This article was originally published in the September/October 1994 issue of Home Energy Magazine. Some formatting inconsistencies may be evident in older archive content.
| Home Energy Home Page | Back Issues of Home Energy |
Home Energy Magazine Online September/October 1994
Superwindow Retrofits Show Significant Energy Savings
by Mark A. Jackson
Mark A. Jackson is a senior engineer for residential research and demonstration at Bonneville Power Administration in Portland, Oregon.
A pilot project conducted in the Pacific Northwest stimulated demand for high-performance windows, providing cost and energy performance data along the way.
Windows account for a fairly large percentage of the heat loss of houses. Even in new homes built to stringent energy code, windows still account for about 25% of the overall conductive heat loss. In the mid to late 1980's, Bonneville Power Administration (BPA) attempted to increase the energy-efficiency of new homes by setting prescriptive U-values for windows and doors, and setting minimum R-values for other components. In addition, we set an upper limit of 15% of floor area on the allowable area of windows.
As our programs matured, we found that window area is somewhat self-limiting due to cost, so we eliminated the window area restrictions and reduced the new construction reference U-value to 0.35 Btu/hr/ft2deg.F, which is somewhat lower than our current state energy codes. Even at U=0.35, windows still account for a substantial portion of conductive heat loss.
To see if it was technically feasible and economically cost-effective to further improve the efficiency of windows, BPA initiated a pilot superwindow commercialization project in 1992.
Technologies to make residential windows more energy-efficient have advanced tremendously in the past ten years. A decade ago in the Pacific Northwest, triple-glazed wood windows (with U-values usually no lower than 0.40) were considered state-of-the-art and were significantly more expensive than commonly used thermally broken aluminum double-glazed windows with U-values in the range of 0.6-0.7. In the past five years, though, the use of vinyl as a window framing material has increased tremendously, mostly because of its cost-effective energy efficiency. Both wood and vinyl produce frames with U-values of about 0.40. Improvements to low-E glass and the use of Argon as an insulating gas between the panes has resulted in affordable products with overall product U-values in the 0.32 to 0.35 range.
Even-more-efficient superwindows have been commercially available for a couple of years now. Consisting of two low-E coated mylar sheets between two panes of glass, with the interpane spaces filled with Krypton or a Krypton-Argon mixture, the advanced four layer design means low U-values. Yet most available products have been produced with wood frames by manufactures who are familiar with wood. The wood frames have good thermal properties, but the overall products are quite expensive and usually are out of reach for low- and moderate-income homeowners considering window replacement or choosing windows for new construction.
BPA saw a need for affordable high-performance windows for both retrofit and new construction applications to enhance the flexibility of its demand-side management programs. For instance, builders are often reluctant to use advanced framing techniques or use insulating foam sheeting on exterior walls to meet the prescriptive requirements of our Super Good Cents program. By installing superwindows, builders can meet the program's energy budget and still use the standard framing techniques and insulating methods they are comfortable with. BPA also realized that developing and marketing such a product would be expensive for window manufacturers. Through a pilot retrofit project, however, BPA was able to stimulate the market toward higher-performance products, obtaining cost and field performance data in the process.
Based on research performed by Dariush Arasteh at Lawrence Berkeley Laboratory, we knew it would be possible to significantly improve the thermal properties of vinyl-framed windows by using high-performance glazing options, thermally-improved glazing spacers, and low-conductivity gas fills. The LBL research also showed that vinyl-frame performance can be further improved by filling chambers within the frame with insulating materials. We didn't know how much each of these options would cost. But based on the price difference between U=0.40 and U=0.20 wood-framed windows, high-performance glazing options could be a cost-effective improvement for vinyl windows, as well as wood windows.
To provide an incentive for manufacturers to develop lower-cost superwindows, BPA decided to purchase retrofit windows and patio doors for 100 houses. We put a window list for the 100 houses out for bid and specified a target U-value of 0.20. We did not restrict manufacturers to any particular frame material or glazing option, but encouraged them to find the least-cost combination that would meet or approach the target performance level.
We chose the winning windows based on simulated product thermal performance (U-value), using National Fenestration Rating Council (NFRC) procedures and bid price. The chosen window has a vinyl frame, three glazing layers with low-E coating on surfaces 2 and 5 (the exterior glass surface of the exterior pane is surface 1), a krypton gas fill, and standard aluminum glazing spacers. The insulating glass unit is one-inch thick and fits into a standard-width frame. The pyrolytic, or hard-coat low-E, used for these windows, has an emissivity of less than 0.20. Pyrolytic coatings are applied to the surface of glass while the glass is still somewhat molten and the coating becomes part of the glass. The overall average heat loss of the superwindow yields a simulated and thermally tested U-value of 0.23. The manufacturer of the product--Viking Industries, Portland, Oregon--recently began offering the window commercially, along with a soft-coat low-E option that brings the average U-value down to 0.17. Soft-coat low-E coatings are deposited onto the surface of solid glass, and usually have lower emissivities than do hard-coats.
We were comfortable allowing vinyl to compete with other materials for the superwindow retrofits because of vinyl's proven track record in recent years. New material blends, ultraviolet stabilizers, and improvements in quality control have resulted in strong, good-looking and durable products.
Once the manufacturer was selected, actual products were produced and tested for thermal performance, structural properties, air infiltration and water intrusion. NFRC-certified laboratories performed the thermal testing, and the tested U-values were within 0.01 of simulated values. The manufactured product also passed all air, water and structural requirements.
The participating utilities in Clark County and in the Pasco-Kenniwick-Richland area of Washington State were asked to select candidate houses from weatherization waiting lists. We wanted to keep the variability in the housing sample to a minimum, so we intended to limit the study to single-story houses built between 1960 and 1978 with crawlspace foundations, single-glazed windows, and no previous weatherization. There are hundreds of thousands of such houses in the Northwest, so the results of the study are applicable to many residences. Most of the houses ultimately selected for the project met the criteria, but, many houses in the Pasco-Kenniwick-Richland area were accepted into the project with double-glazed aluminum-framed windows.
The climate in Clark County is moderate, with about 5,000 heating degree-days and a 97.5% winter design temperature of 23deg.F. The Pasco-Kenniwick-Richland area can be somewhat cooler at times, but the total heating degree-days are 4,700, with a 97.5% winter design temperature of 11deg.F.
BPA offered to provide the superwindows to the homeowners, and participating utilities offered to pay 40% of the installation costs, under the BPA-sponsored Weatherwise program. In exchange, the homeowners agreed to allow testing to their homes, as needed. The superwindows were installed in selected homes in summer and fall of 1993.
At the onset of this project, we decided to use three different approaches for evaluating the energy performance of the superwindows. The first method used short-term pre- and post-installation testing to establish building thermal characteristics. The second method used field-estimated heat-loss characteristics of envelope components as inputs for thermal simulation modeling. The third method used billing-history analysis to determine the energy impacts. The intent of using three methods was not to determine which method is correct, but rather to help determine the applicability and limitations of each method.
To date, only the short-term testing has been completed, but the results from these tests establish a good baseline for energy performance. Simulations based on site audits have been completed for the sites used for short-term testing. The billing history analysis cannot be completed until one year after superwindow installation, and thermal simulations based on site characteristics will be performed for the remaining houses concurrent with the billing history analysis.
Kris Subbarao of Macrodyne Energy International performed the short-term tests using a technique he refers to as Primary and Secondary Terms Analysis and Renormalization (PSTAR). The PSTAR technique derives inputs to a predictive building energy use model from several days of dynamic performance data. Eight houses in Clark County and twelve in the Tricities were recruited to participate in short-term testing. All Clark County houses originally had single-glazed windows, while most of the houses in the Tricities had double-glazed aluminum windows.
The tests were performed over four days and three nights at each unoccupied house. The first two nights involved electric co-heating tests, where the internal temperature of each house was maintained using computer-controlled electric heaters. The final night of testing measured the cool-down rate of the houses. In addition to numerous indoor and two outside temperature sensors, the testing protocol required global horizontal and principal glazing orientation solar radiation, whole-building power consumption and power consumption outside the building envelope, and relative humidity. We performed blower door tests to determine changes in effective leakage area resulting from superwindow installation, and we measured infiltration directly by monitoring the dilution rate of sulfurhexafluoride tracer gas injected into the air within the houses.
The building load coefficient, effective heat capacitance, and solar heat gain were calculated from the data for each of the 20 houses both before and after superwindow installation. The differences after accounting for normalization factors were differences due to the superwindows.
Results of the field measurements indicate that the difference in U-value between single-glazed aluminum windows and vinyl-framed superwindows is 0.50 +/- 0.04 Btu/hr/ft2deg.F, and the difference between double glazed aluminum framed windows and the superwindows is 0.29 +/- 0.04 Btu/hr/ft2deg.F. These differences are less than the calculated differences, when ASHRAE values (Table 6, chapter 27, 1993 Fundamentals) are used. ASHRAE values suggest that the differences would be 0.88 Btu/hr/ft2deg.F between single-glazed and superwindows, and 0.48 Btu/hr/ft2deg.F between double-glazed aluminum and superwindows. The difference between monitored and predicted U-factors may suggest that some of our old assumptions about window performance in houses are not valid.
Possible explanations for the differences include:
* U-values for a single-glazed window, and to a lesser extent, for a double-glazed window, are sensitive to interior and exterior film coefficients. Recent research has indicated that the values in the ASHRAE Handbook of Fundamentals significantly overpredict single-glazed U-values.
* Bug screens cover a significant fraction of the windows, and these screens have an effect on convective and radiative heat transfer.
* Windows are thinner than the walls in which they are placed, and they tend to be recessed both inside and outside. This affects the surface convection and radiation characteristics, particularly for the single-glazed products.
Some superwindows were mounted so that the amount of window sill, or the setback from the interior wall surface, would be the same for the superwindows as it was for the original windows. Since the superwindows are substantially thicker than the original windows, this installation approach required mounting the new windows so they extended out from the siding material. The external frames were then trimmed with lumber to support and finish the frame. This technique could reduce the performance of the window by increasing the exposed area.
Changes in Leakage Area
We did not expect to see big changes in leakage area resulting from superwindow installation, and field testing seems to indicate that the differences, on average, are minimal. For Clark County houses, the leakage area changed from 108 in2 to 110 in2, and for the Tricities, the leakage rate area from 138 in2 to 131 in2.
Analysis of data from short-term testing provided calibrated hourly simulations of space heating energy flows. From the simulations, we obtained the differences in annual and peak space heating energy use. Table 1 compares calibrated hourly simulations based on short-term testing to simulations based on site audits. ASHRAE values for windows were used for the audit based simulations, and we used actual pre- and post-installation infiltration rates from tracer gas testing to minimize confounding effects.
A good way to see the difference in thermal properties between single-glazed windows and superwindows is to look at thermographs. The sets shown here were taken during fairly warm weather, so the temperature differences are not large, but the heat loss differences are significant nonetheless. Note that white (red) indicates warmer surfaces, and green (black) indicates cooler surfaces. The surface of the superwindow is clearly losing less heat than the single-glazed window, and heat-loss differences between the superwindow frame and the glazed area are also visible.
All of the participating homeowners we talked with were more than pleased with the difference the superwindows made to their homes. From a comfort standpoint, the new windows have a higher inside glass surface temperature, so the body does not see or feel its heat radiating toward a large cold surface, as had been the case with the old windows. The noise transmission reduction due to the superwindows is also noticeable, and the occupants reported that the superwindows blocked the bothersome outside noise, so the homes became much quieter inside.
Probably the most reasonable way to evaluate superwindows as a weatherization or new-construction option is to assess the marginal benefits of purchasing superwindows instead of a less-efficient product. In the retrofit case, this assumes that the existing windows in a house are going to be replaced anyway.
The manufacturer of the vinyl-framed superwindow in our project has now made the superwindows available to the public with either soft-coat or hard-coat low-E glass. We used a version of the window with hard-coat low-E, and can compare the expected costs and monitored performance for that product with the estimated costs and performance for an NFRC certified U=0.40 product, which would meet prescriptive energy code requirements for Oregon and Washington.
The exact price a homeowner or window installer will pay for windows can vary by a substantial amount. Discounts may be given to some customers because of volume, payment track record and other factors, so consumers would be well advised to shop around for the best price for product and installation at any given thermal performance level .
The current price increase for U=0.23 windows relative to U=0.40 windows averages about 60%-70%, depending on size and operator. For a typical home in our study in Clark County Washington, this is an averaged cost increase of about $963.
By scaling the energy savings estimates obtained from the model calibrated by short-term testing, we would expect to see savings of about 2,100 kWh per year from installing U=0.40 windows, and an additional 1,085 kWh per year from installing superwindows.
Electricity rates are low in the northwest, and at $0.06 per kWh, the payback period for installing superwindows is fairly long, about 14 years. If a homeowner plans on living in the house for the life of a typical mortgage, the net present value of the investment in superwindows would be positive, and a good investment, even if electricity rates remain stable and low. In colder climates, in places where space heating costs are greater, and to the extent that superwindows are less expensive in the future, the economics will be even more attractive.
Table 1. Estimated Heating Energy and Peak Energy Use from Superwindow Installation
Reduction Estimates Reduction Estimates (kWh) (kWh) (Based on tests) (Based on Audit) ________________ ________________ Annual Peak Annual Peak __________________________________________________________________ Clark County 3,185 1.5 4,600 2.2 Tricities Area 1,787 1.3 3,000 2.1
The thermograph of this single glazed window shows a great deal of escaping heat. (The white areas on the thermogram indicate hotter temperatures.)
The superwindow shows very little escaping heat in the thermographic photo. In addition to cutting down on leakage, program participants found that the superwindows noticeably reduced noise transmission.
| Back to Contents Page | Home Energy Index | About Home Energy |
Home Energy can be reached at: firstname.lastname@example.org
Home Energy magazine -- Please read our Copyright Notice
- FIRST PAGE
- PREVIOUS PAGE