1 and 2).

Cities in Canada generally have much longer heating seasons, and a much higher potential for using waste heat, than cities in the United States, although even in the United States, the potential for using waste heat is substantial in the case of uninsulated houses. This holds true especially for uninsulated houses in the West Coast cities (Vancouver, Seattle, Portland, and San Francisco), because these cities have much longer heating seasons than might be expected from the severity of their climates as measured in degree days. (San Francisco has the third longest heating season of all the cities listed.) However, this effect disappears for insulated houses. Tables 1 and 2 also show that insulation can have a major effect on the length of the heating season, and thus the potential usefulness of waste heat. This is particularly true in milder climates.

The nominal length of the cooling season is 365 minus the length of heating season calculated with the 65°F base. For the Canadian cities, we calculated that the percentage of waste heat that could contribute to cooling load was 13% or less. However, in Canada the added energy benefit from a CFL or an efficient appliance is likely to be less than 13%.  Cooling load depends as much upon solar exposure as it does upon temperature. In relatively cool climates, houses that are properly shaded may not even have air conditioning. This is often the case in many Canadian cities.

The usefulness of the waste heat from a lamp depends upon both the fraction of heat available during the heating season, and the relative heating efficiencies of the electric lights versus the heating system. The electricity to power lightbulbs all ends up as heat eventually, but not all of it necessarily ends up inside conditioned space. The worst case is in older homes when lights are installed in an unsealed recessed fixture that protrudes into an unconditioned attic. The heated fixture can act as a chimney that sucks warm air out of the space, so that the effective heating efficiency of the lightbulb is extremely low, or possibly even negative. In the more usual case, the fixture is sealed, or entirely within the conditioned space. In this situation, the only energy that cannot be used for heat is the energy in the light that escapes through the windows. A window area of 10% of the floor area translates to a window fraction of about 3% of total surface area for a typical house. This gives a minimum utilization efficiency of about 97% during the heating season. If there are curtains, the utilization factor will be a bit higher.

The product of the utilization efficiency and heating season use factors described above gives a seasonal utilization factor of 87%–99% for uninsulated homes in Canada, and 72%–88% for insulated homes. This is an end use efficiency. It can be directly compared to electric-heat efficiencies, but not to fuel-based ones. The burning of fossil fuels generates most of the electricity in North America. Furthermore, in North America, most electric grids are connected. Grid connectivity means that even if a resident gets hydro, nuclear, or even solar-based power, a reduction in that resident’s use of these forms of power means that this power is available to offset fuel-based power elsewhere.

There are four efficiency categories for electric heating, three for fuel-based heating, and two for cooling. The four efficiency categories for electric heating are (1) electric furnace with leaky ducts – 70%; (2) baseboard heaters or electric furnace with sealed ducts – 100%;  (3) old low-efficiency heat pump, heating seasonal performance factor  (HSPF) – 5);  and (4) new high-efficiency Energy Star heat pump, HSPF – 9.  For fuel-based heating, the three efficiency categories are (1) old furnace with  leaky ducts – 49%; (2) old furnace with sealed ducts, or new furnace with somewhat leaky ducts – 70%; and (3) new furnace with sealed ducts – 90%. Finally, for cooling the two efficiency categories are (1) old low-efficiency air conditioner, SEER – 8; and (2) new, high-efficiency air conditioner, SEER – 16. Table 3 shows the expected energy savings or losses to be realized by switching to CFLs under each of the seven heating efficiency categories and the two air conditioning efficiency categories, with and without insulation.

As you can see from Table 3, there are cases, especially in Canada, where switching to the more-efficient CFLs actually increases energy use.  However, energy use only increases when the house is heated with an electric furnace that has leaky ducts (low efficiency). The message here is that CFLs increase energy use only when the heating system needs replacement or repair. It is not just climate that matters. Even for the most severe case in Canada, baseboard heating in an insulated house will allow for 15%–16% of the nominal savings from the CFL, while switching to an efficient heat pump will save almost 50%, and using an efficient fuel-based heating system can save up to 70% (100% is the nominal savings when you ignore heating or cooling effects).

In Canada, the CFL savings in an insulated, energy-efficient house averages from 60% to 77% of nominal savings. In the United States, the average ranges from 81% to 98%. Yes, the effects of heating offset can be very significant, but they are not so significant that they eliminate the possibility of savings. The fact that some of the savings from CFLs is offset during the heating season is not an excuse to avoid using CFLs. It does, however, indicate that the first order of business in heating climates is to upgrade the heating system and insulate the structure as much as possible.

Differenciating Costs

Table 3 addresses only energy. It does not address cost. Parekh’s article gives several examples of ways in which the cost of heating and cooling affects the cost savings realized by switching to CFLs. All of Parekh’s  examples show at least some cost savings. The issue here is differences in the cost of energy.

To get a better handle on this issue, we present the following example. A 60W incandescent lamp with an expected life of 1,000 hours and a first cost of 50¢ per lamp is replaced by a 15W  CFL with an expected life of 4,000 hours (50% of the nominal life) and a cost of $5. Over the life of the CFL, there is a nominal savings of 180 kWh, at an added capital cost of $3.  If heating and cooling effects are ignored, the lamp just pays for itself if electricity costs 1.67¢/kWh. The purchaser makes a 100% profit ($4.50 + $3.00) if electricity costs 4.17¢/kWh. Electricity costs vary widely, due to differences in the cost of production, and to differences in rate schedules that take account of social, environmental, and planning issues. For example, in 2008, Pacific Gas and Electric Company (PG&E), a large utility in California, charged low-income customers as little as 5¢/kWh, while regular customers were charged from 11.6¢/kWh for the lowest usage to 41¢/kWh for the highest usage.  

Figure 1 shows the energy saving percentage (the values listed in Table 3) for the 15W CFL, for breakeven and for 100% profit, as a function of the electricity cost for an electrically heated home. The curves cross the 100% line at 1.67¢/kWh and 4.7¢/kWh, and decline as increasing price gives the same return, even when only some energy is saved. For an uninsulated house in Vancouver (the minimum case in Table 3 for this situation), with baseboard electric heat, the break-even point is 40¢/kWh. The first priority in such a structure is to improve the insulation or the heating system. For the worst-case insulated house (St. John’s), with baseboard electric heat, the break-even point is 11¢/kWh (Vancouver is only 6¢/kWh). Going one more step and converting to a high-efficiency heat pump lowers the break-even point down to 4¢/kWh, and the 100% profit point down to 9¢/kWh.  (In Vancouver, these figures are all the way down to 2¢ and 6¢ respectively).

Adding air conditioning adds an extra dimension to an already complex graph, so we will simply note that the electricity prices to give breakeven or profit are lower with air conditioning than without.

The highest fuel costs are high enough that in the worst-case Canadian scenario, the break-even cost for an insulated house with a high-efficiency furnace is 11¢/kWh, while the break-even cost for an uninsulated house with a low-efficiency heating system is about 20¢/kWh. Clearly, if fuel costs are high enough, especially relative to electricity costs, switching to CFLs may not always be cost-effective.

If absolute fuel costs are high, even higher levels of insulation than were assumed here become cost-effective, which again shifts the cost balance. Note too that even with these extreme costs, the break-even cost for the average Canadian climate with an insulated house and a high-efficiency furnace is down to 9.5¢/kWh, and the 100% profit cost is only 12¢/kWh. In the United States, the extreme case (Minneapolis) is no worse than the average Canadian city, while in the average case, the 100% profit cost is only 9¢/kWh. More common fuel costs dramatically improve the economics of CFL conversion. The worst-case uninsulated house has a break-even cost of 11.5¢/kWh, while the insulated house with a high-efficiency furnace has a break-even cost of less than 6.5¢/kWh.

What can we learn from all these calculations? If there is any question as to whether a CFL conversion will save energy, it means that either the heating system needs to be repaired or replaced, or the insulation needs to be upgraded, or both. Once these things are done, the CFL conversion will save energy. If the climate is so mild that heating and insulation upgrades are not cost-effective, a CFL conversion will almost certainly save energy even without them.  Much the same thing holds true if there is any question as to whether the conversion will save money.  If fuel prices rose much higher than electricity prices, especially in a very cold climate like that of St. John’s, the conversion might not save money; but this situation would probably not last long, because electricity and fuel process tend to rise and fall together.

There is also the question of frequency of use. If you use a lamp so seldom that you expect it to last your lifetime, it might make more economic sense to put your money in the bank—even at today’s low interest rates. In general, however, it makes good economic sense to replace any commonly used incandescent with a CFL, even in Canada, once the appropriate heating and insulation upgrades have been made.  

Robert Clear studies energy and lighting and Francis Rubinstein does research in lighting systems in the Environmental Energy Technology Division at Lawrence Berkeley National Laboratory.

>> For more information:
To learn more about the effect of climate on heat pump performance, see Fairey, Philip, et al. Climate Impacts on Heating Seasonal Performance Factor (HSPF) and Seasonal Energy Efficiency Ratio (SEER) for Air Source Heat Pumps. Cocoa, Florida: Florida Solar Energy Center, 2007. To download this report, go to www.fsec.ucf.edu/en/publications/html/FSEC-PF-413-04/.

• 1
• FIRST PAGE
• PREVIOUS PAGE
• NEXT
• LAST